[PATCH] staging: dgnc: fix 'line over 80 characters'
fix checkpatch.pl warning about 'line over 80 characters'. Signed-off-by: Anson Jacob--- drivers/staging/dgnc/dgnc_driver.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/staging/dgnc/dgnc_driver.h b/drivers/staging/dgnc/dgnc_driver.h index 3092abe..88d2696 100644 --- a/drivers/staging/dgnc/dgnc_driver.h +++ b/drivers/staging/dgnc/dgnc_driver.h @@ -390,7 +390,7 @@ extern uint dgnc_major; /* Our driver/mgmt major */ extern int dgnc_poll_tick; /* Poll interval - 20 ms */ extern spinlock_t dgnc_global_lock; /* Driver global spinlock */ extern spinlock_t dgnc_poll_lock; /* Poll scheduling lock */ -extern uintdgnc_num_boards;/* Total number of boards */ +extern uintdgnc_num_boards;/* Total number of boards */ extern struct dgnc_board *dgnc_board[MAXBOARDS]; /* Array of board * structs */ -- 2.7.4 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH] staging: lustre: fix checkpatch error
convert spaces to tab fix the following error messages from checkpatch.pl ERROR: code indent should use tabs where possible WARNING: please, no spaces at the start of a line Signed-off-by: Anson Jacob--- drivers/staging/lustre/lustre/include/cl_object.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h index 3cd4a25..80c13c9 100644 --- a/drivers/staging/lustre/lustre/include/cl_object.h +++ b/drivers/staging/lustre/lustre/include/cl_object.h @@ -2354,7 +2354,7 @@ struct cl_client_cache { * Waitq for awaiting unstable pages to reach zero. * Used at umounting time and signaled on BRW commit */ -wait_queue_head_t ccc_unstable_waitq; + wait_queue_head_t ccc_unstable_waitq; }; -- 2.7.4 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [PATCH v2 RESEND 2/4] Drivers: hv: balloon: account for gaps in hot add regions
"Alex Ng (LIS)"writes: >> @@ -676,35 +686,63 @@ static void hv_mem_hot_add(unsigned long start, >> unsigned long size, >> >> static void hv_online_page(struct page *pg) { >> -struct list_head *cur; >> struct hv_hotadd_state *has; >> +struct hv_hotadd_gap *gap; >> unsigned long cur_start_pgp; >> unsigned long cur_end_pgp; >> +bool is_gap = false; >> >> list_for_each(cur, _device.ha_region_list) { >> has = list_entry(cur, struct hv_hotadd_state, list); >> cur_start_pgp = (unsigned long) >> +pfn_to_page(has->start_pfn); >> +cur_end_pgp = (unsigned long)pfn_to_page(has->end_pfn); >> + >> +/* The page belongs to a different HAS. */ >> +if (((unsigned long)pg < cur_start_pgp) || >> +((unsigned long)pg >= cur_end_pgp)) >> +continue; >> + >> +cur_start_pgp = (unsigned long) >> pfn_to_page(has->covered_start_pfn); >> cur_end_pgp = (unsigned long)pfn_to_page(has- >> >covered_end_pfn); >> >> -if (((unsigned long)pg >= cur_start_pgp) && >> -((unsigned long)pg < cur_end_pgp)) { >> -/* >> - * This frame is currently backed; online the >> - * page. >> - */ >> -__online_page_set_limits(pg); >> -__online_page_increment_counters(pg); >> -__online_page_free(pg); >> +/* The page is not backed. */ >> +if (((unsigned long)pg < cur_start_pgp) || >> +((unsigned long)pg >= cur_end_pgp)) >> +continue; >> + >> +/* Check for gaps. */ >> +list_for_each_entry(gap, >gap_list, list) { >> +cur_start_pgp = (unsigned long) >> +pfn_to_page(gap->start_pfn); >> +cur_end_pgp = (unsigned long) >> +pfn_to_page(gap->end_pfn); >> +if (((unsigned long)pg >= cur_start_pgp) && >> +((unsigned long)pg < cur_end_pgp)) { >> +is_gap = true; >> +break; >> +} >> } >> + >> +if (is_gap) >> +break; >> + >> +/* This frame is currently backed; online the page. */ >> +__online_page_set_limits(pg); >> +__online_page_increment_counters(pg); >> +__online_page_free(pg); >> +break; >> } >> } >> > > We may need to add similar logic to check for gaps in hv_bring_pgs_online(). > > [...] Yes, probably, I'll take a look and try to refactor the onlinig code in a separate function to avoid duplication. >> static unsigned long handle_pg_range(unsigned long pg_start, @@ -834,13 >> +881,19 @@ static unsigned long process_hot_add(unsigned long pg_start, >> unsigned long rg_size) >> { >> struct hv_hotadd_state *ha_region = NULL; >> +int covered; >> >> if (pfn_cnt == 0) >> return 0; >> >> -if (!dm_device.host_specified_ha_region) >> -if (pfn_covered(pg_start, pfn_cnt)) >> +if (!dm_device.host_specified_ha_region) { >> +covered = pfn_covered(pg_start, pfn_cnt); >> +if (covered < 0) >> +return 0; > > If the hot-add pages aren't covered by any region, then shouldn't it fall > through instead of returning? > That way the new ha_region can be added to the list and we hot-add the > pages accordingly. I was under an impression this is impossible: hot_add_req()/process_hot_add() will create a new region in this case. 'covered < 0' was added to handle one particular error: failure to allocate memory to record gap (struct hv_hotadd_gap) and I don't have a better idea how to handle this: if we can't remember the gap we'll crash later on onlining... -- Vitaly ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [PATCH 1/2] staging: lustre: Add include path to Makefile
On Mon, Aug 15, 2016 at 02:14:18PM -0700, Joe Perches wrote: > On Mon, 2016-08-15 at 23:04 +0200, Greg Kroah-Hartman wrote: > > On Mon, Aug 15, 2016 at 12:33:23PM -0700, Joe Perches wrote: > > > Start to rationalize include paths in source code files. > [] > > > diff --git a/drivers/staging/lustre/Makefile > > > b/drivers/staging/lustre/Makefile > [] > > > @@ -1,2 +1,5 @@ > > > +subdir-ccflags-y += -I$(srctree)/drivers/staging/lustre/include/ > > > +subdir-ccflags-y += -I$(srctree)/drivers/staging/lustre/lustre/include/ > > > + > > > obj-$(CONFIG_LNET) += lnet/ > > > obj-$(CONFIG_LUSTRE_FS) += lustre/ > > This is good, but does this break the subdir make command: > > make M=drivers/staging/lustre/foo_dir/ > > ? > > hmm, yeah, it does. Oh well, nevermind for awhile. > > > I remember the last time I tried to clean this up, it took a while... > > It seems like something the build tools should > handle correctly now, but I'll look at it. Nah, I think it's fine the way it is. What's wrong with the "../" type things in the include directives? The build system works well with it, you can build any single module or file correctly, and it's obvious what the code is including from. Now you might argue that the lustre code shouldn't have such a mess of include files like this, and I would totally agree. I suggest working on reducing the number of .h files they have, and centralizing them such that all of these cross-directory includes aren't needed anymore either in the build directives, or in the .c files. thanks, greg k-h ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [PATCH 2/9] staging: most: hdm-usb: fix race between enqueue and most_stop_enqueue
On Mon, 15 Aug 2016 20:39:38 +0200 Greg KHwrote: > On Mon, Jul 18, 2016 at 05:25:28PM +0200, Christian Gromm wrote: > > The "broken in pipe" handler of the USB-HDM calls most_stop_enqueue() to > > stop the MBO traffic before returning all MBOs back to the Mostcore. As > > the enqueue() call from the Mostcore may run in parallel with the > > most_stop_enqueue(), the HDM may run into the inconsistent state and > > crash the kernel. > > > > This patch synchronizes enqueue(), most_stop_enqueue() and > > most_resume_enqueue() with a mutex, hence avoiding the race condition. > > > > Signed-off-by: Andrey Shvetsov > > Signed-off-by: Christian Gromm > > --- > > drivers/staging/most/hdm-usb/hdm_usb.c | 3 +- > > drivers/staging/most/mostcore/core.c | 53 > > -- > > 2 files changed, 38 insertions(+), 18 deletions(-) > > Doesn't apply to my tree anymore. Can you fix this up and resend it and > the rest in this series so I can apply them? According to git you didn't apply the patchset I sent in on Friday 6/17/16, did you? >>> "[PATCH 0/9] staging: most: fix hdm-usb issues" <<< Those patches need to be applied before the latest set. This is the cover letter of the patchset: This patch set is needed to fix issues of the hdm-usb module. Christian Gromm (9): staging: most: hdm-usb: remove unused macro HW_RESYNC staging: most: hdm-usb: provide MBO status when freeing buffers staging: most: hdm-usb: fix clear halt processing staging: most: hdm-usb: stop core from submitting buffers in case of broken pipe staging: most: hdm-usb: add USB product id staging: most: hdm-usb: rename ID_INIC for consistency staging: most: hdm-usb: make use of is_channel_healthy flag staging: most: hdm-usb: remove redundant conditions staging: most: hdm-usb: simplify initialization of mbo->status. drivers/staging/most/hdm-usb/hdm_usb.c | 109 - 1 file changed, 52 insertions(+), 57 deletions(-) -- 1.9.1 Should I resend those or do still have them? regards, Chris > > thanks, > > greg k-h ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [PATCH 1/2] staging: lustre: Add include path to Makefile
On Tue, 2016-08-16 at 12:53 +0200, Greg Kroah-Hartman wrote: > On Mon, Aug 15, 2016 at 02:14:18PM -0700, Joe Perches wrote: > > On Mon, 2016-08-15 at 23:04 +0200, Greg Kroah-Hartman wrote: > > > On Mon, Aug 15, 2016 at 12:33:23PM -0700, Joe Perches wrote: > > > > Start to rationalize include paths in source code files. [] > Nah, I think it's fine the way it is. What's wrong with the "../" type > things in the include directives? .. path specific includes are unnecessarily file layout sensitive. There are identically named .h files in separate directories. The idea was to isolate the uapi type .h files from local ones. I'll just give this up and let the lustre folks handle it. ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 29/80] staging: lustre: lmv: lookup remote migrating object in LMV
From: wang diIf remote object is being found in a migrating directory, it should continue to lookup the object in remote MDT, instead of return. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4805 Reviewed-on: http://review.whamcloud.com/9806 Reviewed-by: John L. Hammond Reviewed-by: Nathaniel Clark Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_intent.c |1 - 1 files changed, 0 insertions(+), 1 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c index 51b7048..a38d343 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c @@ -471,7 +471,6 @@ static int lmv_intent_lookup(struct obd_export *exp, it->it_disposition &= ~DISP_ENQ_COMPLETE; rc = md_intent_lock(tgt->ltd_exp, op_data, lmm, lmmsize, it, flags, reqp, cb_blocking, extra_lock_flags); - return rc; } /* -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 17/80] staging: lustre: add ability to migrate inodes.
From: wang diAdd client support to migrate the individual inodes from one MDT to another MDT, and this functionality will only migrate inode layout on MDT but not touch data object on OST. The directory will be migrated from top to the bottom, i.e. migrating parent first, then migrating the child. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2430 Reviewed-on: http://review.whamcloud.com/6662 Reviewed-by: Alex Zhuravlev Reviewed-by: John L. Hammond Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h | 11 +- .../lustre/lustre/include/lustre/lustre_user.h |1 + drivers/staging/lustre/lustre/include/lustre_lmv.h |2 + drivers/staging/lustre/lustre/include/obd.h| 12 +-- drivers/staging/lustre/lustre/llite/dir.c | 43 +- drivers/staging/lustre/lustre/llite/file.c | 113 - .../staging/lustre/lustre/llite/llite_internal.h | 14 ++- drivers/staging/lustre/lustre/llite/llite_lib.c| 33 - drivers/staging/lustre/lustre/llite/namei.c|3 +- drivers/staging/lustre/lustre/llite/rw.c |4 + drivers/staging/lustre/lustre/llite/statahead.c|1 + drivers/staging/lustre/lustre/lmv/lmv_intent.c | 32 +++- drivers/staging/lustre/lustre/lmv/lmv_obd.c| 176 ++-- drivers/staging/lustre/lustre/mdc/mdc_lib.c|2 + drivers/staging/lustre/lustre/ptlrpc/wiretest.c|4 +- 15 files changed, 368 insertions(+), 83 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 0ff30c6..6853f62 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -1482,6 +1482,7 @@ enum obdo_flags { #define LOV_MAGIC LOV_MAGIC_V1 #define LOV_MAGIC_JOIN_V1 0x0BD20BD0 #define LOV_MAGIC_V3 0x0BD30BD0 +#define LOV_MAGIC_MIGRATE 0x0BD40BD0 /* * magic for fully defined striping @@ -1987,7 +1988,7 @@ enum mdt_reint_cmd { REINT_OPEN = 6, REINT_SETXATTR = 7, REINT_RMENTRY = 8, -/* REINT_WRITE= 9, */ + REINT_MIGRATE = 9, REINT_MAX }; @@ -2280,6 +2281,7 @@ enum mds_op_bias { MDS_CREATE_VOLATILE = 1 << 10, MDS_OWNEROVERRIDE = 1 << 11, MDS_HSM_RELEASE = 1 << 12, + MDS_RENAME_MIGRATE = BIT(13), }; /* instance of mdt_reint_rec */ @@ -2488,11 +2490,13 @@ struct lmv_desc { /* lmv structures */ #define LMV_MAGIC_V1 0x0CD10CD0 /* normal stripe lmv magic */ #define LMV_USER_MAGIC 0x0CD20CD0 /* default lmv magic*/ +#define LMV_MAGIC_MIGRATE 0x0CD30CD0 /* migrate stripe lmv magic */ #define LMV_MAGIC LMV_MAGIC_V1 enum lmv_hash_type { LMV_HASH_TYPE_ALL_CHARS = 1, LMV_HASH_TYPE_FNV_1A_64 = 2, + LMV_HASH_TYPE_MIGRATION = 3, }; #define LMV_HASH_NAME_ALL_CHARS"all_char" @@ -2552,7 +2556,8 @@ static inline ssize_t lmv_mds_md_size(int stripe_count, unsigned int lmm_magic) ssize_t len = -EINVAL; switch (lmm_magic) { - case LMV_MAGIC_V1: { + case LMV_MAGIC_V1: + case LMV_MAGIC_MIGRATE: { struct lmv_mds_md_v1 *lmm1; len = sizeof(*lmm1); @@ -2568,6 +2573,7 @@ static inline int lmv_mds_md_stripe_count_get(const union lmv_mds_md *lmm) { switch (le32_to_cpu(lmm->lmv_magic)) { case LMV_MAGIC_V1: + case LMV_MAGIC_MIGRATE: return le32_to_cpu(lmm->lmv_md_v1.lmv_stripe_count); case LMV_USER_MAGIC: return le32_to_cpu(lmm->lmv_user_md.lum_stripe_count); @@ -2583,6 +2589,7 @@ static inline int lmv_mds_md_stripe_count_set(union lmv_mds_md *lmm, switch (le32_to_cpu(lmm->lmv_magic)) { case LMV_MAGIC_V1: + case LMV_MAGIC_MIGRATE: lmm->lmv_md_v1.lmv_stripe_count = cpu_to_le32(stripe_count); break; case LMV_USER_MAGIC: diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h index 26dbda0..4746320 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h @@ -243,6 +243,7 @@ struct ost_id { #define LL_IOC_GET_LEASE _IO('f', 244) #define LL_IOC_HSM_IMPORT _IOWR('f', 245, struct hsm_user_import) #define LL_IOC_LMV_SET_DEFAULT_STRIPE _IOWR('f', 246, struct lmv_user_md) +#define LL_IOC_MIGRATE _IOR('f', 247, int) #define LL_STATFS_LMV 1 #define LL_STATFS_LOV 2 diff --git a/drivers/staging/lustre/lustre/include/lustre_lmv.h
[PATCH 19/80] staging: lustre: libcfs: Only dump log once per sec. to avoid EEXIST
From: Ryan HaaskenSince the log file name contains the current time in seconds, dumping the logs more than once per second causes EEXIST errors to be emitted. Add a static variable to libcfs_debug_dumplog_internal that records the time of the last Lustre log dump. If the current time in seconds is equal to the last time, do not dump logs again. Note that this is not thread-safe. However, in the rare case that two threads try to access last_dump_time simultaneously, the worst thing that could happen is that one of the threads will get an EEXIST error when trying to write the log file. This is no worse than the current situation, and it is not likely to happen. Signed-off-by: Ryan Haasken Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4129 Reviewed-on: http://review.whamcloud.com/8964 Reviewed-by: Andreas Dilger Reviewed-by: Bob Glossman Signed-off-by: James Simmons --- drivers/staging/lustre/lnet/libcfs/debug.c |9 +++-- 1 files changed, 7 insertions(+), 2 deletions(-) diff --git a/drivers/staging/lustre/lnet/libcfs/debug.c b/drivers/staging/lustre/lnet/libcfs/debug.c index 42b15a7..23b36b8 100644 --- a/drivers/staging/lustre/lnet/libcfs/debug.c +++ b/drivers/staging/lustre/lnet/libcfs/debug.c @@ -328,15 +328,20 @@ libcfs_debug_str2mask(int *mask, const char *str, int is_subsys) */ void libcfs_debug_dumplog_internal(void *arg) { + static time64_t last_dump_time; + time64_t current_time; void *journal_info; journal_info = current->journal_info; current->journal_info = NULL; + current_time = ktime_get_real_seconds(); - if (strncmp(libcfs_debug_file_path_arr, "NONE", 4) != 0) { + if (strncmp(libcfs_debug_file_path_arr, "NONE", 4) && + current_time > last_dump_time) { + last_dump_time = current_time; snprintf(debug_file_name, sizeof(debug_file_name) - 1, "%s.%lld.%ld", libcfs_debug_file_path_arr, -(s64)ktime_get_real_seconds(), (long_ptr_t)arg); +(s64)current_time, (long_ptr_t)arg); pr_alert("LustreError: dumping log to %s\n", debug_file_name); cfs_tracefile_dump_all_pages(debug_file_name); libcfs_run_debug_log_upcall(debug_file_name); -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 51/80] staging: lustre: lmv: Match MDT where the FID locates first
From: wang diWith DNE every object can have two locks in different namespaces: lookup lock in space of MDT storing direntry and update/open lock in space of MDT storing inode. In lmv_find_cbdata/lmv_lock_lock, it should try the MDT that the FID maps to first, since this can be easily found, and only try others if that fails. In the error handler of lmv_add_targets, it should check whether ld_tgt_count is being increased before ld_tgt_count is being -1. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4098 Reviewed-on: http://review.whamcloud.com/8019 Reviewed-by: Andreas Dilger Reviewed-by: Fan Yong Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_internal.h | 45 +- drivers/staging/lustre/lustre/lmv/lmv_obd.c | 57 +++--- 2 files changed, 73 insertions(+), 29 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_internal.h b/drivers/staging/lustre/lustre/lmv/lmv_internal.h index dbd1da6..faf6a7b 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_internal.h +++ b/drivers/staging/lustre/lustre/lmv/lmv_internal.h @@ -64,35 +64,56 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody, int extra_lock_flags); static inline struct lmv_tgt_desc * -lmv_get_target(struct lmv_obd *lmv, u32 mds) +lmv_get_target(struct lmv_obd *lmv, u32 mdt_idx, int *index) { - int count = lmv->desc.ld_tgt_count; int i; - for (i = 0; i < count; i++) { + for (i = 0; i < lmv->desc.ld_tgt_count; i++) { if (!lmv->tgts[i]) continue; - if (lmv->tgts[i]->ltd_idx == mds) + if (lmv->tgts[i]->ltd_idx == mdt_idx) { + if (index) + *index = i; return lmv->tgts[i]; + } } return ERR_PTR(-ENODEV); } -static inline struct lmv_tgt_desc * -lmv_find_target(struct lmv_obd *lmv, const struct lu_fid *fid) +static inline int +lmv_find_target_index(struct lmv_obd *lmv, const struct lu_fid *fid) { - u32 mds = 0; - int rc; + struct lmv_tgt_desc *ltd; + u32 mdt_idx = 0; + int index = 0; if (lmv->desc.ld_tgt_count > 1) { - rc = lmv_fld_lookup(lmv, fid, ); - if (rc) - return ERR_PTR(rc); + int rc; + + rc = lmv_fld_lookup(lmv, fid, _idx); + if (rc < 0) + return rc; } - return lmv_get_target(lmv, mds); + ltd = lmv_get_target(lmv, mdt_idx, ); + if (IS_ERR(ltd)) + return PTR_ERR(ltd); + + return index; +} + +static inline struct lmv_tgt_desc * +lmv_find_target(struct lmv_obd *lmv, const struct lu_fid *fid) +{ + int index; + + index = lmv_find_target_index(lmv, fid); + if (index < 0) + return ERR_PTR(index); + + return lmv->tgts[index]; } static inline int lmv_stripe_md_size(int stripe_count) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index 9821f69..6917a03 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -480,6 +480,7 @@ static int lmv_add_target(struct obd_device *obd, struct obd_uuid *uuidp, { struct lmv_obd *lmv = >u.lmv; struct lmv_tgt_desc *tgt; + int orig_tgt_count = 0; int rc = 0; CDEBUG(D_CONFIG, "Target uuid: %s. index %d\n", uuidp->uuid, index); @@ -549,14 +550,17 @@ static int lmv_add_target(struct obd_device *obd, struct obd_uuid *uuidp, tgt->ltd_uuid = *uuidp; tgt->ltd_active = 0; lmv->tgts[index] = tgt; - if (index >= lmv->desc.ld_tgt_count) + if (index >= lmv->desc.ld_tgt_count) { + orig_tgt_count = lmv->desc.ld_tgt_count; lmv->desc.ld_tgt_count = index + 1; + } if (lmv->connected) { rc = lmv_connect_mdc(obd, tgt); if (rc) { spin_lock(>lmv_lock); - lmv->desc.ld_tgt_count--; + if (lmv->desc.ld_tgt_count == index + 1) + lmv->desc.ld_tgt_count = orig_tgt_count; memset(tgt, 0, sizeof(*tgt)); spin_unlock(>lmv_lock); } else { @@ -1263,7 +1267,7 @@ int __lmv_fid_alloc(struct lmv_obd *lmv, struct lu_fid *fid, u32 mds) struct lmv_tgt_desc *tgt; int rc; - tgt = lmv_get_target(lmv, mds); + tgt = lmv_get_target(lmv, mds, NULL); if (IS_ERR(tgt)) return PTR_ERR(tgt); @@ -1610,6 +1614,7 @@ static int lmv_find_cbdata(struct
[PATCH 38/80] staging: lustre: move ioctls to lustre_ioctl.h
From: John L. HammondMove ioctl definitions and related functions from lustre_dlm.h, lustre_lib.h, obd.h, to lustre_ioctl.h. Replace the definitions of retired ioctls with comment. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4961 Reviewed-on: http://review.whamcloud.com/10139 Reviewed-by: Andreas Dilger Reviewed-by: Robert Read Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_ioctl.h| 412 .../lustre/lustre/include/lustre/lustre_user.h | 21 +- drivers/staging/lustre/lustre/include/lustre_dlm.h | 10 - drivers/staging/lustre/lustre/include/lustre_lib.h | 284 -- drivers/staging/lustre/lustre/include/obd.h| 10 - drivers/staging/lustre/lustre/llite/dir.c |9 +- drivers/staging/lustre/lustre/llite/file.c |1 + drivers/staging/lustre/lustre/llite/llite_lib.c|1 + drivers/staging/lustre/lustre/lmv/lmv_obd.c|1 + drivers/staging/lustre/lustre/lov/lov_obd.c|1 + drivers/staging/lustre/lustre/mdc/mdc_request.c|1 + drivers/staging/lustre/lustre/obdclass/class_obd.c |8 +- .../lustre/lustre/obdclass/linux/linux-module.c|1 + .../staging/lustre/lustre/obdclass/obd_config.c|1 + .../staging/lustre/lustre/obdecho/echo_client.c|1 + drivers/staging/lustre/lustre/osc/osc_request.c|1 + 16 files changed, 430 insertions(+), 333 deletions(-) create mode 100644 drivers/staging/lustre/lustre/include/lustre/lustre_ioctl.h diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_ioctl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_ioctl.h new file mode 100644 index 000..f3d7c94 --- /dev/null +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_ioctl.h @@ -0,0 +1,412 @@ +/* + * GPL HEADER START + * + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 only, + * as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License version 2 for more details (a copy is included + * in the LICENSE file that accompanied this code). + * + * You should have received a copy of the GNU General Public License + * version 2 along with this program; If not, see + * http://www.gnu.org/licenses/gpl-2.0.html + * + * GPL HEADER END + */ +/* + * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved. + * Use is subject to license terms. + * + * Copyright (c) 2011, 2015, Intel Corporation. + */ +#ifndef LUSTRE_IOCTL_H_ +#define LUSTRE_IOCTL_H_ + +#include +#include "../../../include/linux/libcfs/libcfs.h" +#include "lustre_idl.h" + +#ifdef __KERNEL__ +# include +# include +# include "../obd_support.h" +#else /* __KERNEL__ */ +# include +# include +#include +#endif /* !__KERNEL__ */ + +#if !defined(__KERNEL__) && !defined(LUSTRE_UTILS) +# error This file is for Lustre internal use only. +#endif + +enum md_echo_cmd { + ECHO_MD_CREATE = 1, /* Open/Create file on MDT */ + ECHO_MD_MKDIR = 2, /* Mkdir on MDT */ + ECHO_MD_DESTROY = 3, /* Unlink file on MDT */ + ECHO_MD_RMDIR = 4, /* Rmdir on MDT */ + ECHO_MD_LOOKUP = 5, /* Lookup on MDT */ + ECHO_MD_GETATTR = 6, /* Getattr on MDT */ + ECHO_MD_SETATTR = 7, /* Setattr on MDT */ + ECHO_MD_ALLOC_FID = 8, /* Get FIDs from MDT */ +}; + +#define OBD_DEV_ID 1 +#define OBD_DEV_NAME "obd" +#define OBD_DEV_PATH "/dev/" OBD_DEV_NAME +#define OBD_DEV_MAJOR 10 +#define OBD_DEV_MINOR 241 + +#define OBD_IOCTL_VERSION 0x00010004 +#define OBD_DEV_BY_DEVNAME 0xd0de +#define OBD_MAX_IOCTL_BUFFER CONFIG_LUSTRE_OBD_MAX_IOCTL_BUFFER + +struct obd_ioctl_data { + __u32 ioc_len; + __u32 ioc_version; + + union { + __u64 ioc_cookie; + __u64 ioc_u64_1; + }; + union { + __u32 ioc_conn1; + __u32 ioc_u32_1; + }; + union { + __u32 ioc_conn2; + __u32 ioc_u32_2; + }; + + struct obdo ioc_obdo1; + struct obdo ioc_obdo2; + + __u64 ioc_count; + __u64 ioc_offset; + __u32 ioc_dev; + __u32 ioc_command; + + __u64 ioc_nid; + __u32 ioc_nal; + __u32 ioc_type; + + /* buffers the kernel will treat as user pointers */ + __u32 ioc_plen1; +
[PATCH 44/80] staging: lustre: Remove static declaration in anonymous union
From: Christopher J. MorroneIt is not permitted in C++ to have a static declaration inside of an anonymous union. The g++ compiler will complaine with an error like this: error: struct ost_idostid invalid; an anonymous union can only have non-static data members [-fpermissive] This patch changes the code to use an unnamed struct in place of "struct ostid" inside of the anonymous union. That name declaration was completely unnecessary anyway, since it was not used anywhere else. Signed-off-by: Christopher J. Morrone Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4987 Reviewed-on: http://review.whamcloud.com/10176 Reviewed-by: Robert Read Reviewed-by: Andreas Dilger Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_user.h |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h index 4b2553c..59d45de 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h @@ -167,7 +167,7 @@ struct lustre_mdt_attrs { */ struct ost_id { union { - struct ostid { + struct { __u64 oi_id; __u64 oi_seq; } oi; -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 01/80] staging: lustre: llite: add md_op_data parameter to ll_get_dir_page
From: wang diPass in struct md_op_data for ll_get_dir_page function. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/dir.c |8 .../staging/lustre/lustre/llite/llite_internal.h |4 ++-- drivers/staging/lustre/lustre/llite/statahead.c| 15 +++ 3 files changed, 17 insertions(+), 10 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index 031c9e4..82c7f88 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -322,8 +322,8 @@ static struct page *ll_dir_page_locate(struct inode *dir, __u64 *hash, return page; } -struct page *ll_get_dir_page(struct inode *dir, __u64 hash, -struct ll_dir_chain *chain) +struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data, +__u64 hash, struct ll_dir_chain *chain) { ldlm_policy_data_t policy = {.l_inodebits = {MDS_INODELOCK_UPDATE} }; struct address_space *mapping = dir->i_mapping; @@ -503,7 +503,7 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, ll_dir_chain_init(); - page = ll_get_dir_page(inode, pos, ); + page = ll_get_dir_page(inode, op_data, pos, ); while (rc == 0 && !done) { struct lu_dirpage *dp; @@ -585,7 +585,7 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, le32_to_cpu(dp->ldp_flags) & LDF_COLLIDE); next = pos; - page = ll_get_dir_page(inode, pos, + page = ll_get_dir_page(inode, op_data, pos, ); } } diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h index dc15957..fc0c72c 100644 --- a/drivers/staging/lustre/lustre/llite/llite_internal.h +++ b/drivers/staging/lustre/lustre/llite/llite_internal.h @@ -651,8 +651,8 @@ void ll_rw_stats_tally(struct ll_sb_info *sbi, pid_t pid, void ll_release_page(struct page *page, int remove); extern const struct file_operations ll_dir_operations; extern const struct inode_operations ll_dir_inode_operations; -struct page *ll_get_dir_page(struct inode *dir, __u64 hash, -struct ll_dir_chain *chain); +struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data, +__u64 hash, struct ll_dir_chain *chain); int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, struct dir_context *ctx); diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c index 54ed52e..1b222c7 100644 --- a/drivers/staging/lustre/lustre/llite/statahead.c +++ b/drivers/staging/lustre/lustre/llite/statahead.c @@ -1067,7 +1067,7 @@ static int ll_statahead_thread(void *arg) wake_up(>t_ctl_waitq); ll_dir_chain_init(); - page = ll_get_dir_page(dir, pos, ); + page = ll_get_dir_page(dir, op_data, pos, ); while (1) { struct lu_dirpage *dp; @@ -1232,7 +1232,7 @@ do_it: ll_release_page(page, le32_to_cpu(dp->ldp_flags) & LDF_COLLIDE); sai->sai_in_readpage = 1; - page = ll_get_dir_page(dir, pos, ); + page = ll_get_dir_page(dir, op_data, pos, ); sai->sai_in_readpage = 0; } } @@ -1344,13 +1344,19 @@ static int is_first_dirent(struct inode *dir, struct dentry *dentry) { struct ll_dir_chain chain; const struct qstr *target = >d_name; + struct md_op_data *op_data; struct page *page; __u64pos= 0; intdot_de; intrc = LS_NONE_FIRST_DE; + op_data = ll_prep_md_op_data(NULL, dir, dir, NULL, 0, 0, +LUSTRE_OPC_ANY, dir); + if (IS_ERR(op_data)) + return PTR_ERR(op_data); + ll_dir_chain_init(); - page = ll_get_dir_page(dir, pos, ); + page = ll_get_dir_page(dir, op_data, pos, ); while (1) { struct lu_dirpage *dp; @@ -1438,12 +1444,13 @@ static int is_first_dirent(struct inode *dir, struct dentry
[PATCH 16/80] staging: lustre: obdclass: bug fixes for lu_device_type handling
From: Fan YongThere was no protection when inc/dec lu_device_type::ldt_device_nr, which may caused the ldt_device_nr to be wrong and trigger assert. This patch redefine lu_device_type::ldt_device_nr as atomic type. There was no protection when add/del lu_device_type::ldt_linkage into/from the global lu_device_types list, which may caused bad address accessing. This patch uses the existing obd_types_lock to protect related operations. We do NOT need lu_types_stop() any longer. Such function scans the global lu_device_types list, and for each type item on it which has zerod lu_device_type::ldt_device_nr, call its stop() method. In fact, the lu_device_type::ldt_device_nr only will be zero when the last lu_device_fini() is called, and at that time, inside the lu_device_fini(), its stop() method will be called. So it is unnecessary to call the stop() again via lu_types_stop(). Signed-off-by: Fan Yong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4604 Reviewed-on: http://review.whamcloud.com/8694 Reviewed-by: Jian Yu Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/include/lu_object.h |3 +- drivers/staging/lustre/lustre/llite/vvp_dev.c |6 --- drivers/staging/lustre/lustre/obdclass/lu_object.c | 34 ++-- drivers/staging/lustre/lustre/obdclass/obd_mount.c |1 - 4 files changed, 18 insertions(+), 26 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lu_object.h b/drivers/staging/lustre/lustre/include/lu_object.h index 6e25c1b..25c12d8 100644 --- a/drivers/staging/lustre/lustre/include/lu_object.h +++ b/drivers/staging/lustre/lustre/include/lu_object.h @@ -327,7 +327,7 @@ struct lu_device_type { /** * Number of existing device type instances. */ - unsignedldt_device_nr; + atomic_tldt_device_nr; /** * Linkage into a global list of all device types. * @@ -673,7 +673,6 @@ void lu_object_add(struct lu_object *before, struct lu_object *o); int lu_device_type_init(struct lu_device_type *ldt); void lu_device_type_fini(struct lu_device_type *ldt); -void lu_types_stop(void); /** @} ctors */ diff --git a/drivers/staging/lustre/lustre/llite/vvp_dev.c b/drivers/staging/lustre/lustre/llite/vvp_dev.c index e623216..771c0bd 100644 --- a/drivers/staging/lustre/lustre/llite/vvp_dev.c +++ b/drivers/staging/lustre/lustre/llite/vvp_dev.c @@ -368,12 +368,6 @@ int cl_sb_fini(struct super_block *sb) CERROR("Cannot cleanup cl-stack due to memory shortage.\n"); result = PTR_ERR(env); } - /* -* If mount failed (sbi->ll_cl == NULL), and this there are no other -* mounts, stop device types manually (this usually happens -* automatically when last device is destroyed). -*/ - lu_types_stop(); return result; } diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c index 9b03059..0c00bf8 100644 --- a/drivers/staging/lustre/lustre/obdclass/lu_object.c +++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c @@ -726,34 +726,31 @@ int lu_device_type_init(struct lu_device_type *ldt) { int result = 0; + atomic_set(>ldt_device_nr, 0); INIT_LIST_HEAD(>ldt_linkage); if (ldt->ldt_ops->ldto_init) result = ldt->ldt_ops->ldto_init(ldt); - if (result == 0) + + if (!result) { + spin_lock(_types_lock); list_add(>ldt_linkage, _device_types); + spin_unlock(_types_lock); + } + return result; } EXPORT_SYMBOL(lu_device_type_init); void lu_device_type_fini(struct lu_device_type *ldt) { + spin_lock(_types_lock); list_del_init(>ldt_linkage); + spin_unlock(_types_lock); if (ldt->ldt_ops->ldto_fini) ldt->ldt_ops->ldto_fini(ldt); } EXPORT_SYMBOL(lu_device_type_fini); -void lu_types_stop(void) -{ - struct lu_device_type *ldt; - - list_for_each_entry(ldt, _device_types, ldt_linkage) { - if (ldt->ldt_device_nr == 0 && ldt->ldt_ops->ldto_stop) - ldt->ldt_ops->ldto_stop(ldt); - } -} -EXPORT_SYMBOL(lu_types_stop); - /** * Global list of all sites on this node */ @@ -1082,8 +1079,10 @@ EXPORT_SYMBOL(lu_device_put); */ int lu_device_init(struct lu_device *d, struct lu_device_type *t) { - if (t->ldt_device_nr++ == 0 && t->ldt_ops->ldto_start) + if (atomic_inc_return(>ldt_device_nr) == 1 && + t->ldt_ops->ldto_start) t->ldt_ops->ldto_start(t); + memset(d, 0, sizeof(*d)); atomic_set(>ld_ref, 0); d->ld_type = t; @@ -1098,9 +1097,8 @@
[PATCH 03/80] staging: lustre: llite: style cleanup for llite_internal.h
From: wang diGroup function prototypes together related to dir.c. Move ll_release_page to be with function declarations. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../staging/lustre/lustre/llite/llite_internal.h |8 1 files changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h index fc0c72c..1ced397 100644 --- a/drivers/staging/lustre/lustre/llite/llite_internal.h +++ b/drivers/staging/lustre/lustre/llite/llite_internal.h @@ -648,15 +648,15 @@ void ll_rw_stats_tally(struct ll_sb_info *sbi, pid_t pid, size_t count, int rw); /* llite/dir.c */ -void ll_release_page(struct page *page, int remove); extern const struct file_operations ll_dir_operations; extern const struct inode_operations ll_dir_inode_operations; -struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data, -__u64 hash, struct ll_dir_chain *chain); int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, struct dir_context *ctx); - int ll_get_mdt_idx(struct inode *inode); +struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data, +__u64 hash, struct ll_dir_chain *chain); +void ll_release_page(struct page *page, int remove); + /* llite/namei.c */ extern const struct inode_operations ll_special_inode_operations; -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 14/80] staging: lustre: delete striped directory
From: wang diAdd delete striped directory, it includes 1. enable sync log between MDTs, so slave objects will be delete by unlink log, which is similar as deleting ost object. 2. retrieve layout information of striped directory on MDT, then lock all of the slave objects before unlink. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-on: http://review.whamcloud.com/7445 Reviewed-by: Alex Zhuravlev Reviewed-by: John L. Hammond Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/include/lustre_dlm.h |1 + drivers/staging/lustre/lustre/include/lustre_fid.h |1 + drivers/staging/lustre/lustre/llite/dir.c | 10 --- drivers/staging/lustre/lustre/lmv/lmv_intent.c |5 ++ drivers/staging/lustre/lustre/lmv/lmv_obd.c| 75 +++- 5 files changed, 65 insertions(+), 27 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre_dlm.h b/drivers/staging/lustre/lustre/include/lustre_dlm.h index 60051a5..f7805cc 100644 --- a/drivers/staging/lustre/lustre/include/lustre_dlm.h +++ b/drivers/staging/lustre/lustre/include/lustre_dlm.h @@ -968,6 +968,7 @@ struct ldlm_enqueue_info { void *ei_cb_cp; /** lock completion callback */ void *ei_cb_gl; /** lock glimpse callback */ void *ei_cbdata; /** Data to be passed into callbacks. */ + unsigned int ei_enq_slave:1; /* whether enqueue slave stripes */ }; extern struct obd_ops ldlm_obd_ops; diff --git a/drivers/staging/lustre/lustre/include/lustre_fid.h b/drivers/staging/lustre/lustre/include/lustre_fid.h index 743671a..61f3930 100644 --- a/drivers/staging/lustre/lustre/include/lustre_fid.h +++ b/drivers/staging/lustre/lustre/include/lustre_fid.h @@ -229,6 +229,7 @@ enum local_oid { MDD_LOV_OBJ_OSEQ= 4121UL, LFSCK_NAMESPACE_OID = 4122UL, REMOTE_PARENT_DIR_OID = 4123UL, + SLAVE_LLOG_CATALOGS_OID = 4124UL, }; static inline void lu_local_obj_fid(struct lu_fid *fid, __u32 oid) diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index 5288750..96ae7d5 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -679,16 +679,6 @@ static int ll_dir_setdirstripe(struct inode *dir, struct lmv_user_md *lump, if (unlikely(lump->lum_magic != LMV_USER_MAGIC)) return -EINVAL; - if (lump->lum_stripe_offset == (__u32)-1) { - int mdtidx; - - mdtidx = ll_get_mdt_idx(dir); - if (mdtidx < 0) - return mdtidx; - - lump->lum_stripe_offset = mdtidx; - } - CDEBUG(D_VFSTRACE, "VFS Op:inode="DFID"(%p) name %s stripe_offset %d, stripe_count: %u\n", PFID(ll_inode2fid(dir)), dir, filename, (int)lump->lum_stripe_offset, lump->lum_stripe_count); diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c index 1b9bbb2..5313dfc 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c @@ -244,6 +244,11 @@ update: if (req) ptlrpc_req_finished(req); + if (it.it_lock_mode && lockh) { + ldlm_lock_decref(lockh, it.it_lock_mode); + it.it_lock_mode = 0; + } + rc = -EIO; goto cleanup; } diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index da4855d..81dcc0a 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -101,6 +101,9 @@ int lmv_name_to_stripe_index(enum lmv_hash_type hashtype, return -EINVAL; } + CDEBUG(D_INFO, "name %.*s hash_type %d idx %d\n", namelen, name, + hashtype, idx); + LASSERT(idx < max_mdt_index); return idx; } @@ -1230,7 +1233,16 @@ static int lmv_placement_policy(struct obd_device *obd, struct lmv_user_md *lum; lum = op_data->op_data; - *mds = lum->lum_stripe_offset; + if (lum->lum_stripe_offset != (__u32)-1) { + *mds = lum->lum_stripe_offset; + } else { + /* +* -1 means default, which will be in the same MDT with +* the stripe +*/ + *mds = op_data->op_mds; + lum->lum_stripe_offset =
[PATCH 28/80] staging: lustre: lmv: access lum_stripe_offset as little endian
From: John L. HammondBy the time that a struct lmv_user_md reaches lmv_placement_policy() it has already been converted to little endian. Therefore use the appropriate macros around accesses to this this field. This issue was found by rewriting the definition of struct lmv_user_md to use the __leXX typedefs and running sparse. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4738 Reviewed-on: http://review.whamcloud.com/9671 Reviewed-by: Dmitry Eremin Reviewed-by: Swapnil Pimpale Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_obd.c |6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index c005a66..5929994 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -1242,15 +1242,15 @@ static int lmv_placement_policy(struct obd_device *obd, struct lmv_user_md *lum; lum = op_data->op_data; - if (lum->lum_stripe_offset != (__u32)-1) { - *mds = lum->lum_stripe_offset; + if (le32_to_cpu(lum->lum_stripe_offset) != (__u32)-1) { + *mds = le32_to_cpu(lum->lum_stripe_offset); } else { /* * -1 means default, which will be in the same MDT with * the stripe */ *mds = op_data->op_mds; - lum->lum_stripe_offset = op_data->op_mds; + lum->lum_stripe_offset = cpu_to_le32(op_data->op_mds); } } else { /* -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 11/80] staging: lustre: lmv: replace obd_free_memmd with lmv_free_memmd
From: wang diUse lmv_free_memmd for proper cleanup instead of the generic obd_free_memmd. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_obd.c |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index 0b1260d..6be2afc 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -2692,7 +2692,7 @@ static int lmv_free_lustre_md(struct obd_export *exp, struct lustre_md *md) struct lmv_tgt_desc *tgt = lmv->tgts[0]; if (md->lmv) - obd_free_memmd(exp, (void *)>lmv); + lmv_free_memmd(md->lmv); if (!tgt || !tgt->ltd_exp) return -EINVAL; return md_free_lustre_md(tgt->ltd_exp, md); -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 07/80] staging: lustre: lmv: remove unused lmv_get_mea function
From: wang diThe function lmv_get_mea() is not used so remove it. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_internal.h | 24 -- 1 files changed, 0 insertions(+), 24 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_internal.h b/drivers/staging/lustre/lustre/lmv/lmv_internal.h index 471470b..ab01560 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_internal.h +++ b/drivers/staging/lustre/lustre/lmv/lmv_internal.h @@ -55,30 +55,6 @@ int __lmv_fid_alloc(struct lmv_obd *lmv, struct lu_fid *fid, u32 mds); int lmv_fid_alloc(struct obd_export *exp, struct lu_fid *fid, struct md_op_data *op_data); -static inline struct lmv_stripe_md *lmv_get_mea(struct ptlrpc_request *req) -{ - struct mdt_body *body; - struct lmv_stripe_md*mea; - - LASSERT(req); - - body = req_capsule_server_get(>rq_pill, _MDT_BODY); - - if (!body || !S_ISDIR(body->mode) || !body->eadatasize) - return NULL; - - mea = req_capsule_server_sized_get(>rq_pill, _MDT_MD, - body->eadatasize); - if (mea->mea_count == 0) - return NULL; - if (mea->mea_magic != MEA_MAGIC_LAST_CHAR && - mea->mea_magic != MEA_MAGIC_ALL_CHARS && - mea->mea_magic != MEA_MAGIC_HASH_SEGMENT) - return NULL; - - return mea; -} - static inline int lmv_get_easize(struct lmv_obd *lmv) { return sizeof(struct lmv_stripe_md) + -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 15/80] staging: lustre: obdclass: fix lmd_parse() to handle comma-separated NIDs
From: Jian YuThis patch handles the upgrade situation that old mountdata already contains comma-separated NIDs. The correct way to fix the original issue is to parse comma-separated NIDs in lmd_parse(). Signed-off-by: Jian Yu Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4460 Reviewed-on: http://review.whamcloud.com/8918 Reviewed-by: Niu Yawei Reviewed-by: Bobi Jam Reviewed-by: Sebastien Buisson Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/obdclass/obd_mount.c | 20 1 files changed, 16 insertions(+), 4 deletions(-) diff --git a/drivers/staging/lustre/lustre/obdclass/obd_mount.c b/drivers/staging/lustre/lustre/obdclass/obd_mount.c index aa84a50..4931e37 100644 --- a/drivers/staging/lustre/lustre/obdclass/obd_mount.c +++ b/drivers/staging/lustre/lustre/obdclass/obd_mount.c @@ -880,7 +880,7 @@ static int lmd_parse_mgs(struct lustre_mount_data *lmd, char **ptr) */ static int lmd_parse(char *options, struct lustre_mount_data *lmd) { - char *s1, *s2, *devname = NULL; + char *s1, *s2, *s3, *devname = NULL; struct lustre_mount_data *raw = (struct lustre_mount_data *)options; int rc = 0; @@ -913,6 +913,7 @@ static int lmd_parse(char *options, struct lustre_mount_data *lmd) /* Skip whitespace and extra commas */ while (*s1 == ' ' || *s1 == ',') s1++; + s3 = s1; /* Client options are parsed in ll_options: eg. flock, * user_xattr, acl @@ -970,6 +971,7 @@ static int lmd_parse(char *options, struct lustre_mount_data *lmd) rc = lmd_parse_mgssec(lmd, s1 + 7); if (rc) goto invalid; + s3 = s2; clear++; /* ost exclusion list */ } else if (strncmp(s1, "exclude=", 8) == 0) { @@ -990,10 +992,19 @@ static int lmd_parse(char *options, struct lustre_mount_data *lmd) size_t length, params_length; char *tail = strchr(s1 + 6, ','); - if (!tail) + if (!tail) { length = strlen(s1); - else - length = tail - s1; + } else { + lnet_nid_t nid; + char *param_str = tail + 1; + int supplementary = 1; + + while (!class_parse_nid_quiet(param_str, , + _str)) { + supplementary = 0; + } + length = param_str - s1 - supplementary; + } length -= 6; params_length = strlen(lmd->lmd_params); if (params_length + length + 1 >= LMD_PARAMS_MAXLEN) @@ -1001,6 +1012,7 @@ static int lmd_parse(char *options, struct lustre_mount_data *lmd) strncat(lmd->lmd_params, s1 + 6, length); lmd->lmd_params[params_length + length] = '\0'; strlcat(lmd->lmd_params, " ", LMD_PARAMS_MAXLEN); + s3 = s1 + 6 + length; clear++; } else if (strncmp(s1, "osd=", 4) == 0) { rc = lmd_parse_string(>lmd_osd_type, s1 + 4); -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 18/80] staging: lustre: lmv: fix issue found by Klocwork Insight tool
From: Dmitry Eremin'plock.cookie' might be used uninitialized in this function. sscanf format specification '%d' expects type 'int *' for 'd', but parameter 3 has a different type '__u32*' Signed-off-by: Dmitry Eremin Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4629 Reviewed-on: http://review.whamcloud.com/9390 Reviewed-by: John L. Hammond Reviewed-by: James Simmons Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_intent.c |6 -- drivers/staging/lustre/lustre/lmv/lmv_obd.c|2 +- 2 files changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c index 2bc1098..51b7048 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c @@ -137,8 +137,10 @@ static int lmv_intent_remote(struct obd_export *exp, void *lmm, it->it_remote_lock_mode = it->it_lock_mode; } - it->it_lock_handle = plock.cookie; - it->it_lock_mode = pmode; + if (pmode) { + it->it_lock_handle = plock.cookie; + it->it_lock_mode = pmode; + } out_free_op_data: kfree(op_data); diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index 09b2efe..c005a66 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -1419,7 +1419,7 @@ static int lmv_process_config(struct obd_device *obd, u32 len, void *buf) obd_str2uuid(_uuid, lustre_cfg_buf(lcfg, 1)); - if (sscanf(lustre_cfg_buf(lcfg, 2), "%d", ) != 1) { + if (sscanf(lustre_cfg_buf(lcfg, 2), "%u", ) != 1) { rc = -EINVAL; goto out; } -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 35/80] staging: lustre: obdclass: unified flow control interfaces
From: Fan YongUnify the flow control interfaces for MDC RPC and FLD RPC. We allow to adjust the maximum inflight RPCs count via /sys interface. Signed-off-by: Fan Yong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4687 Reviewed-on: http://review.whamcloud.com/9562 Reviewed-by: Niu Yawei Reviewed-by: Alex Zhuravlev Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/fld/fld_request.c| 55 + drivers/staging/lustre/lustre/include/lustre_mdc.h |5 - drivers/staging/lustre/lustre/include/obd.h| 14 +-- drivers/staging/lustre/lustre/include/obd_class.h |5 + drivers/staging/lustre/lustre/ldlm/ldlm_lib.c |4 +- drivers/staging/lustre/lustre/mdc/lproc_mdc.c | 17 +-- drivers/staging/lustre/lustre/mdc/mdc_internal.h |2 - drivers/staging/lustre/lustre/mdc/mdc_lib.c| 64 -- drivers/staging/lustre/lustre/mdc/mdc_locks.c | 10 +- drivers/staging/lustre/lustre/mdc/mdc_request.c|6 +- drivers/staging/lustre/lustre/obdclass/genops.c| 132 11 files changed, 161 insertions(+), 153 deletions(-) diff --git a/drivers/staging/lustre/lustre/fld/fld_request.c b/drivers/staging/lustre/lustre/fld/fld_request.c index e59d626..ed7962e 100644 --- a/drivers/staging/lustre/lustre/fld/fld_request.c +++ b/drivers/staging/lustre/lustre/fld/fld_request.c @@ -53,57 +53,6 @@ #include "../include/lustre_mdc.h" #include "fld_internal.h" -/* TODO: these 3 functions are copies of flow-control code from mdc_lib.c - * It should be common thing. The same about mdc RPC lock - */ -static int fld_req_avail(struct client_obd *cli, struct mdc_cache_waiter *mcw) -{ - int rc; - - spin_lock(>cl_loi_list_lock); - rc = list_empty(>mcw_entry); - spin_unlock(>cl_loi_list_lock); - return rc; -}; - -static void fld_enter_request(struct client_obd *cli) -{ - struct mdc_cache_waiter mcw; - struct l_wait_info lwi = { 0 }; - - spin_lock(>cl_loi_list_lock); - if (cli->cl_r_in_flight >= cli->cl_max_rpcs_in_flight) { - list_add_tail(_entry, >cl_cache_waiters); - init_waitqueue_head(_waitq); - spin_unlock(>cl_loi_list_lock); - l_wait_event(mcw.mcw_waitq, fld_req_avail(cli, ), ); - } else { - cli->cl_r_in_flight++; - spin_unlock(>cl_loi_list_lock); - } -} - -static void fld_exit_request(struct client_obd *cli) -{ - struct list_head *l, *tmp; - struct mdc_cache_waiter *mcw; - - spin_lock(>cl_loi_list_lock); - cli->cl_r_in_flight--; - list_for_each_safe(l, tmp, >cl_cache_waiters) { - if (cli->cl_r_in_flight >= cli->cl_max_rpcs_in_flight) { - /* No free request slots anymore */ - break; - } - - mcw = list_entry(l, struct mdc_cache_waiter, mcw_entry); - list_del_init(>mcw_entry); - cli->cl_r_in_flight++; - wake_up(>mcw_waitq); - } - spin_unlock(>cl_loi_list_lock); -} - static int fld_rrb_hash(struct lu_client_fld *fld, u64 seq) { LASSERT(fld->lcf_count > 0); @@ -439,9 +388,9 @@ int fld_client_rpc(struct obd_export *exp, req->rq_reply_portal = MDC_REPLY_PORTAL; ptlrpc_at_set_req_timeout(req); - fld_enter_request(>exp_obd->u.cli); + obd_get_request_slot(>exp_obd->u.cli); rc = ptlrpc_queue_wait(req); - fld_exit_request(>exp_obd->u.cli); + obd_put_request_slot(>exp_obd->u.cli); if (rc) goto out_req; diff --git a/drivers/staging/lustre/lustre/include/lustre_mdc.h b/drivers/staging/lustre/lustre/include/lustre_mdc.h index 0a8c639..bf6f87a 100644 --- a/drivers/staging/lustre/lustre/include/lustre_mdc.h +++ b/drivers/staging/lustre/lustre/include/lustre_mdc.h @@ -179,11 +179,6 @@ static inline void mdc_update_max_ea_from_body(struct obd_export *exp, } } -struct mdc_cache_waiter { - struct list_head mcw_entry; - wait_queue_head_tmcw_waitq; -}; - /* mdc/mdc_locks.c */ int it_open_error(int phase, struct lookup_intent *it); diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h index f5eeb05..cacd472 100644 --- a/drivers/staging/lustre/lustre/include/obd.h +++ b/drivers/staging/lustre/lustre/include/obd.h @@ -211,11 +211,12 @@ struct timeout_item { struct list_head ti_chain; }; -#define OSC_MAX_RIF_DEFAULT 8 -#define OSC_MAX_RIF_MAX 256 -#define OSC_MAX_DIRTY_DEFAULT (OSC_MAX_RIF_DEFAULT * 4) -#define OSC_MAX_DIRTY_MB_MAX 2048 /* arbitrary, but < MAX_LONG bytes */ -#define OSC_DEFAULT_RESENDS 10 +#define OBD_MAX_RIF_DEFAULT8 +#define
[PATCH 25/80] staging: lustre: const correct FID/OSTID/... helpers
From: John L. HammondAdd a const qualifier wherever possible to the pointer parameters of the inline helper functions in lustre_idl.h and lustre_fid.h. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675 Reviewed-on: http://review.whamcloud.com/8641 Reviewed-by: wangdi Reviewed-by: Fan Yong Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h | 28 ++-- drivers/staging/lustre/lustre/include/lustre_fid.h |9 +++--- 2 files changed, 19 insertions(+), 18 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 87e79b9..c932e20 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -1033,7 +1033,7 @@ static inline int lu_dirent_calc_size(int namelen, __u16 attr) return (size + 7) & ~7; } -static inline int lu_dirent_size(struct lu_dirent *ent) +static inline int lu_dirent_size(const struct lu_dirent *ent) { if (le16_to_cpu(ent->lde_reclen) == 0) { return lu_dirent_calc_size(le16_to_cpu(ent->lde_namelen), @@ -1067,7 +1067,7 @@ struct lustre_handle { #define DEAD_HANDLE_MAGIC 0xdeadbeefcafebabeULL -static inline int lustre_handle_is_used(struct lustre_handle *lh) +static inline int lustre_handle_is_used(const struct lustre_handle *lh) { return lh->cookie != 0ull; } @@ -1079,7 +1079,7 @@ static inline int lustre_handle_equal(const struct lustre_handle *lh1, } static inline void lustre_handle_copy(struct lustre_handle *tgt, - struct lustre_handle *src) + const struct lustre_handle *src) { tgt->cookie = src->cookie; } @@ -1570,25 +1570,25 @@ static inline void lmm_oi_set_id(struct ost_id *oi, __u64 oid) oi->oi.oi_id = oid; } -static inline __u64 lmm_oi_id(struct ost_id *oi) +static inline __u64 lmm_oi_id(const struct ost_id *oi) { return oi->oi.oi_id; } -static inline __u64 lmm_oi_seq(struct ost_id *oi) +static inline __u64 lmm_oi_seq(const struct ost_id *oi) { return oi->oi.oi_seq; } static inline void lmm_oi_le_to_cpu(struct ost_id *dst_oi, - struct ost_id *src_oi) + const struct ost_id *src_oi) { dst_oi->oi.oi_id = le64_to_cpu(src_oi->oi.oi_id); dst_oi->oi.oi_seq = le64_to_cpu(src_oi->oi.oi_seq); } static inline void lmm_oi_cpu_to_le(struct ost_id *dst_oi, - struct ost_id *src_oi) + const struct ost_id *src_oi) { dst_oi->oi.oi_id = cpu_to_le64(src_oi->oi.oi_id); dst_oi->oi.oi_seq = cpu_to_le64(src_oi->oi.oi_seq); @@ -2724,15 +2724,15 @@ struct ldlm_extent { #define LDLM_GID_ANY ((__u64)-1) -static inline int ldlm_extent_overlap(struct ldlm_extent *ex1, - struct ldlm_extent *ex2) +static inline int ldlm_extent_overlap(const struct ldlm_extent *ex1, + const struct ldlm_extent *ex2) { return (ex1->start <= ex2->end) && (ex2->start <= ex1->end); } /* check if @ex1 contains @ex2 */ -static inline int ldlm_extent_contain(struct ldlm_extent *ex1, - struct ldlm_extent *ex2) +static inline int ldlm_extent_contain(const struct ldlm_extent *ex1, + const struct ldlm_extent *ex2) { return (ex1->start <= ex2->start) && (ex1->end >= ex2->end); } @@ -3092,7 +3092,7 @@ enum agent_req_status { ARS_SUCCEED, }; -static inline char *agent_req_status2name(enum agent_req_status ars) +static inline const char *agent_req_status2name(const enum agent_req_status ars) { switch (ars) { case ARS_WAITING: @@ -3268,7 +3268,7 @@ struct obdo { #define o_cksum o_nlink #define o_grant_used o_data_version -static inline void lustre_set_wire_obdo(struct obd_connect_data *ocd, +static inline void lustre_set_wire_obdo(const struct obd_connect_data *ocd, struct obdo *wobdo, const struct obdo *lobdo) { @@ -3287,7 +3287,7 @@ static inline void lustre_set_wire_obdo(struct obd_connect_data *ocd, } } -static inline void lustre_get_wire_obdo(struct obd_connect_data *ocd, +static inline void lustre_get_wire_obdo(const struct obd_connect_data *ocd, struct obdo *lobdo, const struct obdo *wobdo) { diff --git a/drivers/staging/lustre/lustre/include/lustre_fid.h b/drivers/staging/lustre/lustre/include/lustre_fid.h
[PATCH 40/80] staging: lustre: ptlrpc: Early replies need to honor at_max
From: Chris HornWhen determining whether an early reply can be sent the server will calculate the new deadline based on an offset from the request arrival time. However, when actually setting the new deadline the server offsets the current time. This can result in deadlines being extended more than at_max seconds past the request arrival time. Instead, the server should offset the arrival time when updating its request timeout. When a client receives an early reply it doesn't know the server side arrival time so we use the original sent time as an approximation. Signed-off-by: Chris Horn Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4578 Reviewed-on: http://review.whamcloud.com/9100 Reviewed-by: James Simmons Reviewed-by: Christopher J. Morrone Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/ptlrpc/client.c |8 +--- drivers/staging/lustre/lustre/ptlrpc/import.c | 11 +++ drivers/staging/lustre/lustre/ptlrpc/service.c | 18 +- 3 files changed, 25 insertions(+), 12 deletions(-) diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c index 549c62c..f2e71b4 100644 --- a/drivers/staging/lustre/lustre/ptlrpc/client.c +++ b/drivers/staging/lustre/lustre/ptlrpc/client.c @@ -385,10 +385,12 @@ static int ptlrpc_at_recv_early_reply(struct ptlrpc_request *req) spin_lock(>rq_lock); olddl = req->rq_deadline; /* -* server assumes it now has rq_timeout from when it sent the -* early reply, so client should give it at least that long. +* server assumes it now has rq_timeout from when the request +* arrived, so the client should give it at least that long. +* since we don't know the arrival time we'll use the original +* sent time */ - req->rq_deadline = ktime_get_real_seconds() + req->rq_timeout + + req->rq_deadline = req->rq_sent + req->rq_timeout + ptlrpc_at_get_net_latency(req); DEBUG_REQ(D_ADAPTTO, req, diff --git a/drivers/staging/lustre/lustre/ptlrpc/import.c b/drivers/staging/lustre/lustre/ptlrpc/import.c index 3292e6e..af8ffbc 100644 --- a/drivers/staging/lustre/lustre/ptlrpc/import.c +++ b/drivers/staging/lustre/lustre/ptlrpc/import.c @@ -1497,10 +1497,13 @@ EXPORT_SYMBOL(ptlrpc_disconnect_import); /* Adaptive Timeout utils */ extern unsigned int at_min, at_max, at_history; -/* Bin into timeslices using AT_BINS bins. - * This gives us a max of the last binlimit*AT_BINS secs without the storage, - * but still smoothing out a return to normalcy from a slow response. - * (E.g. remember the maximum latency in each minute of the last 4 minutes.) +/* + *Update at_current with the specified value (bounded by at_min and at_max), + * as well as the AT history "bins". + * - Bin into timeslices using AT_BINS bins. + * - This gives us a max of the last at_history seconds without the storage, + *but still smoothing out a return to normalcy from a slow response. + * - (E.g. remember the maximum latency in each minute of the last 4 minutes.) */ int at_measured(struct adaptive_timeout *at, unsigned int val) { diff --git a/drivers/staging/lustre/lustre/ptlrpc/service.c b/drivers/staging/lustre/lustre/ptlrpc/service.c index 4788c49..30d8b72 100644 --- a/drivers/staging/lustre/lustre/ptlrpc/service.c +++ b/drivers/staging/lustre/lustre/ptlrpc/service.c @@ -1005,13 +1005,16 @@ ptlrpc_at_remove_timed(struct ptlrpc_request *req) array->paa_count--; } +/* + * Attempt to extend the request deadline by sending an early reply to the + * client. + */ static int ptlrpc_at_send_early_reply(struct ptlrpc_request *req) { struct ptlrpc_service_part *svcpt = req->rq_rqbd->rqbd_svcpt; struct ptlrpc_request *reqcopy; struct lustre_msg *reqmsg; long olddl = req->rq_deadline - ktime_get_real_seconds(); - time64_t newdl; int rc; /* deadline is when the client expects us to reply, margin is the @@ -1039,8 +1042,13 @@ static int ptlrpc_at_send_early_reply(struct ptlrpc_request *req) return -ENOSYS; } - /* Fake our processing time into the future to ask the clients -* for some extra amount of time + /* +* We want to extend the request deadline by at_extra seconds, +* so we set our service estimate to reflect how much time has +* passed since this request arrived plus an additional +* at_extra seconds. The client will calculate the new deadline +* based on this service estimate (plus some additional time to +* account for network latency). See ptlrpc_at_recv_early_reply */ at_measured(>scp_at_estimate, at_extra + ktime_get_real_seconds() -
[PATCH 33/80] staging: lustre: lmv: cleanup req in lmv_getattr_name()
From: John L. HammondIn lmv_getattr_name() don't return a freed request in the error path. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4826 Reviewed-on: http://review.whamcloud.com/9863 Reviewed-by: Nathaniel Clark Reviewed-by: Bob Glossman Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_obd.c |1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index 5929994..f1b8ae9 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -1920,6 +1920,7 @@ lmv_getattr_name(struct obd_export *exp, struct md_op_data *op_data, tgt = lmv_find_target(lmv, ); if (IS_ERR(tgt)) { ptlrpc_req_finished(*request); + *request = NULL; return PTR_ERR(tgt); } -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 37/80] staging: lustre: ldlm: flock completion fixes.
From: Vitaly FertmanMove checks for FAILED, DESTROYED flags under ldlm spinlock, destroy flock atomically with the check it is not destroyed yet. Do not put the granted flock into the resource if this is UNLOCK, TEST, or DEADLOCK'ed flock. Later a regression for this patch was reported under LU-7626. The refcount nonzero (1) after lock cleanup errors was reported. The reason is that the case LCK_NL was not handled for obdecho. Patch 17791 resolved this issue which has been combined into this upstream patch. Signed-off-by: Vitaly Fertman Signed-off-by: Andriy Skulysh Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2177 Reviewed-by: Alexey Lyashkov Reviewed-by: Andriy Skulysh Reviewed-by: Vitaly Fertman Xyratex-bug-id: MRP-1588 Reviewed-on: http://review.whamcloud.com/10005 Reviewed-by: Bobi Jam Reviewed-by: Oleg Drokin Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7626 Reviewed-by: Mirza Arshad Mirza Hussain Reviewed-by: Alexey Leonidovich Lyashkov Reviewed-on: http://review.whamcloud.com/17791 Reviewed-by: Niu Yawei Reviewed-by: Alex Zhuravlev Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/include/linux/libcfs/libcfs_fail.h |3 + drivers/staging/lustre/lnet/libcfs/fail.c |6 +- .../lustre/lustre/include/lustre_dlm_flags.h | 36 --- .../staging/lustre/lustre/include/obd_support.h|4 + drivers/staging/lustre/lustre/ldlm/ldlm_flock.c| 98 ++-- drivers/staging/lustre/lustre/ldlm/ldlm_lock.c | 23 - drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 16 ++-- drivers/staging/lustre/lustre/ldlm/ldlm_resource.c |8 ++- drivers/staging/lustre/lustre/llite/file.c | 20 +++-- 9 files changed, 148 insertions(+), 66 deletions(-) diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs_fail.h b/drivers/staging/lustre/include/linux/libcfs/libcfs_fail.h index d3f9a60..bdbbe93 100644 --- a/drivers/staging/lustre/include/linux/libcfs/libcfs_fail.h +++ b/drivers/staging/lustre/include/linux/libcfs/libcfs_fail.h @@ -143,6 +143,9 @@ static inline int cfs_fail_timeout_set(__u32 id, __u32 value, int ms, int set) #define CFS_FAIL_TIMEOUT_ORSET(id, value, secs) \ cfs_fail_timeout_set(id, value, secs * 1000, CFS_FAIL_LOC_ORSET) +#define CFS_FAIL_TIMEOUT_RESET(id, value, secs) \ + cfs_fail_timeout_set(id, value, secs * 1000, CFS_FAIL_LOC_RESET) + #define CFS_FAIL_TIMEOUT_MS_ORSET(id, value, ms) \ cfs_fail_timeout_set(id, value, ms, CFS_FAIL_LOC_ORSET) diff --git a/drivers/staging/lustre/lnet/libcfs/fail.c b/drivers/staging/lustre/lnet/libcfs/fail.c index 9288ee0..e4b1a0a 100644 --- a/drivers/staging/lustre/lnet/libcfs/fail.c +++ b/drivers/staging/lustre/lnet/libcfs/fail.c @@ -90,8 +90,10 @@ int __cfs_fail_check_set(__u32 id, __u32 value, int set) } } - if ((set == CFS_FAIL_LOC_ORSET || set == CFS_FAIL_LOC_RESET) && - (value & CFS_FAIL_ONCE)) + /* Take into account the current call for FAIL_ONCE for ORSET only, +* as RESET is a new fail_loc, it does not change the current call +*/ + if ((set == CFS_FAIL_LOC_ORSET) && (value & CFS_FAIL_ONCE)) set_bit(CFS_FAIL_ONCE_BIT, _fail_loc); /* Lost race to set CFS_FAILED_BIT. */ if (test_and_set_bit(CFS_FAILED_BIT, _fail_loc)) { diff --git a/drivers/staging/lustre/lustre/include/lustre_dlm_flags.h b/drivers/staging/lustre/lustre/include/lustre_dlm_flags.h index e7e0c21..a0f064d 100644 --- a/drivers/staging/lustre/lustre/include/lustre_dlm_flags.h +++ b/drivers/staging/lustre/lustre/include/lustre_dlm_flags.h @@ -28,21 +28,6 @@ /** l_flags bits marked as "all_flags" bits */ #define LDLM_FL_ALL_FLAGS_MASK 0x00FFC08F932FULL -/** l_flags bits marked as "ast" bits */ -#define LDLM_FL_AST_MASK0x80008000ULL - -/** l_flags bits marked as "blocked" bits */ -#define LDLM_FL_BLOCKED_MASK0x000EULL - -/** l_flags bits marked as "gone" bits */ -#define LDLM_FL_GONE_MASK 0x00060040ULL - -/** l_flags bits marked as "inherit" bits */ -#define LDLM_FL_INHERIT_MASK0x0080ULL - -/** l_flags bits marked as "off_wire" bits */ -#define LDLM_FL_OFF_WIRE_MASK 0x00FFULL - /** extent, mode, or resource changed */ #define LDLM_FL_LOCK_CHANGED0x0001ULL /* bit 0 */ #define ldlm_is_lock_changed(_l)LDLM_TEST_FLAG((_l), 1ULL << 0) @@ -372,6 +357,27 @@ #define ldlm_set_excl(_l) LDLM_SET_FLAG((_l), 1ULL << 55)
[PATCH 47/80] staging: lustre: uapi: reduce scope of lustre_idl.h
From: John L. HammondMove the definition of OBD_OCD_VERSION() and similar macros from lustre_idl.h to lustre_ver.h. These macros are primarily used in comparisons to LUSTRE_VERSION_CODE which is defined in lustre_ver.h and so should be defined there as well. Move a few definitions (related to FIDs, quota and striping) from lustre_idl.h to lustre_user.h. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5065 Reviewed-on: http://review.whamcloud.com/10336 Reviewed-by: Andreas Dilger Reviewed-by: Bob Glossman Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h | 38 +-- .../lustre/lustre/include/lustre/lustre_user.h | 32 +++-- drivers/staging/lustre/lustre/include/lustre_ver.h | 13 +-- 3 files changed, 41 insertions(+), 42 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 8736826..69bed64 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -93,6 +93,7 @@ /* Defn's shared with user-space. */ #include "lustre_user.h" #include "lustre_errno.h" +#include "../lustre_ver.h" /* * GENERAL STUFF @@ -846,11 +847,6 @@ static inline bool fid_is_sane(const struct lu_fid *fid) fid_seq_is_rsvd(fid_seq(fid))); } -static inline bool fid_is_zero(const struct lu_fid *fid) -{ - return fid_seq(fid) == 0 && fid_oid(fid) == 0; -} - void lustre_swab_lu_fid(struct lu_fid *fid); void lustre_swab_lu_seq_range(struct lu_seq_range *range); @@ -1318,14 +1314,6 @@ void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); #define CLIENT_CONNECT_MDT_REQD (OBD_CONNECT_IBITS | OBD_CONNECT_FID | \ OBD_CONNECT_FULL20) -#define OBD_OCD_VERSION(major, minor, patch, fix) (((major)<<24) + \ - ((minor)<<16) + \ - ((patch)<<8) + (fix)) -#define OBD_OCD_VERSION_MAJOR(version) ((int)((version)>>24)&255) -#define OBD_OCD_VERSION_MINOR(version) ((int)((version)>>16)&255) -#define OBD_OCD_VERSION_PATCH(version) ((int)((version)>>8)&255) -#define OBD_OCD_VERSION_FIX(version) ((int)(version)&255) - /* This structure is used for both request and reply. * * If we eventually have separate connect data for different types, which we @@ -1509,14 +1497,6 @@ enum obdo_flags { #define LOV_MAGIC_V1_DEF 0x0CD10BD0 #define LOV_MAGIC_V3_DEF 0x0CD30BD0 -#define LOV_PATTERN_RAID0 0x001 /* stripes are used round-robin */ -#define LOV_PATTERN_RAID1 0x002 /* stripes are mirrors of each other */ -#define LOV_PATTERN_FIRST 0x100 /* first stripe is not in round-robin */ -#define LOV_PATTERN_CMOBD 0x200 - -#define LOV_PATTERN_F_MASK 0x -#define LOV_PATTERN_F_RELEASED 0x8000 /* HSM released file */ - #define lov_pattern(pattern) (pattern & ~LOV_PATTERN_F_MASK) #define lov_pattern_flags(pattern) (pattern & LOV_PATTERN_F_MASK) @@ -1796,7 +1776,7 @@ void lustre_swab_obd_statfs(struct obd_statfs *os); * it to sync quickly */ -#define OBD_OBJECT_EOF 0xULL +#define OBD_OBJECT_EOF LUSTRE_EOF #define OST_MIN_PRECREATE 32 #define OST_MAX_PRECREATE 2 @@ -1892,12 +1872,6 @@ struct obd_quotactl { void lustre_swab_obd_quotactl(struct obd_quotactl *q); -#define Q_QUOTACHECK 0x800100 /* deprecated as of 2.4 */ -#define Q_INITQUOTA0x800101 /* deprecated as of 2.4 */ -#define Q_GETOINFO 0x800102 /* get obd quota info */ -#define Q_GETOQUOTA0x800103 /* get obd quotas */ -#define Q_FINVALIDATE 0x800104 /* deprecated as of 2.4 */ - #define Q_COPY(out, in, member) (out)->member = (in)->member #define QCTL_COPY(out, in) \ @@ -2533,19 +2507,11 @@ struct lmv_mds_md_v1 { * for example the object is being migrated. And the hash function * might be interpreted differently with different flags. */ -enum lmv_hash_type { - LMV_HASH_TYPE_ALL_CHARS = 1, - LMV_HASH_TYPE_FNV_1A_64 = 2, -}; - #define LMV_HASH_TYPE_MASK 0x #define LMV_HASH_FLAG_MIGRATION0x8000 #define LMV_HASH_FLAG_DEAD 0x4000 -#define LMV_HASH_NAME_ALL_CHARS"all_char" -#define LMV_HASH_NAME_FNV_1A_64"fnv_1a_64" - /** * The FNV-1a hash algorithm is as follows: * hash = FNV_offset_basis diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h index 59d45de..8398c4f 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h +++
[PATCH 52/80] staging: lustre: llite: use the correct mode for striped directory
From: wang diCreate striped directory with correct mode, which should be handling same as mkdir. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4929 Reviewed-on: http://review.whamcloud.com/10028 Reviewed-by: Andreas Dilger Reviewed-by: John L. Hammond Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/dir.c | 40 +++- 1 files changed, 27 insertions(+), 13 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index 3fed80d..a1b5143 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -694,28 +694,40 @@ static int ll_send_mgc_param(struct obd_export *mgc, char *string) return rc; } -static int ll_dir_setdirstripe(struct inode *dir, struct lmv_user_md *lump, - const char *filename) +/** + * Create striped directory with specified stripe(@lump) + * + * param[in] parentthe parent of the directory. + * param[in] lump the specified stripes. + * param[in] dirname the name of the directory. + * param[in] mode the specified mode of the directory. + * + * retval =0 if striped directory is being created successfully. + * <0 if the creation is failed. + */ +static int ll_dir_setdirstripe(struct inode *parent, struct lmv_user_md *lump, + const char *dirname, umode_t mode) { struct ptlrpc_request *request = NULL; struct md_op_data *op_data; - struct ll_sb_info *sbi = ll_i2sbi(dir); - int mode; + struct ll_sb_info *sbi = ll_i2sbi(parent); int err; if (unlikely(lump->lum_magic != LMV_USER_MAGIC)) return -EINVAL; CDEBUG(D_VFSTRACE, "VFS Op:inode="DFID"(%p) name %s stripe_offset %d, stripe_count: %u\n", - PFID(ll_inode2fid(dir)), dir, filename, + PFID(ll_inode2fid(parent)), parent, dirname, (int)lump->lum_stripe_offset, lump->lum_stripe_count); if (lump->lum_magic != cpu_to_le32(LMV_USER_MAGIC)) lustre_swab_lmv_user_md(lump); - mode = (~current_umask() & 0755) | S_IFDIR; - op_data = ll_prep_md_op_data(NULL, dir, NULL, filename, -strlen(filename), mode, LUSTRE_OPC_MKDIR, + if (!IS_POSIXACL(parent) || !exp_connect_umask(ll_i2mdexp(parent))) + mode &= ~current_umask(); + mode = (mode & (S_IRWXUGO | S_ISVTX)) | S_IFDIR; + op_data = ll_prep_md_op_data(NULL, parent, NULL, dirname, +strlen(dirname), mode, LUSTRE_OPC_MKDIR, lump); if (IS_ERR(op_data)) { err = PTR_ERR(op_data); @@ -1379,6 +1391,7 @@ out_free: char*filename; int namelen = 0; int lumlen = 0; + umode_t mode; int len; int rc; @@ -1412,11 +1425,12 @@ out_free: goto lmv_out_free; } - /** -* ll_dir_setdirstripe will be used to set dir stripe -* mdc_create--->mdt_reint_create (with dirstripe) -*/ - rc = ll_dir_setdirstripe(inode, lum, filename); +#if OBD_OCD_VERSION(2, 9, 50, 0) > LUSTRE_VERSION_CODE + mode = data->ioc_type != 0 ? data->ioc_type : S_IRWXUGO; +#else + mode = data->ioc_type; +#endif + rc = ll_dir_setdirstripe(inode, lum, filename, mode); lmv_out_free: obd_ioctl_freedata(buf, len); return rc; -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 50/80] staging: lustre: lov: new pattern flag for partially repaired file
From: Fan YongWhen the layout LFSCK repairs orphan OST-object, if the parent MDT-object was lost, then it will re-create the MDT-object and regenerate the LOV EA and fill the target LOV EA slot with the orphan information, and fill other slots with zero (LOV hole); if related LOV EA slot is invalid or hole, then it will refill the target LOV EA slot; if the target slot exceeds current LOV EA tail, then extend the LOV EA, and fill the gaps as zero. Some of the LOV EA holes may cannot be re-filled finally because of lost some OST-objects. And even if they can be re-filled, but there are still some possible race accessings from client before the re-filling. If the client access the LOV EA with hole(s), it may cause some strange behaviour, such as trigger LBUG()/LASSERT() on the client. So we will make the client to be aware of the LOV EA is incomplete. We introduce a new LOV EA pattern flag LOV_PATTERN_F_HOLE for that: any time when the LFSCK repairs the LOV EA with hole(s), the LOV EA will be marked as LOV_PATTERN_F_HOLE; when all the holes in the LOV EA are refilled, the LOV_PATTERN_F_HOLE will be dropped. For a new client, it recongizes the pattern flag LOV_PATTERN_F_HOLE, then it can permit/forbid some opertions on the file with LOV holes: 1) Normal read/write the file with LOV EA hole is permitted, but the application will get EIO error when read data from the dummy slot or write data to the dummy slot. 2) The users can dump the recovered data via some common read tools, such as "dd conv=sync,noerror". 3) Append data to the file which has LOV EA hole will get EIO failure. 4) Other operations will skip the LOV EA hole(s), and will not get failures, such as {s,g}etattr, {s,g}getxattr, stat, chown/chgrp, chmod, touch, unlink, and so on. For an old client, since it will not recognize the new pattern flag LOV_PATTERN_F_HOLE. So the LOV EA with hole will be dicarded with failure, but it will not cause the client to be crashed. Signed-off-by: Fan Yong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4675 Reviewed-on: http://review.whamcloud.com/10042 Reviewed-by: Andreas Dilger Reviewed-by: Alex Zhuravlev Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h |1 + .../lustre/lustre/include/lustre/lustre_user.h |2 +- drivers/staging/lustre/lustre/llite/llite_lib.c|2 +- drivers/staging/lustre/lustre/lov/lov_io.c | 16 +--- .../lustre/lustre/obdclass/lprocfs_status.c|2 ++ drivers/staging/lustre/lustre/ptlrpc/wiretest.c|2 ++ 6 files changed, 20 insertions(+), 5 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 69bed64..87eef4c 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -1289,6 +1289,7 @@ void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); #define OBD_CONNECT_OPEN_BY_FID0x20ULL /* open by fid won't pack * name in request */ +#define OBD_CONNECT_LFSCK 0x40ULL/* support online LFSCK */ /* XXX README XXX: * Please DO NOT add flag values here before first ensuring that this same diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h index 8398c4f..9e38ed3 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h @@ -284,9 +284,9 @@ struct ost_id { #define LOV_PATTERN_CMOBD 0x200 #define LOV_PATTERN_F_MASK 0x +#define LOV_PATTERN_F_HOLE 0x4000 /* there is hole in LOV EA */ #define LOV_PATTERN_F_RELEASED 0x8000 /* HSM released file */ - #define LOV_MAXPOOLNAME 16 #define LOV_POOLNAMEF "%.16s" diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c index ac59cd6..dd44ee8 100644 --- a/drivers/staging/lustre/lustre/llite/llite_lib.c +++ b/drivers/staging/lustre/lustre/llite/llite_lib.c @@ -189,7 +189,7 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt, OBD_CONNECT_PINGLESS | OBD_CONNECT_MAX_EASIZE | OBD_CONNECT_FLOCK_DEAD | - OBD_CONNECT_DISP_STRIPE; + OBD_CONNECT_DISP_STRIPE | OBD_CONNECT_LFSCK; if (sbi->ll_flags & LL_SBI_SOM_PREVIEW) data->ocd_connect_flags |= OBD_CONNECT_SOM; diff --git
[PATCH 42/80] staging: lustre: llite: validate names
From: John L. HammondIn ll_prep_md_op_data() validate names according to the same formula used in mdd_name_check(). Add mdc_pack_name() to validate the name actually packed in the request. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4992 Reviewed-on: http://review.whamcloud.com/10198 Reviewed-by: wangdi Reviewed-by: Andreas Dilger Signed-off-by: James Simmons --- .../lustre/include/linux/libcfs/libcfs_private.h |9 --- drivers/staging/lustre/lustre/include/lu_object.h | 16 + drivers/staging/lustre/lustre/llite/llite_lib.c| 13 +++- drivers/staging/lustre/lustre/mdc/mdc_lib.c| 67 +--- 4 files changed, 70 insertions(+), 35 deletions(-) diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h b/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h index 4daa382..d401ae1 100644 --- a/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h +++ b/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h @@ -360,13 +360,4 @@ do { \ ptr += cfs_size_round(len); \ } while (0) -#define LOGL0(var, len, ptr) \ -do { \ - if (!len) \ - break;\ - memcpy((char *)ptr, (const char *)var, len);\ - *((char *)(ptr) + len) = 0; \ - ptr += cfs_size_round(len + 1); \ -} while (0) - #endif diff --git a/drivers/staging/lustre/lustre/include/lu_object.h b/drivers/staging/lustre/lustre/include/lu_object.h index 25c12d8..6ab1782 100644 --- a/drivers/staging/lustre/lustre/include/lu_object.h +++ b/drivers/staging/lustre/lustre/include/lu_object.h @@ -1263,6 +1263,22 @@ struct lu_name { }; /** + * Validate names (path components) + * + * To be valid \a name must be non-empty, '\0' terminated of length \a + * name_len, and not contain '/'. The maximum length of a name (before + * say -ENAMETOOLONG will be returned) is really controlled by llite + * and the server. We only check for something insane coming from bad + * integer handling here. + */ +static inline bool lu_name_is_valid_2(const char *name, size_t name_len) +{ + return name && name_len > 0 && name_len < INT_MAX && + name[name_len] == '\0' && strlen(name) == name_len && + !memchr(name, '/', name_len); +} + +/** * Common buffer structure to be passed around for various xattr_{s,g}et() * methods. */ diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c index 2f6e770..a3b4c97 100644 --- a/drivers/staging/lustre/lustre/llite/llite_lib.c +++ b/drivers/staging/lustre/lustre/llite/llite_lib.c @@ -2304,8 +2304,17 @@ struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data, const char *name, int namelen, int mode, __u32 opc, void *data) { - if (namelen > ll_i2sbi(i1)->ll_namelen) - return ERR_PTR(-ENAMETOOLONG); + if (!name) { + /* Do not reuse namelen for something else. */ + if (namelen) + return ERR_PTR(-EINVAL); + } else { + if (namelen > ll_i2sbi(i1)->ll_namelen) + return ERR_PTR(-ENAMETOOLONG); + + if (!lu_name_is_valid_2(name, namelen)) + return ERR_PTR(-EINVAL); + } if (!op_data) op_data = kzalloc(sizeof(*op_data), GFP_NOFS); diff --git a/drivers/staging/lustre/lustre/mdc/mdc_lib.c b/drivers/staging/lustre/lustre/mdc/mdc_lib.c index b532623..16c3571 100644 --- a/drivers/staging/lustre/lustre/mdc/mdc_lib.c +++ b/drivers/staging/lustre/lustre/mdc/mdc_lib.c @@ -87,6 +87,37 @@ void mdc_pack_body(struct ptlrpc_request *req, const struct lu_fid *fid, } } +/** + * Pack a name (path component) into a request + * + * \param[in] req request + * \param[in] fieldrequest field (usually RMF_NAME) + * \param[in] name path component + * \param[in] name_len length of path component + * + * \a field must be present in \a req and of size \a name_len + 1. + * + * \a name must be '\0' terminated of length \a name_len and represent + * a single path component (not contain '/'). + */ +static void mdc_pack_name(struct ptlrpc_request *req, + const struct req_msg_field *field, + const char *name, size_t name_len) +{ + size_t buf_size; + size_t cpy_len; + char *buf; + + buf = req_capsule_client_get(>rq_pill, field); + buf_size = req_capsule_get_size(>rq_pill, field, RCL_CLIENT); + + LASSERT(name
[PATCH 54/80] staging: lustre: llite: set dir LOV xattr length variable
From: Hongchao Zhangthe LOV xattr of directory could be either lov_user_md_v1 (size is 32) or lov_user_md_v3 (size is 48), then the actual size of the LOV xattr should be return. Signed-off-by: Hongchao Zhang Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5100 Reviewed-on: http://review.whamcloud.com/10453 Reviewed-by: James Simmons Reviewed-by: Alex Zhuravlev Reviewed-by: jacques-Charles Lafoucriere Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/xattr.c |8 1 files changed, 0 insertions(+), 8 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/xattr.c b/drivers/staging/lustre/lustre/llite/xattr.c index aa0738b..146da6b 100644 --- a/drivers/staging/lustre/lustre/llite/xattr.c +++ b/drivers/staging/lustre/lustre/llite/xattr.c @@ -379,14 +379,6 @@ static int ll_xattr_get(const struct xattr_handler *handler, if (!S_ISREG(inode->i_mode) && !S_ISDIR(inode->i_mode)) return -ENODATA; - if (size == 0 && S_ISDIR(inode->i_mode)) { - /* XXX directory EA is fix for now, optimize to save -* RPC transfer -*/ - rc = sizeof(struct lov_user_md); - goto out; - } - lsm = ccc_inode_lsm_get(inode); if (!lsm) { if (S_ISDIR(inode->i_mode)) { -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 53/80] staging: lustre: obd: rename lsr_padding to lsr_valid
From: Niu YaweiSimple variable rename. Signed-off-by: Niu Yawei Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4345 Reviewed-on: http://review.whamcloud.com/10223 Reviewed-by: Andreas Dilger Reviewed-by: Mike Pershin Reviewed-by: Alex Zhuravlev Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h |2 +- drivers/staging/lustre/lustre/obdclass/llog_swab.c |1 + drivers/staging/lustre/lustre/ptlrpc/wiretest.c|8 3 files changed, 6 insertions(+), 5 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 87eef4c..bbf0c8d 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -3036,7 +3036,7 @@ struct llog_setattr64_rec { __u32 lsr_uid_h; __u32 lsr_gid; __u32 lsr_gid_h; - __u64 lsr_padding; + __u64 lsr_valid; struct llog_rec_taillsr_tail; } __packed; diff --git a/drivers/staging/lustre/lustre/obdclass/llog_swab.c b/drivers/staging/lustre/lustre/obdclass/llog_swab.c index f7b9b19..0ec6361 100644 --- a/drivers/staging/lustre/lustre/obdclass/llog_swab.c +++ b/drivers/staging/lustre/lustre/obdclass/llog_swab.c @@ -224,6 +224,7 @@ void lustre_swab_llog_rec(struct llog_rec_hdr *rec) __swab32s(>lsr_uid_h); __swab32s(>lsr_gid); __swab32s(>lsr_gid_h); + __swab64s(>lsr_valid); tail = >lsr_tail; break; } diff --git a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c index 9d5d2c8..8dbaf32 100644 --- a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c +++ b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c @@ -3170,10 +3170,10 @@ void lustre_assert_wire_constants(void) (long long)(int)offsetof(struct llog_setattr64_rec, lsr_gid_h)); LASSERTF((int)sizeof(((struct llog_setattr64_rec *)0)->lsr_gid_h) == 4, "found %lld\n", (long long)(int)sizeof(((struct llog_setattr64_rec *)0)->lsr_gid_h)); - LASSERTF((int)offsetof(struct llog_setattr64_rec, lsr_padding) == 48, "found %lld\n", -(long long)(int)offsetof(struct llog_setattr64_rec, lsr_padding)); - LASSERTF((int)sizeof(((struct llog_setattr64_rec *)0)->lsr_padding) == 8, "found %lld\n", -(long long)(int)sizeof(((struct llog_setattr64_rec *)0)->lsr_padding)); + LASSERTF((int)offsetof(struct llog_setattr64_rec, lsr_valid) == 48, "found %lld\n", +(long long)(int)offsetof(struct llog_setattr64_rec, lsr_valid)); + LASSERTF((int)sizeof(((struct llog_setattr64_rec *)0)->lsr_valid) == 8, "found %lld\n", +(long long)(int)sizeof(((struct llog_setattr64_rec *)0)->lsr_valid)); LASSERTF((int)offsetof(struct llog_setattr64_rec, lsr_tail) == 56, "found %lld\n", (long long)(int)offsetof(struct llog_setattr64_rec, lsr_tail)); LASSERTF((int)sizeof(((struct llog_setattr64_rec *)0)->lsr_tail) == 8, "found %lld\n", -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 43/80] staging: lustre: llite: fix inconsistencies of root squash feature
From: Gregoire PichonRoot squash exhibits inconsistent behaviour on a client when enabled. If a file is not cached on the client, then root will get a permission denied error when accessing the file. When the file has recently been accessed by a regular user and is still in cache, root will be able to access the file without error because the permission check is only done by the client that isn't aware of root squash. While the only real security benefit from root squash is to deny clients access to files owned by root itself, it also makes sense to treat file access on the client in a consistent manner regardless of whether the file is in cache or not. This patch adds root squash settings to llite so that client is able to apply root squashing when it is relevant. Configuration of MDT root squash settings will automatically be applied to llite config log as well. Update cfs_str2num_check() routine by removing any modification of the specified string parameter. Since string can come from ls_str field of a lstr structure, this avoids inconsistent ls_len field. Signed-off-by: Gregoire Pichon Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1778 Reviewed-on: http://review.whamcloud.com/5700 Reviewed-by: Andreas Dilger Reviewed-by: Niu Yawei Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lnet/libcfs/libcfs_string.c |2 - .../staging/lustre/lustre/include/lprocfs_status.h |6 + drivers/staging/lustre/lustre/include/obd_class.h |9 ++ drivers/staging/lustre/lustre/llite/file.c | 44 ++ .../staging/lustre/lustre/llite/llite_internal.h |6 + drivers/staging/lustre/lustre/llite/llite_lib.c| 47 +++ drivers/staging/lustre/lustre/llite/lproc_llite.c | 68 ++ .../lustre/lustre/obdclass/lprocfs_status.c| 140 8 files changed, 320 insertions(+), 2 deletions(-) diff --git a/drivers/staging/lustre/lnet/libcfs/libcfs_string.c b/drivers/staging/lustre/lnet/libcfs/libcfs_string.c index fc697cd..56a614d 100644 --- a/drivers/staging/lustre/lnet/libcfs/libcfs_string.c +++ b/drivers/staging/lustre/lnet/libcfs/libcfs_string.c @@ -229,8 +229,6 @@ cfs_str2num_check(char *str, int nob, unsigned *num, char *endp, cache; int rc; - str = cfs_trimwhite(str); - /** * kstrouint can only handle strings composed * of only numbers. We need to scan the string diff --git a/drivers/staging/lustre/lustre/include/lprocfs_status.h b/drivers/staging/lustre/lustre/include/lprocfs_status.h index d68e60e..ff35e63 100644 --- a/drivers/staging/lustre/lustre/include/lprocfs_status.h +++ b/drivers/staging/lustre/lustre/include/lprocfs_status.h @@ -681,6 +681,12 @@ static struct lustre_attr lustre_attr_##name = __ATTR(name, mode, show, store) extern const struct sysfs_ops lustre_sysfs_ops; +struct root_squash_info; +int lprocfs_wr_root_squash(const char *buffer, unsigned long count, + struct root_squash_info *squash, char *name); +int lprocfs_wr_nosquash_nids(const char *buffer, unsigned long count, +struct root_squash_info *squash, char *name); + /* all quota proc functions */ int lprocfs_quota_rd_bunit(char *page, char **start, loff_t off, int count, diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h index a288995..e86961c 100644 --- a/drivers/staging/lustre/lustre/include/obd_class.h +++ b/drivers/staging/lustre/lustre/include/obd_class.h @@ -1759,4 +1759,13 @@ extern spinlock_t obd_types_lock; /* prng.c */ #define ll_generate_random_uuid(uuid_out) cfs_get_random_bytes(uuid_out, sizeof(class_uuid_t)) +/* root squash info */ +struct rw_semaphore; +struct root_squash_info { + uid_t rsi_uid; + gid_t rsi_gid; + struct list_headrsi_nosquash_nids; + struct rw_semaphore rsi_sem; +}; + #endif /* __LINUX_OBD_CLASS_H */ diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c index 519db53..90a7170 100644 --- a/drivers/staging/lustre/lustre/llite/file.c +++ b/drivers/staging/lustre/lustre/llite/file.c @@ -41,6 +41,7 @@ #include "../include/lustre_lite.h" #include #include +#include #include #include "llite_internal.h" #include "../include/lustre/ll_fiemap.h" @@ -3289,6 +3290,12 @@ struct posix_acl *ll_get_acl(struct inode *inode, int type) int ll_inode_permission(struct inode *inode, int mask) { + struct ll_sb_info *sbi; + struct root_squash_info *squash; + const struct cred *old_cred = NULL; + struct cred *cred = NULL; + bool squash_id = false; + cfs_cap_t cap; int rc = 0; if (mask &
[PATCH 48/80] staging: lustre: llite: a few fixes about readdir of striped dir.
From: wang diNormally we know the value of op_mea1 when ll_readdir is called. In the case of '.' or '..' op_mea1 is unknown so for that case fetch the real parents FID. Signed-off-by: wang di Signed-off-by: Li Xi Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4603 Reviewed-on: http://review.whamcloud.com/9191 Reviewed-by: John L. Hammond Reviewed-by: Andreas Dilger Reviewed-by: Fan Yong Reviewed-by: Li Xi Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/dir.c | 27 + .../staging/lustre/lustre/llite/llite_internal.h |1 + drivers/staging/lustre/lustre/llite/llite_nfs.c| 31 ++-- 3 files changed, 50 insertions(+), 9 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index 924b5df..3fed80d 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -622,6 +622,33 @@ static int ll_readdir(struct file *filp, struct dir_context *ctx) goto out; } + if (unlikely(op_data->op_mea1)) { + /* +* This is only needed for striped dir to fill .., +* see lmv_read_page +*/ + if (file_dentry(filp)->d_parent && + file_dentry(filp)->d_parent->d_inode) { + __u64 ibits = MDS_INODELOCK_UPDATE; + struct inode *parent; + + parent = file_dentry(filp)->d_parent->d_inode; + if (ll_have_md_lock(parent, , LCK_MINMODE)) + op_data->op_fid3 = *ll_inode2fid(parent); + } + + /* +* If it can not find in cache, do lookup .. on the master +* object +*/ + if (fid_is_zero(_data->op_fid3)) { + rc = ll_dir_get_parent_fid(inode, _data->op_fid3); + if (rc) { + ll_finish_md_op_data(op_data); + return rc; + } + } + } ctx->pos = pos; rc = ll_dir_read(inode, , op_data, ctx); pos = ctx->pos; diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h index 3d7fa9a..43269aa 100644 --- a/drivers/staging/lustre/lustre/llite/llite_internal.h +++ b/drivers/staging/lustre/lustre/llite/llite_internal.h @@ -812,6 +812,7 @@ __u32 get_uuid2int(const char *name, int len); void get_uuid2fsid(const char *name, int len, __kernel_fsid_t *fsid); struct inode *search_inode_for_lustre(struct super_block *sb, const struct lu_fid *fid); +int ll_dir_get_parent_fid(struct inode *dir, struct lu_fid *parent_fid); /* llite/symlink.c */ extern const struct inode_operations ll_fast_symlink_inode_operations; diff --git a/drivers/staging/lustre/lustre/llite/llite_nfs.c b/drivers/staging/lustre/lustre/llite/llite_nfs.c index ab9d5cc..06a8199 100644 --- a/drivers/staging/lustre/lustre/llite/llite_nfs.c +++ b/drivers/staging/lustre/lustre/llite/llite_nfs.c @@ -302,14 +302,12 @@ static struct dentry *ll_fh_to_parent(struct super_block *sb, struct fid *fid, return ll_iget_for_nfs(sb, _fid->lnf_parent, NULL); } -static struct dentry *ll_get_parent(struct dentry *dchild) +int ll_dir_get_parent_fid(struct inode *dir, struct lu_fid *parent_fid) { struct ptlrpc_request *req = NULL; - struct inode *dir = d_inode(dchild); struct ll_sb_info *sbi; - struct dentry*result = NULL; struct mdt_body *body; - static chardotdot[] = ".."; + static const char dotdot[] = ".."; struct md_op_data *op_data; intrc; int lmmsize; @@ -324,13 +322,13 @@ static struct dentry *ll_get_parent(struct dentry *dchild) rc = ll_get_default_mdsize(sbi, ); if (rc != 0) - return ERR_PTR(rc); + return rc; op_data = ll_prep_md_op_data(NULL, dir, NULL, dotdot, strlen(dotdot), lmmsize, LUSTRE_OPC_ANY, NULL); if (IS_ERR(op_data)) - return (void *)op_data; + return PTR_ERR(op_data); rc = md_getattr_name(sbi->ll_md_exp, op_data, ); ll_finish_md_op_data(op_data); @@ -338,7 +336,7 @@ static struct dentry *ll_get_parent(struct dentry *dchild) CERROR("%s: failure inode "DFID" get parent: rc = %d\n", ll_get_fsname(dir->i_sb, NULL, 0),
[PATCH 45/80] staging: lustre: llite: Fix the deadlock in balance_dirty_pages()
From: Jinshan XiongIf the page is already dirtied in ll_write_end() and kernel tries to call balance_dirty_pages() to write back dirty pages in the same thread, this is deadlock case if the page is already held by clio. This can also fix the issue of LU-4873. Signed-off-by: Jinshan Xiong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4977 Reviewed-on: http://review.whamcloud.com/10149 Reviewed-by: Bobi Jam Reviewed-by: Lai Siyao Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/rw26.c | 12 +--- 1 files changed, 9 insertions(+), 3 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c index c14a1b6..8c8c100 100644 --- a/drivers/staging/lustre/lustre/llite/rw26.c +++ b/drivers/staging/lustre/lustre/llite/rw26.c @@ -506,9 +506,8 @@ static int ll_write_begin(struct file *file, struct address_space *mapping, env = lcc->lcc_env; io = lcc->lcc_io; - if (likely(to == PAGE_SIZE)) /* LU-4873 */ - /* To avoid deadlock, try to lock page first. */ - vmpage = grab_cache_page_nowait(mapping, index); + /* To avoid deadlock, try to lock page first. */ + vmpage = grab_cache_page_nowait(mapping, index); if (unlikely(!vmpage || PageDirty(vmpage) || PageWriteback(vmpage))) { struct vvp_io *vio = vvp_env_io(env); struct cl_page_list *plist = >u.write.vui_queue; @@ -617,6 +616,13 @@ static int ll_write_end(struct file *file, struct address_space *mapping, LASSERT(from == 0); vio->u.write.vui_to = from + copied; + /* +* To address the deadlock in balance_dirty_pages() where +* this dirty page may be written back in the same thread. +*/ + if (PageDirty(vmpage)) + unplug = true; + /* We may have one full RPC, commit it soon */ if (plist->pl_nr >= PTLRPC_MAX_BRW_PAGES) unplug = true; -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [RFCv2][PATCH 2/5] arm: Implement ARCH_HAS_FORCE_CACHE
On 08/10/2016 04:22 PM, Russell King - ARM Linux wrote: On Mon, Aug 08, 2016 at 10:49:34AM -0700, Laura Abbott wrote: +/* + * Make an area consistent for devices. + * Note: Drivers should NOT use this function directly, as it will break + * platforms with CONFIG_DMABOUNCE. + * Use the driver DMA support - see dma-mapping.h (dma_sync_*) + */ +void __dma_page_cpu_to_dev(struct page *page, unsigned long off, + size_t size, enum dma_data_direction dir) +{ + phys_addr_t paddr; + + dma_cache_maint_page(page, off, size, dir, dmac_map_area); + + paddr = page_to_phys(page) + off; + if (dir == DMA_FROM_DEVICE) { + outer_inv_range(paddr, paddr + size); + } else { + outer_clean_range(paddr, paddr + size); + } + /* FIXME: non-speculating: flush on bidirectional mappings? */ +} + +void __dma_page_dev_to_cpu(struct page *page, unsigned long off, + size_t size, enum dma_data_direction dir) +{ + phys_addr_t paddr = page_to_phys(page) + off; + + /* FIXME: non-speculating: not required */ + /* in any case, don't bother invalidating if DMA to device */ + if (dir != DMA_TO_DEVICE) { + outer_inv_range(paddr, paddr + size); + + dma_cache_maint_page(page, off, size, dir, dmac_unmap_area); + } + + /* +* Mark the D-cache clean for these pages to avoid extra flushing. +*/ + if (dir != DMA_TO_DEVICE && size >= PAGE_SIZE) { + unsigned long pfn; + size_t left = size; + + pfn = page_to_pfn(page) + off / PAGE_SIZE; + off %= PAGE_SIZE; + if (off) { + pfn++; + left -= PAGE_SIZE - off; + } + while (left >= PAGE_SIZE) { + page = pfn_to_page(pfn++); + set_bit(PG_dcache_clean, >flags); + left -= PAGE_SIZE; + } + } +} I _really_ don't want these exposed in any shape or form to driver code. I've seen too many hacks out there where people have gone under the cover of the APIs they should be using, and headed straight for the low-level functionality - adding function prototypes to get at stuff they have no business doing. Moving this here is just asking for it to be abused. + +void kernel_force_cache_clean(struct page *page, size_t size) +{ + __dma_page_cpu_to_dev(page, 0, size, DMA_BIDIRECTIONAL); +} + +void kernel_force_cache_invalidate(struct page *page, size_t size) +{ + __dma_page_dev_to_cpu(page, 0, size, DMA_BIDIRECTIONAL); +} Nothing in our implementation of these DMA operations guarantees that those mean "clean" and "invalidate". The DMA operations are there so that CPUs can implement whatever they need at the map and unmap times - and I've been very careful not to specify which cache operations are involved. For example, on older CPUs, __dma_page_dev_to_cpu() is almost always a no-op. If you want something that does something specific, then we need something designed to do something specific. Please don't re-use what you think will fit. I see what you are saying. What I really wanted was to re-use some of the code that dma_cache_maint_page was doing for highmem handling but it looks like I picked the wrong layer to make common. I'll give this some thought. Thanks, Laura ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
RE: [PATCH v2 RESEND 2/4] Drivers: hv: balloon: account for gaps in hot add regions
> >> static unsigned long handle_pg_range(unsigned long pg_start, @@ - > 834,13 > >> +881,19 @@ static unsigned long process_hot_add(unsigned long > pg_start, > >>unsigned long rg_size) > >> { > >>struct hv_hotadd_state *ha_region = NULL; > >> + int covered; > >> > >>if (pfn_cnt == 0) > >>return 0; > >> > >> - if (!dm_device.host_specified_ha_region) > >> - if (pfn_covered(pg_start, pfn_cnt)) > >> + if (!dm_device.host_specified_ha_region) { > >> + covered = pfn_covered(pg_start, pfn_cnt); > >> + if (covered < 0) > >> + return 0; > > > > If the hot-add pages aren't covered by any region, then shouldn't it fall > through instead of returning? > > That way the new ha_region can be added to the list and we hot-add the > > pages accordingly. > > I was under an impression this is impossible: > hot_add_req()/process_hot_add() will create a new region in this > case. 'covered < 0' was added to handle one particular error: failure to > allocate memory to record gap (struct hv_hotadd_gap) and I don't have a > better idea how to handle this: if we can't remember the gap we'll crash > later on onlining... > You are correct. I misread your patch thinking "covered < 0" means the page range is not covered; when it is actually handling an error case. > -- > Vitaly ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 55/80] staging: lustre: mdt: add mbo_ prefix to members of struct mdt_body
From: John L. HammondRename each member of struct mdt_body, adding the prefix mbo_. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675 Reviewed-on: http://review.whamcloud.com/10202 Reviewed-by: Bob Glossman Reviewed-by: Andreas Dilger Reviewed-by: Mike Pershin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h | 74 +++--- drivers/staging/lustre/lustre/include/lustre_mdc.h | 14 +- drivers/staging/lustre/lustre/llite/dir.c | 30 +- drivers/staging/lustre/lustre/llite/file.c | 20 +- drivers/staging/lustre/lustre/llite/lcommon_cl.c |2 +- drivers/staging/lustre/lustre/llite/llite_lib.c| 110 drivers/staging/lustre/lustre/llite/llite_nfs.c|6 +- drivers/staging/lustre/lustre/llite/namei.c| 44 ++-- drivers/staging/lustre/lustre/llite/statahead.c|4 +- drivers/staging/lustre/lustre/llite/symlink.c |6 +- drivers/staging/lustre/lustre/llite/xattr.c| 14 +- drivers/staging/lustre/lustre/llite/xattr_cache.c | 12 +- drivers/staging/lustre/lustre/lmv/lmv_intent.c | 38 ++-- drivers/staging/lustre/lustre/lmv/lmv_obd.c| 16 +- drivers/staging/lustre/lustre/mdc/mdc_lib.c| 62 +++--- drivers/staging/lustre/lustre/mdc/mdc_locks.c | 32 ++-- drivers/staging/lustre/lustre/mdc/mdc_reint.c |4 +- drivers/staging/lustre/lustre/mdc/mdc_request.c| 52 ++-- .../staging/lustre/lustre/ptlrpc/pack_generic.c| 56 ++-- drivers/staging/lustre/lustre/ptlrpc/wiretest.c| 268 ++-- 20 files changed, 432 insertions(+), 432 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index bbf0c8d..400ab3c 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -2097,43 +2097,43 @@ enum md_transient_state { }; struct mdt_body { - struct lu_fid fid1; - struct lu_fid fid2; - struct lustre_handle handle; - __u64 valid; - __u64 size; /* Offset, in the case of MDS_READPAGE */ - __s64 mtime; - __s64 atime; - __s64 ctime; - __u64 blocks; /* XID, in the case of MDS_READPAGE */ - __u64 ioepoch; - __u64 t_state; /* transient file state defined in - * enum md_transient_state - * was "ino" until 2.4.0 - */ - __u32 fsuid; - __u32 fsgid; - __u32 capability; - __u32 mode; - __u32 uid; - __u32 gid; - __u32 flags; /* from vfs for pin/unpin, LUSTRE_BFLAG close */ - __u32 rdev; - __u32 nlink; /* #bytes to read in the case of MDS_READPAGE */ - __u32 unused2; /* was "generation" until 2.4.0 */ - __u32 suppgid; - __u32 eadatasize; - __u32 aclsize; - __u32 max_mdsize; - __u32 max_cookiesize; - __u32 uid_h; /* high 32-bits of uid, for FUID */ - __u32 gid_h; /* high 32-bits of gid, for FUID */ - __u32 padding_5; /* also fix lustre_swab_mdt_body */ - __u64 padding_6; - __u64 padding_7; - __u64 padding_8; - __u64 padding_9; - __u64 padding_10; + struct lu_fid mbo_fid1; + struct lu_fid mbo_fid2; + struct lustre_handle mbo_handle; + __u64 mbo_valid; + __u64 mbo_size; /* Offset, in the case of MDS_READPAGE */ + __s64 mbo_mtime; + __s64 mbo_atime; + __s64 mbo_ctime; + __u64 mbo_blocks; /* XID, in the case of MDS_READPAGE */ + __u64 mbo_ioepoch; + __u64 mbo_t_state;/* transient file state defined in +* enum md_transient_state +* was "ino" until 2.4.0 +*/ + __u32 mbo_fsuid; + __u32 mbo_fsgid; + __u32 mbo_capability; + __u32 mbo_mode; + __u32 mbo_uid; + __u32 mbo_gid; + __u32 mbo_flags; + __u32 mbo_rdev; + __u32 mbo_nlink; /* #bytes to read in the case of MDS_READPAGE */ + __u32 mbo_unused2;/* was "generation" until 2.4.0 */ + __u32 mbo_suppgid; + __u32 mbo_eadatasize; + __u32 mbo_aclsize; + __u32 mbo_max_mdsize; + __u32 mbo_max_cookiesize; + __u32 mbo_uid_h; /* high 32-bits of uid, for FUID */ + __u32 mbo_gid_h; /* high 32-bits of gid, for FUID */ + __u32 mbo_padding_5; /* also fix lustre_swab_mdt_body */ + __u64 mbo_padding_6; + __u64
[PATCH 61/80] staging: lustre: obdclass: compile issues with variable not being initialized
One of the versions of gcc I have refuses to build obd_mount.c due to index not be initialized in function lmd_make_exclusion before it is used. Signed-off-by: James SimmonsIntel-bug-id: https://jira.hpdd.intel.com/browse/LU-4629 Reviewed-on: http://review.whamcloud.com/10705 Reviewed-by: John L. Hammond Reviewed-by: Bob Glossman Reviewed-by: Dmitry Eremin Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/obdclass/obd_mount.c |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/drivers/staging/lustre/lustre/obdclass/obd_mount.c b/drivers/staging/lustre/lustre/obdclass/obd_mount.c index 33d6c42..595ea1f 100644 --- a/drivers/staging/lustre/lustre/obdclass/obd_mount.c +++ b/drivers/staging/lustre/lustre/obdclass/obd_mount.c @@ -730,7 +730,7 @@ int lustre_check_exclusion(struct super_block *sb, char *svname) static int lmd_make_exclusion(struct lustre_mount_data *lmd, const char *ptr) { const char *s1 = ptr, *s2; - __u32 index, *exclude_list; + __u32 index = 0, *exclude_list; int rc = 0, devmax; /* The shortest an ost name can be is 8 chars: -OST. -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 63/80] staging: lustre: fid: do open-by-fid by default
From: Lai SiyaoCurrently client open-by-fid often packs name into the request, but the name may be invalid, eg. NFS export, and even if it's valid, it may cause inconsistency because this operation is done on this fid, which is globally unique, but name not. Since open-by-fid doesn't pack name, for striped dir we can't know parent stripe fid on client, so we set parent fid the same as child fid, and MDT has to find its parent fid from linkea (this is already supported by MDT). M_CHECK_STALE becomes obsolete. Unset MDS_OPEN_FL_INTERNAL from open syscall flags, because these flags are internally used, and should not be set from user space. It's not necessary to store parent fid in lli_pfid, because MDT can get it's parent fid from linkea, and now that DNE stripe directory stores master inode fid in lli_pfid, stop storing parent fid to avoid conflict. Signed-off-by: Lai Siyao Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3544 Reviewed-on: http://review.whamcloud.com/7476 Reviewed-on: http://review.whamcloud.com/10692 Reviewed-by: Fan Yong Reviewed-by: Nathaniel Clark Reviewed-by: wangdi Reviewed-by: John L. Hammond Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h |5 ++ .../staging/lustre/lustre/include/lustre_lite.h|1 - drivers/staging/lustre/lustre/include/lustre_mds.h |3 - drivers/staging/lustre/lustre/llite/file.c | 71 +--- .../staging/lustre/lustre/llite/llite_internal.h |4 +- drivers/staging/lustre/lustre/llite/llite_lib.c| 17 + drivers/staging/lustre/lustre/llite/llite_nfs.c| 14 +++- drivers/staging/lustre/lustre/llite/namei.c|1 + drivers/staging/lustre/lustre/lmv/lmv_intent.c | 41 +-- drivers/staging/lustre/lustre/mdc/mdc_internal.h |1 - drivers/staging/lustre/lustre/mdc/mdc_lib.c|5 +- drivers/staging/lustre/lustre/mdc/mdc_locks.c | 21 -- .../lustre/lustre/obdclass/lprocfs_status.c|2 +- 13 files changed, 71 insertions(+), 115 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 400ab3c..a9661c0 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -2252,6 +2252,11 @@ void lustre_swab_mdt_rec_setattr(struct mdt_rec_setattr *sa); */ #define MDS_OPEN_RELEASE 02ULL /* Open the file for HSM release */ +#define MDS_OPEN_FL_INTERNAL (MDS_OPEN_HAS_EA | MDS_OPEN_HAS_OBJS |\ + MDS_OPEN_OWNEROVERRIDE | MDS_OPEN_LOCK | \ + MDS_OPEN_BY_FID | MDS_OPEN_LEASE |\ + MDS_OPEN_RELEASE) + enum mds_op_bias { MDS_CHECK_SPLIT = 1 << 0, MDS_CROSS_REF = 1 << 1, diff --git a/drivers/staging/lustre/lustre/include/lustre_lite.h b/drivers/staging/lustre/lustre/include/lustre_lite.h index b168977..a3d7573 100644 --- a/drivers/staging/lustre/lustre/include/lustre_lite.h +++ b/drivers/staging/lustre/lustre/include/lustre_lite.h @@ -42,7 +42,6 @@ #include "obd_class.h" #include "lustre_net.h" -#include "lustre_mds.h" #include "lustre_ha.h" /* 4UL * 1024 * 1024 */ diff --git a/drivers/staging/lustre/lustre/include/lustre_mds.h b/drivers/staging/lustre/lustre/include/lustre_mds.h index 4104bd9..23a7e4f 100644 --- a/drivers/staging/lustre/lustre/include/lustre_mds.h +++ b/drivers/staging/lustre/lustre/include/lustre_mds.h @@ -58,9 +58,6 @@ struct mds_group_info { #define MDD_OBD_NAME "mdd_obd" #define MDD_OBD_UUID "mdd_obd_uuid" -/* these are local flags, used only on the client, private */ -#define M_CHECK_STALE 02 - /** @} mds */ #endif diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c index 563cdf6..015b0ab 100644 --- a/drivers/staging/lustre/lustre/llite/file.c +++ b/drivers/staging/lustre/lustre/llite/file.c @@ -379,53 +379,35 @@ int ll_file_release(struct inode *inode, struct file *file) return rc; } -static int ll_intent_file_open(struct dentry *dentry, void *lmm, - int lmmsize, struct lookup_intent *itp) +static int ll_intent_file_open(struct dentry *de, void *lmm, int lmmsize, + struct lookup_intent *itp) { - struct inode *inode = d_inode(dentry); + struct inode *inode = d_inode(de); struct ll_sb_info *sbi = ll_i2sbi(inode); - struct dentry *parent = dentry->d_parent; - const char *name = dentry->d_name.name; - const int len = dentry->d_name.len; + struct
[PATCH 68/80] staging: lustre: lmv: build master LMV EA dynamically build via readdir
From: Fan YongWhen creating a striped directory, the master object saves the slave objects (or shards) as internal sub-directories. The sub-directory's name is composed of ${shard_FID}:${shard_idx}. With the name, we can easily to know what the shard is and where it should be. On the other hand, we need to store some information related with the striped directory, such as magic, hash type, shards count, and so on. That is the LMV EA (header). We do NOT store the FID of each shard in the LMV EA. Instead, when we need the shards' FIDs (such as readdir() on client-side), we can build the entrie LMV EA on the MDT (in RAM) by iterating the sub-directory entries that are contained in the master object of the striped directroy. Above mechanism can simplify the striped directory create operation. For very large striped directory, logging the FIDs array in the LMV EA will be trouble. It also simplify the LFSCK for verifying striped directory, because it reduces the inconsistency sources. Another fixing is about the lmv_master_fid in master LMV EA header, it is redundant information, and may become one of the inconsistency sources. So replace it with two __u64 padding fields. Signed-off-by: Fan Yong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5223 Reviewed-on: http://review.whamcloud.com/10751 Reviewed-by: Andreas Dilger Reviewed-by: Alex Zhuravlev Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h |7 +-- drivers/staging/lustre/lustre/include/lustre_lmv.h | 30 drivers/staging/lustre/lustre/lmv/lmv_obd.c|4 -- drivers/staging/lustre/lustre/ptlrpc/wiretest.c| 49 4 files changed, 52 insertions(+), 38 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 05fe359..17581ba 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -2494,10 +2494,9 @@ struct lmv_mds_md_v1 { * for example migrating or dead. */ __u32 lmv_layout_version; /* Used for directory restriping */ - __u32 lmv_padding; - struct lu_fid lmv_master_fid; /* The FID of the master object, which -* is the namespace-visible dir FID -*/ + __u32 lmv_padding1; + __u64 lmv_padding2; + __u64 lmv_padding3; char lmv_pool_name[LOV_MAXPOOLNAME];/* pool name */ struct lu_fid lmv_stripe_fids[0]; /* FIDs for each stripe */ }; diff --git a/drivers/staging/lustre/lustre/include/lustre_lmv.h b/drivers/staging/lustre/lustre/include/lustre_lmv.h index 1dd3e92..085e596 100644 --- a/drivers/staging/lustre/lustre/include/lustre_lmv.h +++ b/drivers/staging/lustre/lustre/include/lustre_lmv.h @@ -48,7 +48,6 @@ struct lmv_stripe_md { __u32 lsm_md_layout_version; __u32 lsm_md_default_count; __u32 lsm_md_default_index; - struct lu_fid lsm_md_master_fid; charlsm_md_pool_name[LOV_MAXPOOLNAME]; struct lmv_oinfo lsm_md_oinfo[0]; }; @@ -90,23 +89,6 @@ static inline void lmv_free_memmd(struct lmv_stripe_md *lsm) lmv_unpack_md(NULL, , NULL, 0); } -static inline void lmv1_cpu_to_le(struct lmv_mds_md_v1 *lmv_dst, - const struct lmv_mds_md_v1 *lmv_src) -{ - int i; - - lmv_dst->lmv_magic = cpu_to_le32(lmv_src->lmv_magic); - lmv_dst->lmv_stripe_count = cpu_to_le32(lmv_src->lmv_stripe_count); - lmv_dst->lmv_master_mdt_index = - cpu_to_le32(lmv_src->lmv_master_mdt_index); - lmv_dst->lmv_hash_type = cpu_to_le32(lmv_src->lmv_hash_type); - lmv_dst->lmv_layout_version = cpu_to_le32(lmv_src->lmv_layout_version); - - for (i = 0; i < lmv_src->lmv_stripe_count; i++) - fid_cpu_to_le(_dst->lmv_stripe_fids[i], - _src->lmv_stripe_fids[i]); -} - static inline void lmv1_le_to_cpu(struct lmv_mds_md_v1 *lmv_dst, const struct lmv_mds_md_v1 *lmv_src) { @@ -124,18 +106,6 @@ static inline void lmv1_le_to_cpu(struct lmv_mds_md_v1 *lmv_dst, _src->lmv_stripe_fids[i]); } -static inline void lmv_cpu_to_le(union lmv_mds_md *lmv_dst, -const union lmv_mds_md *lmv_src) -{ - switch (lmv_src->lmv_magic) { - case LMV_MAGIC_V1: - lmv1_cpu_to_le(_dst->lmv_md_v1, _src->lmv_md_v1); - break; - default: - break; - } -} - static inline void lmv_le_to_cpu(union lmv_mds_md *lmv_dst,
[PATCH 72/80] staging: lustre: obd: validate open handle cookies
From: John L. HammondAdd a const void *h_owner member to struct portals_handle. Add a const void *owner parameter to class_handle2object() which must be matched by the h_owner member of the handle in addition to the cookie. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3233 Reviewed-on: http://review.whamcloud.com/6938 Reviewed-by: Andreas Dilger Reviewed-by: Fan Yong Reviewed-by: Mike Pershin Signed-off-by: James Simmons --- .../staging/lustre/lustre/include/lustre_handles.h |3 ++- drivers/staging/lustre/lustre/ldlm/ldlm_lock.c |2 +- drivers/staging/lustre/lustre/obdclass/genops.c|2 +- .../lustre/lustre/obdclass/lustre_handles.c|4 ++-- 4 files changed, 6 insertions(+), 5 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre_handles.h b/drivers/staging/lustre/lustre/include/lustre_handles.h index 1a63a6b..bc1dd46 100644 --- a/drivers/staging/lustre/lustre/include/lustre_handles.h +++ b/drivers/staging/lustre/lustre/include/lustre_handles.h @@ -66,6 +66,7 @@ struct portals_handle_ops { struct portals_handle { struct list_headh_link; __u64 h_cookie; + const void *h_owner; struct portals_handle_ops *h_ops; /* newly added fields to handle the RCU issue. -jxiong */ @@ -83,7 +84,7 @@ struct portals_handle { void class_handle_hash(struct portals_handle *, struct portals_handle_ops *ops); void class_handle_unhash(struct portals_handle *); -void *class_handle2object(__u64 cookie); +void *class_handle2object(__u64 cookie, const void *owner); void class_handle_free_cb(struct rcu_head *rcu); int class_handle_init(void); void class_handle_cleanup(void); diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c index a91cdb4..7a34caf 100644 --- a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c +++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c @@ -542,7 +542,7 @@ struct ldlm_lock *__ldlm_handle2lock(const struct lustre_handle *handle, LASSERT(handle); - lock = class_handle2object(handle->cookie); + lock = class_handle2object(handle->cookie, NULL); if (!lock) return NULL; diff --git a/drivers/staging/lustre/lustre/obdclass/genops.c b/drivers/staging/lustre/lustre/obdclass/genops.c index be25434..a739eb1 100644 --- a/drivers/staging/lustre/lustre/obdclass/genops.c +++ b/drivers/staging/lustre/lustre/obdclass/genops.c @@ -618,7 +618,7 @@ struct obd_export *class_conn2export(struct lustre_handle *conn) } CDEBUG(D_INFO, "looking for export cookie %#llx\n", conn->cookie); - export = class_handle2object(conn->cookie); + export = class_handle2object(conn->cookie, NULL); return export; } EXPORT_SYMBOL(class_conn2export); diff --git a/drivers/staging/lustre/lustre/obdclass/lustre_handles.c b/drivers/staging/lustre/lustre/obdclass/lustre_handles.c index 082f530..7ca68ae 100644 --- a/drivers/staging/lustre/lustre/obdclass/lustre_handles.c +++ b/drivers/staging/lustre/lustre/obdclass/lustre_handles.c @@ -130,7 +130,7 @@ void class_handle_unhash(struct portals_handle *h) } EXPORT_SYMBOL(class_handle_unhash); -void *class_handle2object(__u64 cookie) +void *class_handle2object(__u64 cookie, const void *owner) { struct handle_bucket *bucket; struct portals_handle *h; @@ -145,7 +145,7 @@ void *class_handle2object(__u64 cookie) rcu_read_lock(); list_for_each_entry_rcu(h, >head, h_link) { - if (h->h_cookie != cookie) + if (h->h_cookie != cookie || h->h_owner != owner) continue; spin_lock(>h_lock); -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 59/80] staging: lustre: fld: add fld description documentation
From: Patrick FarrellAdd subsystem description from Di Wang to header file. Signed-off-by: Patrick Farrell Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5153 Reviewed-on: http://review.whamcloud.com/10631 Reviewed-by: wangdi Reviewed-by: John L. Hammond Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/fld/fld_internal.h | 19 +++ 1 files changed, 19 insertions(+), 0 deletions(-) diff --git a/drivers/staging/lustre/lustre/fld/fld_internal.h b/drivers/staging/lustre/lustre/fld/fld_internal.h index f0efe5b..08eaec7 100644 --- a/drivers/staging/lustre/lustre/fld/fld_internal.h +++ b/drivers/staging/lustre/lustre/fld/fld_internal.h @@ -31,6 +31,25 @@ * * lustre/fld/fld_internal.h * + * Subsystem Description: + * FLD is FID Location Database, which stores where (IE, on which MDT) + * FIDs are located. + * The database is basically a record file, each record consists of a FID + * sequence range, MDT/OST index, and flags. The FLD for the whole FS + * is only stored on the sequence controller(MDT0) right now, but each target + * also has its local FLD, which only stores the local sequence. + * + * The FLD subsystem usually has two tasks: + * 1. maintain the database, i.e. when the sequence controller allocates + * new sequence ranges to some nodes, it will call the FLD API to insert the + * location information in FLDB. + * + * 2. Handle requests from other nodes, i.e. if client needs to know where + * the FID is located, if it can not find the information in the local cache, + * it will send a FLD lookup RPC to the FLD service, and the FLD service will + * look up the FLDB entry and return the location information to client. + * + * * Author: Yury Umanets * Author: Tom WangDi */ -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 62/80] staging: lustre: obd: limit lu_object cache
From: Brian BehlendorfAs the LU cache grows it can consume large enough chunks of memory that ends up preventing buffers for other objects, such as the OIs, from being cached and severely impacting the performance for FID lookups. Limit the lu_object cache to a maximum of lu_cache_nr objects. NOTES: * In order to be able to quickly determine the number of objects in the hash table the CFS_HASH_COUNTER flag is added. This adds an atomic_inc/dec to the hash insert/remove paths but is not expected to have any measurable impact of performance. Signed-off-by: Brian Behlendorf Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5164 Reviewed-on: http://review.whamcloud.com/10237 Reviewed-by: Andreas Dilger Reviewed-by: Alex Zhuravlev Reviewed-by: Isaac Huang Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/obdclass/lu_object.c | 91 ++-- 1 files changed, 64 insertions(+), 27 deletions(-) diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c index 0c00bf8..9d1c96b 100644 --- a/drivers/staging/lustre/lustre/obdclass/lu_object.c +++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c @@ -55,6 +55,34 @@ #include "../include/lu_ref.h" #include +enum { + LU_CACHE_PERCENT_MAX = 50, + LU_CACHE_PERCENT_DEFAULT = 20 +}; + +#define LU_CACHE_NR_MAX_ADJUST 128 +#define LU_CACHE_NR_UNLIMITED -1 +#define LU_CACHE_NR_DEFAULTLU_CACHE_NR_UNLIMITED +#define LU_CACHE_NR_LDISKFS_LIMIT LU_CACHE_NR_UNLIMITED +#define LU_CACHE_NR_ZFS_LIMIT 256 + +#define LU_SITE_BITS_MIN 12 +#define LU_SITE_BITS_MAX 24 +/** + * total 256 buckets, we don't want too many buckets because: + * - consume too much memory + * - avoid unbalanced LRU list + */ +#define LU_SITE_BKT_BITS 8 + +static unsigned int lu_cache_percent = LU_CACHE_PERCENT_DEFAULT; +module_param(lu_cache_percent, int, 0644); +MODULE_PARM_DESC(lu_cache_percent, "Percentage of memory to be used as lu_object cache"); + +static long lu_cache_nr = LU_CACHE_NR_DEFAULT; +module_param(lu_cache_nr, long, 0644); +MODULE_PARM_DESC(lu_cache_nr, "Maximum number of objects in lu_object cache"); + static void lu_object_free(const struct lu_env *env, struct lu_object *o); static __u32 ls_stats_read(struct lprocfs_stats *stats, int idx); @@ -573,6 +601,27 @@ static struct lu_object *lu_object_find(const struct lu_env *env, return lu_object_find_at(env, dev->ld_site->ls_top_dev, f, conf); } +/* + * Limit the lu_object cache to a maximum of lu_cache_nr objects. Because + * the calculation for the number of objects to reclaim is not covered by + * a lock the maximum number of objects is capped by LU_CACHE_MAX_ADJUST. + * This ensures that many concurrent threads will not accidentally purge + * the entire cache. + */ +static void lu_object_limit(const struct lu_env *env, struct lu_device *dev) +{ + __u64 size, nr; + + if (lu_cache_nr == LU_CACHE_NR_UNLIMITED) + return; + + size = cfs_hash_size_get(dev->ld_site->ls_obj_hash); + nr = (__u64)lu_cache_nr; + if (size > nr) + lu_site_purge(env, dev->ld_site, + min_t(__u64, size - nr, LU_CACHE_NR_MAX_ADJUST)); +} + static struct lu_object *lu_object_new(const struct lu_env *env, struct lu_device *dev, const struct lu_fid *f, @@ -590,6 +639,9 @@ static struct lu_object *lu_object_new(const struct lu_env *env, cfs_hash_bd_get_and_lock(hs, (void *)f, , 1); cfs_hash_bd_add_locked(hs, , >lo_header->loh_hash); cfs_hash_bd_unlock(hs, , 1); + + lu_object_limit(env, dev); + return o; } @@ -656,6 +708,9 @@ static struct lu_object *lu_object_find_try(const struct lu_env *env, if (likely(PTR_ERR(shadow) == -ENOENT)) { cfs_hash_bd_add_locked(hs, , >lo_header->loh_hash); cfs_hash_bd_unlock(hs, , 1); + + lu_object_limit(env, dev); + return o; } @@ -805,20 +860,12 @@ void lu_site_print(const struct lu_env *env, struct lu_site *s, void *cookie, } EXPORT_SYMBOL(lu_site_print); -enum { - LU_CACHE_PERCENT_MAX = 50, - LU_CACHE_PERCENT_DEFAULT = 20 -}; - -static unsigned int lu_cache_percent = LU_CACHE_PERCENT_DEFAULT; -module_param(lu_cache_percent, int, 0644); -MODULE_PARM_DESC(lu_cache_percent, "Percentage of memory to be used as lu_object cache"); - /** * Return desired hash table order. */ -static int lu_htable_order(void) +static int lu_htable_order(struct lu_device *top) { + unsigned long bits_max = LU_SITE_BITS_MAX; unsigned long cache_size; int
[PATCH 75/80] staging: lustre: llite: set op_max_pages
From: wang diCache the maximum allowed pages supported by the llite layer. This value will be used in the mdc and lmv layer. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/dir.c |1 + drivers/staging/lustre/lustre/llite/llite_nfs.c |1 + drivers/staging/lustre/lustre/llite/statahead.c |6 ++ 3 files changed, 8 insertions(+), 0 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index 9c7fa8f..ed09015 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -649,6 +649,7 @@ static int ll_readdir(struct file *filp, struct dir_context *ctx) } } } + op_data->op_max_pages = sbi->ll_md_brw_pages; ctx->pos = pos; rc = ll_dir_read(inode, , op_data, ctx); pos = ctx->pos; diff --git a/drivers/staging/lustre/lustre/llite/llite_nfs.c b/drivers/staging/lustre/lustre/llite/llite_nfs.c index 2b65240..1e156dc 100644 --- a/drivers/staging/lustre/lustre/llite/llite_nfs.c +++ b/drivers/staging/lustre/lustre/llite/llite_nfs.c @@ -276,6 +276,7 @@ static int ll_get_name(struct dentry *dentry, char *name, goto out; } + op_data->op_max_pages = ll_i2sbi(dir)->ll_md_brw_pages; inode_lock(dir); rc = ll_dir_read(dir, , op_data, ); inode_unlock(dir); diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c index 46b8faf..454c33e 100644 --- a/drivers/staging/lustre/lustre/llite/statahead.c +++ b/drivers/staging/lustre/lustre/llite/statahead.c @@ -1052,6 +1052,8 @@ static int ll_statahead_thread(void *arg) if (IS_ERR(op_data)) return PTR_ERR(op_data); + op_data->op_max_pages = ll_i2sbi(dir)->ll_md_brw_pages; + if (sbi->ll_flags & LL_SBI_AGL_ENABLED) ll_start_agl(parent, sai); @@ -1355,6 +1357,10 @@ static int is_first_dirent(struct inode *dir, struct dentry *dentry) LUSTRE_OPC_ANY, dir); if (IS_ERR(op_data)) return PTR_ERR(op_data); + /** +* FIXME choose the start offset of the readdir +*/ + op_data->op_max_pages = ll_i2sbi(dir)->ll_md_brw_pages; ll_dir_chain_init(); page = ll_get_dir_page(dir, op_data, pos, ); -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 77/80] staging: lustre: lnet: Correct position of lnet_ni_decref()
From: Doug OucharekIn fix http://review.whamcloud.com/#/c/19614/, the call to lnet_ni_decref() should have followed the routines which are using the NI. This patch correct that. Signed-off-by: Doug Oucharek Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8022 Reviewed-on: http://review.whamcloud.com/21001 Reviewed-by: James Simmons Reviewed-by: John L. Hammond Reviewed-by: Dmitry Eremin Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c index 9eb1db6..19c90fc 100644 --- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c +++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c @@ -2526,9 +2526,9 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) failed: if (ni) { - lnet_ni_decref(ni); rej.ibr_cp.ibcp_queue_depth = kiblnd_msg_queue_size(version, ni); rej.ibr_cp.ibcp_max_frags = kiblnd_rdma_frags(version, ni); + lnet_ni_decref(ni); } rej.ibr_version = version; -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH] staging: lustre: fix bare use of 'unsigned'
fix checkpatch.pl warning Prefer 'unsigned int' to bare use of 'unsigned' Signed-off-by: Anson Jacob--- drivers/staging/lustre/lustre/include/cl_object.h | 14 +++--- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h index 3cd4a25..db910ea 100644 --- a/drivers/staging/lustre/lustre/include/cl_object.h +++ b/drivers/staging/lustre/lustre/include/cl_object.h @@ -367,7 +367,7 @@ struct cl_object_operations { * cl_object_operations::coo_attr_get() is used. */ int (*coo_attr_set)(const struct lu_env *env, struct cl_object *obj, - const struct cl_attr *attr, unsigned valid); + const struct cl_attr *attr, unsigned int valid); /** * Update object configuration. Called top-to-bottom to modify object * configuration. @@ -757,7 +757,7 @@ struct cl_page { /** Link to a queue, for debugging. */ struct lu_ref_link cp_queue_ref; /** Per-page flags from enum cl_page_flags. Protected by a VM lock. */ - unsigned cp_flags; + unsigned int cp_flags; /** Assigned if doing a sync_io */ struct cl_sync_io *cp_sync_io; }; @@ -1316,7 +1316,7 @@ do { \ * @{ */ struct cl_page_list { - unsigned pl_nr; + unsigned int pl_nr; struct list_head pl_pages; struct task_struct *pl_owner; }; @@ -1834,7 +1834,7 @@ struct cl_io { /** * Number of pages owned by this IO. For invariant checking. */ - unsigned ci_owned_nr; + unsigned int ci_owned_nr; }; /** @} cl_io */ @@ -1992,11 +1992,11 @@ struct cl_req { /** A list of pages being transferred */ struct list_headcrq_pages; /** Number of pages in cl_req::crq_pages */ - unsigned crq_nrpages; + unsigned int crq_nrpages; /** An array of objects which pages are in ->crq_pages */ struct cl_req_obj*crq_o; /** Number of elements in cl_req::crq_objs[] */ - unsigned crq_nrobjs; + unsigned int crq_nrobjs; struct list_headcrq_layers; }; @@ -2177,7 +2177,7 @@ void cl_object_attr_unlock(struct cl_object *o); int cl_object_attr_get(const struct lu_env *env, struct cl_object *obj, struct cl_attr *attr); int cl_object_attr_set(const struct lu_env *env, struct cl_object *obj, - const struct cl_attr *attr, unsigned valid); + const struct cl_attr *attr, unsigned int valid); int cl_object_glimpse(const struct lu_env *env, struct cl_object *obj, struct ost_lvb *lvb); int cl_conf_set(const struct lu_env *env, struct cl_object *obj, -- 2.7.4 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 74/80] staging: lustre: obd: implement md_read_page
From: wang diThis patch adds md_read_page which is a new more flexiable api that will replace md_readpage. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/10761 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4906 Reviewed-by: John L. Hammond Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/include/obd.h | 13 - drivers/staging/lustre/lustre/include/obd_class.h | 15 +++ 2 files changed, 27 insertions(+), 1 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h index e91f65a..92eebff 100644 --- a/drivers/staging/lustre/lustre/include/obd.h +++ b/drivers/staging/lustre/lustre/include/obd.h @@ -830,6 +830,15 @@ struct md_op_data { struct lustre_handleop_lease_handle; }; +#define op_stripe_offset op_ioepoch +#define op_max_pages op_valid + +struct md_callback { + int (*md_blocking_ast)(struct ldlm_lock *lock, + struct ldlm_lock_desc *desc, + void *data, int flag); +}; + enum op_cli_flags { CLI_SET_MEA = 1 << 0, CLI_RM_ENTRY= 1 << 1, @@ -1039,7 +1048,9 @@ struct md_ops { struct ptlrpc_request **); int (*readpage)(struct obd_export *, struct md_op_data *, struct page **, struct ptlrpc_request **); - + int (*read_page)(struct obd_export *, struct md_op_data *, +struct md_callback *cb_op, __u64 hash_offset, +struct page **ppage); int (*unlink)(struct obd_export *, struct md_op_data *, struct ptlrpc_request **); diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h index 69b628b..daca5a0 100644 --- a/drivers/staging/lustre/lustre/include/obd_class.h +++ b/drivers/staging/lustre/lustre/include/obd_class.h @@ -1535,6 +1535,21 @@ static inline int md_readpage(struct obd_export *exp, struct md_op_data *opdata, return rc; } +static inline int md_read_page(struct obd_export *exp, + struct md_op_data *op_data, + struct md_callback *cb_op, + __u64 hash_offset, + struct page **ppage) +{ + int rc; + + EXP_CHECK_MD_OP(exp, read_page); + EXP_MD_COUNTER_INCREMENT(exp, read_page); + rc = MDP(exp->exp_obd, read_page)(exp, op_data, cb_op, hash_offset, + ppage); + return rc; +} + static inline int md_unlink(struct obd_export *exp, struct md_op_data *op_data, struct ptlrpc_request **request) { -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 73/80] staging: lustre: lmv: build error with gcc 4.7.0 20110509
From: Dmitry EreminFixed comparison between signed and unsigned indexes. Signed-off-by: Dmitry Eremin Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3775 Reviewed-on: http://review.whamcloud.com/7382 Reviewed-by: John L. Hammond Reviewed-by: James Simmons Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_obd.c | 34 ++ 1 files changed, 18 insertions(+), 16 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index b8275e1..3e41f49 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -130,12 +130,12 @@ static void lmv_activate_target(struct lmv_obd *lmv, * -ENOTCONN: The UUID is found, but the target connection is bad (!) * -EBADF : The UUID is found, but the OBD of the wrong type (!) */ -static int lmv_set_mdc_active(struct lmv_obd *lmv, struct obd_uuid *uuid, +static int lmv_set_mdc_active(struct lmv_obd *lmv, const struct obd_uuid *uuid, int activate) { struct lmv_tgt_desc*uninitialized_var(tgt); struct obd_device *obd; - int i; + u32 i; int rc = 0; CDEBUG(D_INFO, "Searching in lmv %p for uuid %s (activate=%d)\n", @@ -307,7 +307,7 @@ static int lmv_connect(const struct lu_env *env, static void lmv_set_timeouts(struct obd_device *obd) { struct lmv_obd *lmv; - int i; + u32 i; lmv = >u.lmv; if (lmv->server_timeout == 0) @@ -333,7 +333,7 @@ static int lmv_init_ea_size(struct obd_export *exp, int easize, { struct obd_device *obd = exp->exp_obd; struct lmv_obd *lmv = >u.lmv; - int i; + u32 i; int rc = 0; int change = 0; @@ -578,7 +578,7 @@ int lmv_check_connect(struct obd_device *obd) { struct lmv_obd *lmv = >u.lmv; struct lmv_tgt_desc *tgt; - inti; + u32 i; intrc; inteasize; @@ -693,7 +693,7 @@ static int lmv_disconnect(struct obd_export *exp) struct obd_device *obd = class_exp2obd(exp); struct lmv_obd *lmv = >u.lmv; int rc; - int i; + u32 i; if (!lmv->tgts) goto out_local; @@ -822,7 +822,7 @@ static int lmv_hsm_req_count(struct lmv_obd *lmv, const struct hsm_user_request *hur, const struct lmv_tgt_desc *tgt_mds) { - int i, nr = 0; + u32 i, nr = 0; struct lmv_tgt_desc*curr_tgt; /* count how many requests must be sent to the given target */ @@ -963,10 +963,10 @@ static int lmv_iocontrol(unsigned int cmd, struct obd_export *exp, struct obd_device*obddev = class_exp2obd(exp); struct lmv_obd *lmv = >u.lmv; struct lmv_tgt_desc *tgt = NULL; - inti = 0; + u32 i = 0; intrc = 0; intset = 0; - intcount = lmv->desc.ld_tgt_count; + u32 count = lmv->desc.ld_tgt_count; if (count == 0) return -ENOTTY; @@ -1444,7 +1444,7 @@ static int lmv_statfs(const struct lu_env *env, struct obd_export *exp, struct lmv_obd *lmv = >u.lmv; struct obd_statfs *temp; int rc = 0; - int i; + u32 i; rc = lmv_check_connect(obd); if (rc) @@ -1586,7 +1586,7 @@ static int lmv_null_inode(struct obd_export *exp, const struct lu_fid *fid) { struct obd_device *obd = exp->exp_obd; struct lmv_obd *lmv = >u.lmv; - int i; + u32 i; int rc; rc = lmv_check_connect(obd); @@ -1615,7 +1615,7 @@ static int lmv_find_cbdata(struct obd_export *exp, const struct lu_fid *fid, struct obd_device *obd = exp->exp_obd; struct lmv_obd *lmv = >u.lmv; int tgt; - int i; + u32 i; int rc; rc = lmv_check_connect(obd); @@ -2923,7 +2923,7 @@ static int lmv_cancel_unused(struct obd_export *exp, const struct lu_fid *fid, struct lmv_obd*lmv = >u.lmv; int rc = 0; int err; - int i; + u32 i; LASSERT(fid); @@ -2966,7 +2966,7 @@ static enum ldlm_mode lmv_lock_match(struct obd_export *exp, __u64 flags, struct lmv_obd*lmv = >u.lmv; enum ldlm_moderc; int tgt; - int
[PATCH 76/80] staging: lustre: lnet: Do not drop message when shutting down LNet
From: Doug OucharekThere is a case in lnet_parse() where we discover that LNet is shutting down but we continue to use the NI when we drop the message and end up calling ko2iblnd_check_send_locked() which tries to allocate from the Tx pool which has been cleaned up already. This triggers a NULL pointer dereference. This fix just returns from lnet_parse() when we disover LNet is shutting down. Signed-off-by: Doug Oucharek Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8106 Reviewed-on: http://review.whamcloud.com/19993 Reviewed-by: James Simmons Reviewed-by: Liang Zhen Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lnet/lnet/lib-move.c |3 +++ 1 files changed, 3 insertions(+), 0 deletions(-) diff --git a/drivers/staging/lustre/lnet/lnet/lib-move.c b/drivers/staging/lustre/lnet/lnet/lib-move.c index 6a3f2e1..5598fa8 100644 --- a/drivers/staging/lustre/lnet/lnet/lib-move.c +++ b/drivers/staging/lustre/lnet/lnet/lib-move.c @@ -2002,6 +2002,9 @@ lnet_parse(lnet_ni_t *ni, lnet_hdr_t *hdr, lnet_nid_t from_nid, libcfs_nid2str(from_nid), libcfs_nid2str(src_nid), lnet_msgtyp2str(type), rc); lnet_msg_free(msg); + if (rc == -ESHUTDOWN) + /* We are shutting down. Don't do anything more */ + return 0; goto drop; } -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 70/80] staging: lustre: include: fix one off errors in lustre_id.h
During inspection of another patch Dan Carpenter noticed some one off errors in lustre_id.h. Fix the condition test for OBIF_MAX_OID. Signed-off-by: James Simmons--- .../lustre/lustre/include/lustre/lustre_idl.h |4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 17581ba..9545451 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -659,7 +659,7 @@ static inline void ostid_set_id(struct ost_id *oi, __u64 oid) oi->oi_fid.f_oid = oid; oi->oi_fid.f_ver = oid >> 48; } else { - if (oid > OBIF_MAX_OID) { + if (oid >= OBIF_MAX_OID) { CERROR("Bad %llu to set " DOSTID "\n", oid, POSTID(oi)); return; } @@ -684,7 +684,7 @@ static inline int fid_set_id(struct lu_fid *fid, __u64 oid) fid->f_oid = oid; fid->f_ver = oid >> 48; } else { - if (oid > OBIF_MAX_OID) { + if (oid >= OBIF_MAX_OID) { CERROR("Too large OID %#llx to set REG "DFID"\n", (unsigned long long)oid, PFID(fid)); return -EBADF; -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 1/9 RESEND] staging: most: hdm-usb: remove unused macro HW_RESYNC
This patch removes the macro HW_RESYNC that is not used anymore. Signed-off-by: Andrey ShvetsovSigned-off-by: Christian Gromm --- drivers/staging/most/hdm-usb/hdm_usb.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/staging/most/hdm-usb/hdm_usb.c b/drivers/staging/most/hdm-usb/hdm_usb.c index aeae071..d2139d1 100644 --- a/drivers/staging/most/hdm-usb/hdm_usb.c +++ b/drivers/staging/most/hdm-usb/hdm_usb.c @@ -44,7 +44,6 @@ #define USB_VENDOR_ID_SMSC 0x0424 /* VID: SMSC */ #define USB_DEV_ID_BRDG0xC001 /* PID: USB Bridge */ #define USB_DEV_ID_INIC0xCF18 /* PID: USB INIC */ -#define HW_RESYNC 0x /* DRCI Addresses */ #define DRCI_REG_NI_STATE 0x0100 #define DRCI_REG_PACKET_BW 0x0101 -- 1.9.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 2/8] staging: most: hdm-usb: assign spinlock to local variable
This patch assigns the spinlock of struct mdev to local spinlock_t variable to get rid of all the ugly dereferencing. Signed-off-by: Andrey ShvetsovSigned-off-by: Christian Gromm --- drivers/staging/most/hdm-usb/hdm_usb.c | 31 +++ 1 file changed, 19 insertions(+), 12 deletions(-) diff --git a/drivers/staging/most/hdm-usb/hdm_usb.c b/drivers/staging/most/hdm-usb/hdm_usb.c index 8d8c72c..4156a30 100644 --- a/drivers/staging/most/hdm-usb/hdm_usb.c +++ b/drivers/staging/most/hdm-usb/hdm_usb.c @@ -206,14 +206,15 @@ static void free_anchored_buffers(struct most_dev *mdev, unsigned int channel, { struct mbo *mbo; struct buf_anchor *anchor, *tmp; + spinlock_t *lock = mdev->anchor_list_lock + channel; /* temp. lock */ unsigned long flags; - spin_lock_irqsave(>anchor_list_lock[channel], flags); + spin_lock_irqsave(lock, flags); list_for_each_entry_safe(anchor, tmp, >anchor_list[channel], list) { struct urb *urb = anchor->urb; - spin_unlock_irqrestore(>anchor_list_lock[channel], flags); + spin_unlock_irqrestore(lock, flags); if (likely(urb)) { mbo = urb->context; if (!irqs_disabled()) { @@ -229,11 +230,11 @@ static void free_anchored_buffers(struct most_dev *mdev, unsigned int channel, } usb_free_urb(urb); } - spin_lock_irqsave(>anchor_list_lock[channel], flags); + spin_lock_irqsave(lock, flags); list_del(>list); kfree(anchor); } - spin_unlock_irqrestore(>anchor_list_lock[channel], flags); + spin_unlock_irqrestore(lock, flags); } /** @@ -397,12 +398,14 @@ static void hdm_write_completion(struct urb *urb) struct device *dev; unsigned int channel; unsigned long flags; + spinlock_t *lock; /* temp. lock */ mbo = urb->context; anchor = mbo->priv; mdev = to_mdev(mbo->ifp); channel = mbo->hdm_channel_id; dev = >usb_device->dev; + lock = mdev->anchor_list_lock + channel; if ((urb->status == -ENOENT) || (urb->status == -ECONNRESET) || (!mdev->is_channel_healthy[channel])) { @@ -433,9 +436,9 @@ static void hdm_write_completion(struct urb *urb) mbo->processed_length = urb->actual_length; } - spin_lock_irqsave(>anchor_list_lock[channel], flags); + spin_lock_irqsave(lock, flags); list_del(>list); - spin_unlock_irqrestore(>anchor_list_lock[channel], flags); + spin_unlock_irqrestore(lock, flags); kfree(anchor); if (likely(mbo->complete)) @@ -559,12 +562,14 @@ static void hdm_read_completion(struct urb *urb) struct device *dev; unsigned long flags; unsigned int channel; + spinlock_t *lock; /* temp. lock */ mbo = urb->context; anchor = mbo->priv; mdev = to_mdev(mbo->ifp); channel = mbo->hdm_channel_id; dev = >usb_device->dev; + lock = mdev->anchor_list_lock + channel; if ((urb->status == -ENOENT) || (urb->status == -ECONNRESET) || (!mdev->is_channel_healthy[channel])) { @@ -601,9 +606,9 @@ static void hdm_read_completion(struct urb *urb) mbo->status = MBO_E_INVAL; } } - spin_lock_irqsave(>anchor_list_lock[channel], flags); + spin_lock_irqsave(lock, flags); list_del(>list); - spin_unlock_irqrestore(>anchor_list_lock[channel], flags); + spin_unlock_irqrestore(lock, flags); kfree(anchor); if (likely(mbo->complete)) @@ -638,6 +643,7 @@ static int hdm_enqueue(struct most_interface *iface, int channel, unsigned long flags; unsigned long length; void *virt_address; + spinlock_t *lock; /* temp. lock */ if (unlikely(!iface || !mbo)) return -EIO; @@ -697,9 +703,10 @@ static int hdm_enqueue(struct most_interface *iface, int channel, } urb->transfer_flags |= URB_NO_TRANSFER_DMA_MAP; - spin_lock_irqsave(>anchor_list_lock[channel], flags); + lock = mdev->anchor_list_lock + channel; + spin_lock_irqsave(lock, flags); list_add_tail(>list, >anchor_list[channel]); - spin_unlock_irqrestore(>anchor_list_lock[channel], flags); + spin_unlock_irqrestore(lock, flags); retval = usb_submit_urb(urb, GFP_KERNEL); if (retval) { @@ -709,9 +716,9 @@ static int hdm_enqueue(struct most_interface *iface, int channel, return 0; _error_1: - spin_lock_irqsave(>anchor_list_lock[channel], flags); + spin_lock_irqsave(lock, flags); list_del(>list); - spin_unlock_irqrestore(>anchor_list_lock[channel], flags); +
[PATCH 4/8] staging: most: hdm-usb: remove completion object
Waiting for the urb_compl object to complete evaluates always as false. This patch removes this unnecessary completion object. Signed-off-by: Andrey ShvetsovSigned-off-by: Christian Gromm --- drivers/staging/most/hdm-usb/hdm_usb.c | 11 +-- 1 file changed, 1 insertion(+), 10 deletions(-) diff --git a/drivers/staging/most/hdm-usb/hdm_usb.c b/drivers/staging/most/hdm-usb/hdm_usb.c index 26b5c1b..29e98dc 100644 --- a/drivers/staging/most/hdm-usb/hdm_usb.c +++ b/drivers/staging/most/hdm-usb/hdm_usb.c @@ -72,7 +72,6 @@ struct buf_anchor { struct urb *urb; struct list_head list; - struct completion urb_compl; }; /** @@ -217,12 +216,7 @@ static void free_anchored_buffers(struct most_dev *mdev, unsigned int channel, spin_unlock_irqrestore(lock, flags); if (likely(urb)) { mbo = urb->context; - if (!irqs_disabled()) { - usb_kill_urb(urb); - } else { - usb_unlink_urb(urb); - wait_for_completion(>urb_compl); - } + usb_kill_urb(urb); if ((mbo) && (mbo->complete)) { mbo->status = status; mbo->processed_length = 0; @@ -416,7 +410,6 @@ static void hdm_write_completion(struct urb *urb) if ((urb->status == -ENOENT) || (urb->status == -ECONNRESET) || (!mdev->is_channel_healthy[channel])) { spin_unlock_irqrestore(lock, flags); - complete(>urb_compl); return; } @@ -582,7 +575,6 @@ static void hdm_read_completion(struct urb *urb) if ((urb->status == -ENOENT) || (urb->status == -ECONNRESET) || (!mdev->is_channel_healthy[channel])) { spin_unlock_irqrestore(lock, flags); - complete(>urb_compl); return; } @@ -678,7 +670,6 @@ static int hdm_enqueue(struct most_interface *iface, int channel, } anchor->urb = urb; - init_completion(>urb_compl); mbo->priv = anchor; if ((mdev->padding_active[channel]) && -- 1.9.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 7/8] staging: most: hdm-usb: remove unnecessary status assignment
The USB completion callbacks set the status field of an MBO object before scheduling the clear_work. This patch removes this redundant assignment as the work_struct does the same for all MBOs. Signed-off-by: Andrey ShvetsovSigned-off-by: Christian Gromm --- drivers/staging/most/hdm-usb/hdm_usb.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/staging/most/hdm-usb/hdm_usb.c b/drivers/staging/most/hdm-usb/hdm_usb.c index 453b641..f44f27e 100644 --- a/drivers/staging/most/hdm-usb/hdm_usb.c +++ b/drivers/staging/most/hdm-usb/hdm_usb.c @@ -412,7 +412,6 @@ static void hdm_write_completion(struct urb *urb) dev_warn(dev, "Broken OUT pipe detected\n"); mdev->is_channel_healthy[channel] = false; spin_unlock_irqrestore(lock, flags); - mbo->status = MBO_E_INVAL; mdev->clear_work[channel].pipe = urb->pipe; schedule_work(>clear_work[channel].ws); return; @@ -570,7 +569,6 @@ static void hdm_read_completion(struct urb *urb) dev_warn(dev, "Broken IN pipe detected\n"); mdev->is_channel_healthy[channel] = false; spin_unlock_irqrestore(lock, flags); - mbo->status = MBO_E_INVAL; mdev->clear_work[channel].pipe = urb->pipe; schedule_work(>clear_work[channel].ws); return; -- 1.9.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 1/8] staging: most: hdm-usb: fix race between enqueue and most_stop_enqueue
The "broken in pipe" handler of the USB-HDM calls most_stop_enqueue() to stop the MBO traffic before returning all MBOs back to the Mostcore. As the enqueue() call from the Mostcore may run in parallel with the most_stop_enqueue(), the HDM may run into the inconsistent state and crash the kernel. This patch synchronizes enqueue(), most_stop_enqueue() and most_resume_enqueue() with a mutex, hence avoiding the race condition. Signed-off-by: Andrey ShvetsovSigned-off-by: Christian Gromm --- drivers/staging/most/hdm-usb/hdm_usb.c | 3 +- drivers/staging/most/mostcore/core.c | 53 -- 2 files changed, 38 insertions(+), 18 deletions(-) diff --git a/drivers/staging/most/hdm-usb/hdm_usb.c b/drivers/staging/most/hdm-usb/hdm_usb.c index 7f00aaf..8d8c72c 100644 --- a/drivers/staging/most/hdm-usb/hdm_usb.c +++ b/drivers/staging/most/hdm-usb/hdm_usb.c @@ -415,7 +415,6 @@ static void hdm_write_completion(struct urb *urb) switch (urb->status) { case -EPIPE: dev_warn(dev, "Broken OUT pipe detected\n"); - most_stop_enqueue(>iface, channel); mdev->is_channel_healthy[channel] = false; mbo->status = MBO_E_INVAL; mdev->clear_work[channel].pipe = urb->pipe; @@ -578,7 +577,6 @@ static void hdm_read_completion(struct urb *urb) switch (urb->status) { case -EPIPE: dev_warn(dev, "Broken IN pipe detected\n"); - most_stop_enqueue(>iface, channel); mdev->is_channel_healthy[channel] = false; mbo->status = MBO_E_INVAL; mdev->clear_work[channel].pipe = urb->pipe; @@ -928,6 +926,7 @@ static void wq_clear_halt(struct work_struct *wq_obj) int pipe = clear_work->pipe; mutex_lock(>io_mutex); + most_stop_enqueue(>iface, channel); free_anchored_buffers(mdev, channel, MBO_E_INVAL); if (usb_clear_halt(mdev->usb_device, pipe)) dev_warn(>usb_device->dev, "Failed to reset endpoint.\n"); diff --git a/drivers/staging/most/mostcore/core.c b/drivers/staging/most/mostcore/core.c index b03cdc9..db0606ca 100644 --- a/drivers/staging/most/mostcore/core.c +++ b/drivers/staging/most/mostcore/core.c @@ -51,6 +51,7 @@ struct most_c_obj { u16 channel_id; bool is_poisoned; struct mutex start_mutex; + struct mutex nq_mutex; /* nq thread synchronization */ int is_starving; struct most_interface *iface; struct most_inst_obj *inst; @@ -1131,18 +1132,18 @@ static inline void trash_mbo(struct mbo *mbo) spin_unlock_irqrestore(>fifo_lock, flags); } -static struct mbo *get_hdm_mbo(struct most_c_obj *c) +static bool hdm_mbo_ready(struct most_c_obj *c) { - unsigned long flags; - struct mbo *mbo; + bool empty; - spin_lock_irqsave(>fifo_lock, flags); - if (c->enqueue_halt || list_empty(>halt_fifo)) - mbo = NULL; - else - mbo = list_pop_mbo(>halt_fifo); - spin_unlock_irqrestore(>fifo_lock, flags); - return mbo; + if (c->enqueue_halt) + return false; + + spin_lock_irq(>fifo_lock); + empty = list_empty(>halt_fifo); + spin_unlock_irq(>fifo_lock); + + return !empty; } static void nq_hdm_mbo(struct mbo *mbo) @@ -1160,20 +1161,32 @@ static int hdm_enqueue_thread(void *data) { struct most_c_obj *c = data; struct mbo *mbo; + int ret; typeof(c->iface->enqueue) enqueue = c->iface->enqueue; while (likely(!kthread_should_stop())) { wait_event_interruptible(c->hdm_fifo_wq, -(mbo = get_hdm_mbo(c)) || +hdm_mbo_ready(c) || kthread_should_stop()); - if (unlikely(!mbo)) + mutex_lock(>nq_mutex); + spin_lock_irq(>fifo_lock); + if (unlikely(c->enqueue_halt || list_empty(>halt_fifo))) { + spin_unlock_irq(>fifo_lock); + mutex_unlock(>nq_mutex); continue; + } + + mbo = list_pop_mbo(>halt_fifo); + spin_unlock_irq(>fifo_lock); if (c->cfg.direction == MOST_CH_RX) mbo->buffer_length = c->cfg.buffer_size; - if (unlikely(enqueue(mbo->ifp, mbo->hdm_channel_id, mbo))) { + ret = enqueue(mbo->ifp, mbo->hdm_channel_id, mbo); + mutex_unlock(>nq_mutex); + + if (unlikely(ret)) { pr_err("hdm enqueue failed\n"); nq_hdm_mbo(mbo); c->hdm_enqueue_task = NULL; @@ -1759,6
[PATCH 41/80] staging: lustre: lmv: separate master object with master stripe
From: wang diSeparate master stripe with master object, so 1. stripeEA only exists on master object. 2. sub-stripe object will be inserted into master object as sub-directory, and it can get the master object by "..". By this, it will remove those specilities for stripe0 in LMV and LOD. And also simplify LFSCK, i.e. consistency check would be easier. When then master object becomes an orphan, we should mark all of its sub-stripes as dead object as well, otherwise client might still be able to create files under these stripes. A few fixes for striped directory layout lock: 1. stripe 0 should be locked as EX, same as other stripes. 2. Acquire the layout for directory, when it is being unliked. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4690 Reviewed-on: http://review.whamcloud.com/9511 Reviewed-by: Andreas Dilger Reviewed-by: John L. Hammond Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h | 64 +- .../lustre/lustre/include/lustre/lustre_user.h |3 +- drivers/staging/lustre/lustre/include/lustre_lmv.h | 25 +- drivers/staging/lustre/lustre/include/obd.h|4 +- drivers/staging/lustre/lustre/include/obd_class.h |5 +- drivers/staging/lustre/lustre/llite/dir.c | 31 ++- drivers/staging/lustre/lustre/llite/llite_lib.c| 89 ++-- drivers/staging/lustre/lustre/lmv/lmv_intent.c | 25 +- drivers/staging/lustre/lustre/lmv/lmv_internal.h |4 +- drivers/staging/lustre/lustre/lmv/lmv_obd.c| 70 drivers/staging/lustre/lustre/mdc/mdc_internal.h |4 +- drivers/staging/lustre/lustre/mdc/mdc_locks.c |2 +- drivers/staging/lustre/lustre/mdc/mdc_reint.c |6 +- drivers/staging/lustre/lustre/mdc/mdc_request.c|8 +- 14 files changed, 174 insertions(+), 166 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 3444add..8736826 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -2497,18 +2497,52 @@ struct lmv_desc { struct obd_uuid ld_uuid; }; -/* lmv structures */ -#define LMV_MAGIC_V1 0x0CD10CD0 /* normal stripe lmv magic */ -#define LMV_USER_MAGIC 0x0CD20CD0 /* default lmv magic*/ -#define LMV_MAGIC_MIGRATE 0x0CD30CD0 /* migrate stripe lmv magic */ -#define LMV_MAGIC LMV_MAGIC_V1 +/* LMV layout EA, and it will be stored both in master and slave object */ +struct lmv_mds_md_v1 { + __u32 lmv_magic; + __u32 lmv_stripe_count; + __u32 lmv_master_mdt_index; /* On master object, it is master +* MDT index, on slave object, it +* is stripe index of the slave obj +*/ + __u32 lmv_hash_type;/* dir stripe policy, i.e. indicate +* which hash function to be used, +* Note: only lower 16 bits is being +* used for now. Higher 16 bits will +* be used to mark the object status, +* for example migrating or dead. +*/ + __u32 lmv_layout_version; /* Used for directory restriping */ + __u32 lmv_padding; + struct lu_fid lmv_master_fid; /* The FID of the master object, which +* is the namespace-visible dir FID +*/ + char lmv_pool_name[LOV_MAXPOOLNAME];/* pool name */ + struct lu_fid lmv_stripe_fids[0]; /* FIDs for each stripe */ +}; +#define LMV_MAGIC_V10x0CD20CD0 /* normal stripe lmv magic */ +#define LMV_MAGIC LMV_MAGIC_V1 + +/* #define LMV_USER_MAGIC 0x0CD30CD0 */ +#define LMV_MAGIC_STRIPE 0x0CD40CD0/* magic for dir sub_stripe */ + +/* + *Right now only the lower part(0-16bits) of lmv_hash_type is being used, + * and the higher part will be the flag to indicate the status of object, + * for example the object is being migrated. And the hash function + * might be interpreted differently with different flags. + */ enum lmv_hash_type { LMV_HASH_TYPE_ALL_CHARS = 1, LMV_HASH_TYPE_FNV_1A_64 = 2, - LMV_HASH_TYPE_MIGRATION = 3, }; +#define LMV_HASH_TYPE_MASK 0x + +#define LMV_HASH_FLAG_MIGRATION0x8000 +#define LMV_HASH_FLAG_DEAD 0x4000 + #define LMV_HASH_NAME_ALL_CHARS"all_char" #define LMV_HASH_NAME_FNV_1A_64
[PATCH 49/80] staging: lustre: lmv: validate lock with correct stripe FID
From: wang diIn ll_lookup_it_finish, we need use the real parent(stripe) FID to validate the parent UPDATE lock. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4925 Reviewed-on: http://review.whamcloud.com/10026 Reviewed-by: John L. Hammond Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/include/obd.h |5 + drivers/staging/lustre/lustre/include/obd_class.h | 13 + drivers/staging/lustre/lustre/llite/namei.c | 15 +-- drivers/staging/lustre/lustre/lmv/lmv_obd.c | 19 ++- 4 files changed, 49 insertions(+), 3 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h index 52020a9..b7bdd07 100644 --- a/drivers/staging/lustre/lustre/include/obd.h +++ b/drivers/staging/lustre/lustre/include/obd.h @@ -1103,6 +1103,11 @@ struct md_ops { ldlm_policy_data_t *, enum ldlm_mode, enum ldlm_cancel_flags flags, void *opaque); + int (*get_fid_from_lsm)(struct obd_export *, + const struct lmv_stripe_md *, + const char *name, int namelen, + struct lu_fid *fid); + int (*intent_getattr_async)(struct obd_export *, struct md_enqueue_info *, struct ldlm_enqueue_info *); diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h index e86961c..69b628b 100644 --- a/drivers/staging/lustre/lustre/include/obd_class.h +++ b/drivers/staging/lustre/lustre/include/obd_class.h @@ -1699,6 +1699,19 @@ static inline int md_revalidate_lock(struct obd_export *exp, return rc; } +static inline int md_get_fid_from_lsm(struct obd_export *exp, + const struct lmv_stripe_md *lsm, + const char *name, int namelen, + struct lu_fid *fid) +{ + int rc; + + EXP_CHECK_MD_OP(exp, get_fid_from_lsm); + EXP_MD_COUNTER_INCREMENT(exp, get_fid_from_lsm); + rc = MDP(exp->exp_obd, get_fid_from_lsm)(exp, lsm, name, namelen, fid); + return rc; +} + /* OBD Metadata Support */ int obd_init_caches(void); diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c index 6e11b99..581b083 100644 --- a/drivers/staging/lustre/lustre/llite/namei.c +++ b/drivers/staging/lustre/lustre/llite/namei.c @@ -487,9 +487,20 @@ static int ll_lookup_it_finish(struct ptlrpc_request *request, struct lookup_intent parent_it = { .it_op = IT_GETATTR, .it_lock_handle = 0 }; + struct lu_fid fid = ll_i2info(parent)->lli_fid; + + /* If it is striped directory, get the real stripe parent */ + if (unlikely(ll_i2info(parent)->lli_lsm_md)) { + rc = md_get_fid_from_lsm(ll_i2mdexp(parent), +ll_i2info(parent)->lli_lsm_md, +(*de)->d_name.name, +(*de)->d_name.len, ); + if (rc) + return rc; + } - if (md_revalidate_lock(ll_i2mdexp(parent), _it, - _i2info(parent)->lli_fid, NULL)) { + if (md_revalidate_lock(ll_i2mdexp(parent), _it, , + NULL)) { d_lustre_revalidate(*de); ll_intent_release(_it); } diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index 03594f0..9821f69 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -2991,6 +2991,22 @@ static int lmv_revalidate_lock(struct obd_export *exp, struct lookup_intent *it, return rc; } +int lmv_get_fid_from_lsm(struct obd_export *exp, +const struct lmv_stripe_md *lsm, +const char *name, int namelen, struct lu_fid *fid) +{ + const struct lmv_oinfo *oinfo; + + LASSERT(lsm); + oinfo = lsm_name_to_stripe_info(lsm, name, namelen); + if (IS_ERR(oinfo)) + return PTR_ERR(oinfo); + + *fid = oinfo->lmo_fid; + + return 0; +} + /** * For lmv, only need to send request to master MDT, and the master MDT will * process with other slave MDTs. The only exception is Q_GETOQUOTA for
[PATCH 39/80] staging: lustre: llite: add error handler in inode prepare phase
From: wang diAdd error handler during inode inialization, so inode will become bad inode if something bad happens during inode prepare phase, otherwise the striped directory will not get its layout and being mis-regarded as normal directory. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4930 Reviewed-on: http://review.whamcloud.com/10170 Reviewed-by: Lai Siyao Reviewed-by: Fan Yong Reviewed-by: John L. Hammond Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../staging/lustre/lustre/llite/llite_internal.h |4 +- drivers/staging/lustre/lustre/llite/llite_lib.c| 66 +++ drivers/staging/lustre/lustre/llite/namei.c| 59 ++ 3 files changed, 72 insertions(+), 57 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h index 120aca3..e101dd8 100644 --- a/drivers/staging/lustre/lustre/llite/llite_internal.h +++ b/drivers/staging/lustre/lustre/llite/llite_internal.h @@ -766,8 +766,8 @@ int ll_setattr(struct dentry *de, struct iattr *attr); int ll_statfs(struct dentry *de, struct kstatfs *sfs); int ll_statfs_internal(struct super_block *sb, struct obd_statfs *osfs, __u64 max_age, __u32 flags); -void ll_update_inode(struct inode *inode, struct lustre_md *md); -void ll_read_inode2(struct inode *inode, void *opaque); +int ll_update_inode(struct inode *inode, struct lustre_md *md); +int ll_read_inode2(struct inode *inode, void *opaque); void ll_delete_inode(struct inode *inode); int ll_iocontrol(struct inode *inode, struct file *file, unsigned int cmd, unsigned long arg); diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c index 111264e..ea79ca3 100644 --- a/drivers/staging/lustre/lustre/llite/llite_lib.c +++ b/drivers/staging/lustre/lustre/llite/llite_lib.c @@ -464,7 +464,7 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt, md_free_lustre_md(sbi->ll_md_exp, ); ptlrpc_req_finished(request); - if (!(root)) { + if (IS_ERR(root)) { if (lmd.lsm) obd_free_memmd(sbi->ll_dt_exp, ); #ifdef CONFIG_FS_POSIX_ACL @@ -1109,11 +1109,11 @@ static inline int lli_lsm_md_eq(const struct lmv_stripe_md *lsm_md1, lsm_md2->lsm_md_pool_name); } -static void ll_update_lsm_md(struct inode *inode, struct lustre_md *md) +static int ll_update_lsm_md(struct inode *inode, struct lustre_md *md) { struct ll_inode_info *lli = ll_i2info(inode); struct lmv_stripe_md *lsm = md->lmv; - int idx; + int idx, rc; LASSERT(S_ISDIR(inode->i_mode)); CDEBUG(D_INODE, "update lsm %p of "DFID"\n", lli->lli_lsm_md, @@ -1122,7 +1122,7 @@ static void ll_update_lsm_md(struct inode *inode, struct lustre_md *md) /* no striped information from request. */ if (!lsm) { if (!lli->lli_lsm_md) { - return; + return 0; } else if (lli->lli_lsm_md->lsm_md_magic == LMV_MAGIC_MIGRATE) { /* * migration is done, the temporay MIGRATE layout has @@ -1132,27 +1132,22 @@ static void ll_update_lsm_md(struct inode *inode, struct lustre_md *md) PFID(ll_inode2fid(inode))); lmv_free_memmd(lli->lli_lsm_md); lli->lli_lsm_md = NULL; - return; + return 0; } else { /* * The lustre_md from req does not include stripeEA, * see ll_md_setattr */ - return; + return 0; } } /* set the directory layout */ if (!lli->lli_lsm_md) { - int rc; - rc = ll_init_lsm_md(inode, md); - if (rc) { - CERROR("%s: init "DFID" failed: rc = %d\n", - ll_get_fsname(inode->i_sb, NULL, 0), - PFID(>lli_fid), rc); - return; - } + if (rc) + return rc; + lli->lli_lsm_md = lsm; /* * set lsm_md to NULL, so the following free lustre_md @@ -1161,7 +1156,7 @@ static void ll_update_lsm_md(struct inode *inode, struct lustre_md *md) md->lmv = NULL; CDEBUG(D_INODE, "Set lsm %p magic %x to "DFID"\n", lsm, lsm->lsm_md_magic, PFID(ll_inode2fid(inode))); - return; +
Re: [PATCH 01/15 RESEND] staging: dgnc: remove redundant NULL checks in
2016-08-16 2:05 GMT+09:00 Greg KH: > On Wed, Jul 06, 2016 at 03:11:13PM +0900, Daeseok Youn wrote: >> The dgnc_block_til_ready() is only used in dgnc_tty_open(). >> The unit data(struct un_t) was stored into tty->driver_data in >> dgnc_tty_open(). >> And also tty and un were tested about NULL so these variables doesn't >> need to check for NULL in dgnc_block_til_ready(). >> >> Signed-off-by: Daeseok Youn >> --- >> RESEND: This patch was not merged for a long time, if there is any reason >> why this patch could NOT be merged into staging tree, let me know. >> There were no comment for this patch. >> I cannot understand why this patch have to wait long time to merge. >> And I also sent emails to mailing-lists for reminding this patch... >> please let me know, what is the problem to merge this patch into staging >> tree. > > Please note, staging patches are at the bottom of my priority queue. > Combined with a vacation, conferences, and a merge window and there are > a lot of pending staging patches in my to-review queue. That's Ok. but I took a long time to wait for merging my patches in this case. :-( > > thanks for your patience. > > greg k-h Thanks. Regards, Daeseok Youn. ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 13/80] staging: lustre: llite: fix "getdirstripe" to show stripe info
From: wang diFix "lfs getdirstripe", so it can show layout information of striped directory [root@testnode tests]# ../utils/lfs getdirstripe /mnt/lustre/test1 /mnt/lustre/test1 lmv_stripe_count: 2 lmv_stripe_offset: 0 mdtidx FID[seq:oid:ver] 0 [0x28400:0x1:0x0] 1 [0x2c400:0x1:0x0] Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-on: http://review.whamcloud.com/7228 Reviewed-by: Andreas Dilger Reviewed-by: John L. Hammond Reviewed-by: Alex Zhuravlev Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h |4 + .../lustre/lustre/include/lustre/lustre_user.h |1 + drivers/staging/lustre/lustre/llite/dir.c | 184 +++- .../staging/lustre/lustre/llite/llite_internal.h |4 +- drivers/staging/lustre/lustre/llite/xattr.c|7 +- .../staging/lustre/lustre/ptlrpc/pack_generic.c| 26 +++ 6 files changed, 180 insertions(+), 46 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index a612080..0ff30c6 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -1728,6 +1728,8 @@ lov_mds_md_max_stripe_count(size_t buf_size, __u32 lmm_magic) #define OBD_MD_FLDATAVERSION (0x0010ULL) /* iversion sum */ #define OBD_MD_FLRELEASED(0x0020ULL) /* file released */ +#define OBD_MD_DEFAULT_MEA (0x0040ULL) /* default MEA */ + #define OBD_MD_FLGETATTR (OBD_MD_FLID| OBD_MD_FLATIME | OBD_MD_FLMTIME | \ OBD_MD_FLCTIME | OBD_MD_FLSIZE | OBD_MD_FLBLKSZ | \ OBD_MD_FLMODE | OBD_MD_FLTYPE | OBD_MD_FLUID | \ @@ -2543,6 +2545,8 @@ union lmv_mds_md { struct lmv_user_md lmv_user_md; }; +void lustre_swab_lmv_mds_md(union lmv_mds_md *lmm); + static inline ssize_t lmv_mds_md_size(int stripe_count, unsigned int lmm_magic) { ssize_t len = -EINVAL; diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h index d496d0e..26dbda0 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h @@ -242,6 +242,7 @@ struct ost_id { #define LL_IOC_SET_LEASE _IOWR('f', 243, long) #define LL_IOC_GET_LEASE _IO('f', 244) #define LL_IOC_HSM_IMPORT _IOWR('f', 245, struct hsm_user_import) +#define LL_IOC_LMV_SET_DEFAULT_STRIPE _IOWR('f', 246, struct lmv_user_md) #define LL_STATFS_LMV 1 #define LL_STATFS_LOV 2 diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index a0560b6..5288750 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -749,6 +749,13 @@ int ll_dir_setstripe(struct inode *inode, struct lov_user_md *lump, lum_size = sizeof(struct lov_user_md_v3); break; } + case LMV_USER_MAGIC: { + if (lump->lmm_magic != cpu_to_le32(LMV_USER_MAGIC)) + lustre_swab_lmv_user_md( + (struct lmv_user_md *)lump); + lum_size = sizeof(struct lmv_user_md); + break; + } default: { CDEBUG(D_IOCTL, "bad userland LOV MAGIC: %#08x != %#08x nor %#08x\n", lump->lmm_magic, LOV_USER_MAGIC_V1, @@ -819,8 +826,16 @@ end: return rc; } -int ll_dir_getstripe(struct inode *inode, struct lov_mds_md **lmmp, -int *lmm_size, struct ptlrpc_request **request) +/** + * This function will be used to get default LOV/LMV/Default LMV + * @valid will be used to indicate which stripe it will retrieve + * OBD_MD_MEA LMV stripe EA + * OBD_MD_DEFAULT_MEA Default LMV stripe EA + * otherwise Default LOV EA. + * Each time, it can only retrieve 1 stripe EA + **/ +int ll_dir_getstripe(struct inode *inode, void **plmm, int *plmm_size, +struct ptlrpc_request **request, u64 valid) { struct ll_sb_info *sbi = ll_i2sbi(inode); struct mdt_body *body; @@ -829,7 +844,7 @@ int ll_dir_getstripe(struct inode *inode, struct lov_mds_md **lmmp, int rc, lmmsize; struct md_op_data *op_data; - rc = ll_get_default_mdsize(sbi, ); + rc = ll_get_max_mdsize(sbi, ); if (rc) return rc; @@
[PATCH 22/80] staging: lustre: llite: a few fixes for migration.
From: wang di1. Clear the client dentry cache before migrating file/directory to the remote MDT. 2. Do not return stripe information to client, if it did not get the layout lock. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4682 Reviewed-on: http://review.whamcloud.com/9522 Reviewed-by: Andreas Dilger Reviewed-by: John L. Hammond Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/dir.c | 22 +++- drivers/staging/lustre/lustre/llite/file.c | 34 +++- .../staging/lustre/lustre/llite/llite_internal.h |2 + drivers/staging/lustre/lustre/lov/lov_object.c |1 + 4 files changed, 28 insertions(+), 31 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index ef7322e..84bec03 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -1318,11 +1318,9 @@ static long ll_dir_ioctl(struct file *file, unsigned int cmd, unsigned long arg) return 0; } case IOC_MDC_LOOKUP: { - struct ptlrpc_request *request = NULL; int namelen, len = 0; char *buf = NULL; char *filename; - struct md_op_data *op_data; rc = obd_ioctl_getdata(, , (void __user *)arg); if (rc) @@ -1338,21 +1336,13 @@ static long ll_dir_ioctl(struct file *file, unsigned int cmd, unsigned long arg) goto out_free; } - op_data = ll_prep_md_op_data(NULL, inode, NULL, filename, namelen, -0, LUSTRE_OPC_ANY, NULL); - if (IS_ERR(op_data)) { - rc = PTR_ERR(op_data); - goto out_free; - } - - op_data->op_valid = OBD_MD_FLID; - rc = md_getattr_name(sbi->ll_md_exp, op_data, ); - ll_finish_md_op_data(op_data); + rc = ll_get_fid_by_name(inode, filename, namelen, NULL); if (rc < 0) { - CDEBUG(D_INFO, "md_getattr_name: %d\n", rc); + CERROR("%s: lookup %.*s failed: rc = %d\n", + ll_get_fsname(inode->i_sb, NULL, 0), namelen, + filename, rc); goto out_free; } - ptlrpc_req_finished(request); out_free: obd_ioctl_freedata(buf, len); return rc; @@ -1981,7 +1971,7 @@ out_quotactl: filename = data->ioc_inlbuf1; namelen = data->ioc_inllen1; - if (namelen < 1) { + if (namelen < 1 || namelen != strlen(filename) + 1) { rc = -EINVAL; goto migrate_free; } @@ -1992,7 +1982,7 @@ out_quotactl: } mdtidx = *(int *)data->ioc_inlbuf2; - rc = ll_migrate(inode, file, mdtidx, filename, namelen); + rc = ll_migrate(inode, file, mdtidx, filename, namelen - 1); migrate_free: obd_ioctl_freedata(buf, len); diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c index 8d98db6..769b028 100644 --- a/drivers/staging/lustre/lustre/llite/file.c +++ b/drivers/staging/lustre/lustre/llite/file.c @@ -2828,8 +2828,8 @@ ll_file_flock(struct file *file, int cmd, struct file_lock *file_lock) return rc; } -static int ll_get_fid_by_name(struct inode *parent, const char *name, - int namelen, struct lu_fid *fid) +int ll_get_fid_by_name(struct inode *parent, const char *name, + int namelen, struct lu_fid *fid) { struct md_op_data *op_data = NULL; struct ptlrpc_request *req; @@ -2843,20 +2843,19 @@ static int ll_get_fid_by_name(struct inode *parent, const char *name, op_data->op_valid = OBD_MD_FLID; rc = md_getattr_name(ll_i2sbi(parent)->ll_md_exp, op_data, ); + ll_finish_md_op_data(op_data); if (rc < 0) - goto out_free; + return rc; body = req_capsule_server_get(>rq_pill, _MDT_BODY); if (!body) { rc = -EFAULT; goto out_req; } - *fid = body->fid1; + if (fid) + *fid = body->fid1; out_req: ptlrpc_req_finished(req); -out_free: - if (op_data) - ll_finish_md_op_data(op_data); return rc; } @@ -2864,12 +2863,13 @@ int ll_migrate(struct inode *parent, struct file *file, int mdtidx, const char *name, int namelen) { struct ptlrpc_request *request = NULL; + struct inode *child_inode = NULL;
[PATCH 27/80] staging: lustre: simplify inline functions in lustre_fid.h
From: John L. HammondSeveral inline functions return a structure that was passed in. Their is no need for this so just make these function void. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675 Reviewed-on: http://review.whamcloud.com/8641 Reviewed-by: wangdi Reviewed-by: Fan Yong Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/include/lustre_fid.h | 16 1 files changed, 4 insertions(+), 12 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre_fid.h b/drivers/staging/lustre/lustre/include/lustre_fid.h index 6f7dc15..f1d5bbd 100644 --- a/drivers/staging/lustre/lustre/include/lustre_fid.h +++ b/drivers/staging/lustre/lustre/include/lustre_fid.h @@ -393,14 +393,12 @@ struct ldlm_namespace; * but was moved into name[1] along with the OID to avoid consuming the * renaming name[2,3] fields that need to be used for the quota identifier. */ -static inline struct ldlm_res_id * +static inline void fid_build_reg_res_name(const struct lu_fid *fid, struct ldlm_res_id *res) { memset(res, 0, sizeof(*res)); res->name[LUSTRE_RES_ID_SEQ_OFF] = fid_seq(fid); res->name[LUSTRE_RES_ID_VER_OID_OFF] = fid_ver_oid(fid); - - return res; } /* @@ -416,29 +414,25 @@ static inline bool fid_res_name_eq(const struct lu_fid *fid, /* * Extract FID from LDLM resource. Reverse of fid_build_reg_res_name(). */ -static inline struct lu_fid * +static inline void fid_extract_from_res_name(struct lu_fid *fid, const struct ldlm_res_id *res) { fid->f_seq = res->name[LUSTRE_RES_ID_SEQ_OFF]; fid->f_oid = (__u32)(res->name[LUSTRE_RES_ID_VER_OID_OFF]); fid->f_ver = (__u32)(res->name[LUSTRE_RES_ID_VER_OID_OFF] >> 32); LASSERT(fid_res_name_eq(fid, res)); - - return fid; } /* * Build (DLM) resource identifier from global quota FID and quota ID. */ -static inline struct ldlm_res_id * +static inline void fid_build_quota_res_name(const struct lu_fid *glb_fid, union lquota_id *qid, struct ldlm_res_id *res) { fid_build_reg_res_name(glb_fid, res); res->name[LUSTRE_RES_ID_QUOTA_SEQ_OFF] = fid_seq(>qid_fid); res->name[LUSTRE_RES_ID_QUOTA_VER_OID_OFF] = fid_ver_oid(>qid_fid); - - return res; } /* @@ -455,14 +449,12 @@ static inline void fid_extract_from_quota_res(struct lu_fid *glb_fid, (__u32)(res->name[LUSTRE_RES_ID_QUOTA_VER_OID_OFF] >> 32); } -static inline struct ldlm_res_id * +static inline void fid_build_pdo_res_name(const struct lu_fid *fid, unsigned int hash, struct ldlm_res_id *res) { fid_build_reg_res_name(fid, res); res->name[LUSTRE_RES_ID_HSH_OFF] = hash; - - return res; } /** -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 23/80] staging: lustre: mdc: fixup MDS_SWAP_LAYOUTS ELC handling
From: John L. HammondIn mdc_ioc_swap_layouts() cancel *any* unused locks with LAYOUT or XATTR IBITS set on the two files. (This matches the locks acquired in mdt_swap_layouts(). Previously only locks that conflicted with a CR LAYOUT lock were cancelled.) Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4320 Reviewed-on: http://review.whamcloud.com/9329 Reviewed-by: Andreas Dilger Reviewed-by: Faccini Bruno Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/mdc/mdc_request.c |6 -- 1 files changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c index 702ced9..030295f 100644 --- a/drivers/staging/lustre/lustre/mdc/mdc_request.c +++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c @@ -1670,9 +1670,11 @@ static int mdc_ioc_swap_layouts(struct obd_export *exp, * with the request RPC to avoid extra RPC round trips */ count = mdc_resource_get_unused(exp, _data->op_fid1, , - LCK_CR, MDS_INODELOCK_LAYOUT); + LCK_CR, MDS_INODELOCK_LAYOUT | + MDS_INODELOCK_XATTR); count += mdc_resource_get_unused(exp, _data->op_fid2, , -LCK_CR, MDS_INODELOCK_LAYOUT); +LCK_CR, MDS_INODELOCK_LAYOUT | +MDS_INODELOCK_XATTR); req = ptlrpc_request_alloc(class_exp2cliimp(exp), _MDS_SWAP_LAYOUTS); -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 10/80] staging: lustre: lmv: remove lmv_get_easize
From: wang diCompletely replace lmv_get_easize with lmv_mds_md_size. With this change we can delete lmv_get_easize. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_internal.h |7 --- drivers/staging/lustre/lustre/lmv/lmv_obd.c |2 +- 2 files changed, 1 insertions(+), 8 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_internal.h b/drivers/staging/lustre/lustre/lmv/lmv_internal.h index 90a9786..f4c917b 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_internal.h +++ b/drivers/staging/lustre/lustre/lmv/lmv_internal.h @@ -55,13 +55,6 @@ int __lmv_fid_alloc(struct lmv_obd *lmv, struct lu_fid *fid, u32 mds); int lmv_fid_alloc(struct obd_export *exp, struct lu_fid *fid, struct md_op_data *op_data); -static inline int lmv_get_easize(struct lmv_obd *lmv) -{ - return sizeof(struct lmv_stripe_md) + - lmv->desc.ld_tgt_count * - sizeof(struct lu_fid); -} - static inline struct lmv_tgt_desc * lmv_get_target(struct lmv_obd *lmv, u32 mds) { diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index 1ba5900..0b1260d 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -558,7 +558,7 @@ int lmv_check_connect(struct obd_device *obd) lmv_set_timeouts(obd); class_export_put(lmv->exp); lmv->connected = 1; - easize = lmv_get_easize(lmv); + easize = lmv_mds_md_size(lmv->desc.ld_tgt_count, LMV_MAGIC); lmv_init_ea_size(obd->obd_self_export, easize, 0, 0, 0); mutex_unlock(>lmv_init_mutex); return 0; -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 08/80] staging: lustre: lmv: remove duplicate MAX_HASH_*
From: wang diThe MAX_HASH_* macros already exist in obd.h. Remove the duplicated defines in lustre_idl.h. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h |4 1 files changed, 0 insertions(+), 4 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 32471a6..5f31724 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -2486,10 +2486,6 @@ struct lmv_desc { #define MEA_MAGIC_ALL_CHARS 0xb222a11c #define MEA_MAGIC_HASH_SEGMENT 0xb222a11b -#define MAX_HASH_SIZE_320x7fffUL -#define MAX_HASH_SIZE 0x7fffULL -#define MAX_HASH_HIGHEST_BIT 0x1000ULL - /* lmv structures */ #define LMV_MAGIC_V1 0x0CD10CD0 /* normal stripe lmv magic */ #define LMV_USER_MAGIC 0x0CD20CD0 /* default lmv magic*/ -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 00/80] staging: lustre: majority of missing fixes for 2.6 release
This is the combination of two previous patch sets that cover missing fixes for lustre version 2.5.56 to the 2.6 release minus a few patches that have some bug regressions. Also in this update are the latest LNet fixes. Alexander Boyko (1): staging: lustre: lnet: make connection more stable with packet loss Andriy Skulysh (1): staging: lustre: ptlrpc: request gets stuck in UNREGISTERING phase Brian Behlendorf (1): staging: lustre: obd: limit lu_object cache Chris Horn (1): staging: lustre: ptlrpc: Early replies need to honor at_max Christopher J. Morrone (1): staging: lustre: Remove static declaration in anonymous union Dmitry Eremin (2): staging: lustre: lmv: fix issue found by Klocwork Insight tool staging: lustre: lmv: build error with gcc 4.7.0 20110509 Doug Oucharek (3): staging: lustre: lnet: Do not drop message when shutting down LNet staging: lustre: lnet: Correct position of lnet_ni_decref() staging: lustre: lnet: Stop Infinite CON RACE Condition Emoly Liu (1): staging: lustre: ldlm: improve ldlm_lock_create() return value Fan Yong (6): staging: lustre: obdclass: bug fixes for lu_device_type handling staging: lustre: llite: enable clients to inject error for lfsck staging: lustre: obdclass: unified flow control interfaces staging: lustre: reorder LOV_MAGIC_* definition staging: lustre: lov: new pattern flag for partially repaired file staging: lustre: lmv: build master LMV EA dynamically build via readdir Gregoire Pichon (1): staging: lustre: llite: fix inconsistencies of root squash feature Hongchao Zhang (2): staging: lustre: llite: set dir LOV xattr length variable staging: lustre: osc: Automatically increase the max_dirty_mb James Simmons (3): staging: lustre: obdclass: compile issues with variable not being initialized staging: lustre: include: fix one off errors in lustre_id.h staging: lustre: llite: remove assert for acl refcount Jian Yu (1): staging: lustre: obdclass: fix lmd_parse() to handle comma-separated NIDs Jinshan Xiong (7): staging: lustre: osc: allow to call brw_commit() multiple times staging: lustre: llite: avoid a deadlock in page write staging: lustre: lov: handle the case of stripe size is not power 2 staging: lustre: llite: Fix the deadlock in balance_dirty_pages() staging: lustre: llite: Change readdir BRW metrics staging: lustre: clio: Reduce memory overhead of per-page allocation staging: lustre: osc: revise unstable pages accounting John L. Hammond (13): staging: lustre: mdc: fixup MDS_SWAP_LAYOUTS ELC handling staging: lustre: don't need to const __u64 parameters for lustre_idl.h staging: lustre: const correct FID/OSTID/... helpers staging: lustre: use bool for several function in lustre_idl.h/lustre_fid.h staging: lustre: simplify inline functions in lustre_fid.h staging: lustre: lmv: access lum_stripe_offset as little endian staging: lustre: lmv: cleanup req in lmv_getattr_name() staging: lustre: lmv: rename request to preq in lmv_getattr_name() staging: lustre: move ioctls to lustre_ioctl.h staging: lustre: llite: validate names staging: lustre: uapi: reduce scope of lustre_idl.h staging: lustre: mdt: add mbo_ prefix to members of struct mdt_body staging: lustre: obd: validate open handle cookies Lai Siyao (2): staging: lustre: fid: do open-by-fid by default staging: lustre: ptlrpc: add OBD_CONNECT_UNLINK_CLOSE flag Liang Zhen (1): staging: lustre: lnet: lock improvement for ko2iblnd Mikhail Pershin (1): staging: lustre: llog: keep llog ctxt indices constant Nathaniel Clark (1): staging: lustre: lmv: Ensure lmv_intent_lookup cleans up reqp Niu Yawei (1): staging: lustre: obd: rename lsr_padding to lsr_valid Patrick Farrell (1): staging: lustre: fld: add fld description documentation Ryan Haasken (1): staging: lustre: libcfs: Only dump log once per sec. to avoid EEXIST Vitaly Fertman (1): staging: lustre: ldlm: flock completion fixes. wang di (27): staging: lustre: llite: add md_op_data parameter to ll_get_dir_page staging: lustre: llite: remove comment from ll_dir_read staging: lustre: llite: style cleanup for llite_internal.h staging: lustre: llite: pass inode to ll_release_page staging: lustre: llite: change remove parameter to bool staging: lustre: mdc: don't take rpc lock for readdir case staging: lustre: lmv: remove unused lmv_get_mea function staging: lustre: lmv: remove duplicate MAX_HASH_* staging: lustre: lmv: change handling of lmv striping information staging: lustre: lmv: remove lmv_get_easize staging: lustre: lmv: replace obd_free_memmd with lmv_free_memmd staging: lustre: create striped directory staging: lustre: llite: fix "getdirstripe" to show stripe info staging: lustre: delete striped directory staging: lustre: add ability to migrate inodes. staging: lustre: llite: a few fixes for migration. staging: lustre: lmv: lookup remote migrating object in LMV staging: lustre: llite: add error
[PATCH 04/80] staging: lustre: llite: pass inode to ll_release_page
From: wang diAdd a inode parameter to ll_release_page. This will be used in the future. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/dir.c | 17 + .../staging/lustre/lustre/llite/llite_internal.h |2 +- drivers/staging/lustre/lustre/llite/statahead.c| 20 +++- 3 files changed, 21 insertions(+), 18 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index d854edd..3a800b2 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -236,7 +236,7 @@ static int ll_dir_filler(void *_hash, struct page *page0) return rc; } -void ll_release_page(struct page *page, int remove) +void ll_release_page(struct inode *inode, struct page *page, int remove) { kunmap(page); if (remove) { @@ -297,7 +297,7 @@ static struct page *ll_dir_page_locate(struct inode *dir, __u64 *hash, CDEBUG(D_VFSTRACE, "page %lu [%llu %llu], hash %llu\n", offset, *start, *end, *hash); if (*hash > *end) { - ll_release_page(page, 0); + ll_release_page(dir, page, 0); page = NULL; } else if (*end != *start && *hash == *end) { /* @@ -306,8 +306,9 @@ static struct page *ll_dir_page_locate(struct inode *dir, __u64 *hash, * ll_get_dir_page() will issue RPC to fetch * the page we want. */ - ll_release_page(page, - le32_to_cpu(dp->ldp_flags) & LDF_COLLIDE); + ll_release_page(dir, page, + le32_to_cpu(dp->ldp_flags) & + LDF_COLLIDE); page = NULL; } } else { @@ -462,7 +463,7 @@ out_unlock: return page; fail: - ll_release_page(page, 1); + ll_release_page(dir, page, 1); page = ERR_PTR(-EIO); goto out_unlock; } @@ -560,7 +561,7 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, if (done) { pos = hash; - ll_release_page(page, 0); + ll_release_page(inode, page, 0); break; } @@ -571,13 +572,13 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, * End of directory reached. */ done = 1; - ll_release_page(page, 0); + ll_release_page(inode, page, 0); } else { /* * Normal case: continue to the next * page. */ - ll_release_page(page, + ll_release_page(inode, page, le32_to_cpu(dp->ldp_flags) & LDF_COLLIDE); next = pos; diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h index 1ced397..4b03a64 100644 --- a/drivers/staging/lustre/lustre/llite/llite_internal.h +++ b/drivers/staging/lustre/lustre/llite/llite_internal.h @@ -655,7 +655,7 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, int ll_get_mdt_idx(struct inode *inode); struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data, __u64 hash, struct ll_dir_chain *chain); -void ll_release_page(struct page *page, int remove); +void ll_release_page(struct inode *inode, struct page *page, int remove); /* llite/namei.c */ extern const struct inode_operations ll_special_inode_operations; diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c index 1b222c7..2949ff6 100644 --- a/drivers/staging/lustre/lustre/llite/statahead.c +++ b/drivers/staging/lustre/lustre/llite/statahead.c @@ -1142,7 +1142,7 @@ interpret_it: ll_post_statahead(sai); if (unlikely(!thread_is_running(thread))) { -
[PATCH 06/80] staging: lustre: mdc: don't take rpc lock for readdir case
From: wang diIf the operation is IT_READDIR don't need to handle the mdc RPC lock. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/10761 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4906 Reviewed-by: John L. Hammond Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/include/lustre_mdc.h |4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre_mdc.h b/drivers/staging/lustre/lustre/include/lustre_mdc.h index fa62b95..0a8c639 100644 --- a/drivers/staging/lustre/lustre/include/lustre_mdc.h +++ b/drivers/staging/lustre/lustre/include/lustre_mdc.h @@ -96,7 +96,7 @@ static inline void mdc_get_rpc_lock(struct mdc_rpc_lock *lck, struct lookup_intent *it) { if (it && (it->it_op == IT_GETATTR || it->it_op == IT_LOOKUP || - it->it_op == IT_LAYOUT)) + it->it_op == IT_LAYOUT || it->it_op == IT_READDIR)) return; /* This would normally block until the existing request finishes. @@ -136,7 +136,7 @@ static inline void mdc_put_rpc_lock(struct mdc_rpc_lock *lck, struct lookup_intent *it) { if (it && (it->it_op == IT_GETATTR || it->it_op == IT_LOOKUP || - it->it_op == IT_LAYOUT)) + it->it_op == IT_LAYOUT || it->it_op == IT_READDIR)) return; if (lck->rpcl_it == MDC_FAKE_RPCL_IT) { /* OBD_FAIL_MDC_RPCS_SEM */ -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 30/80] staging: lustre: lmv: Ensure lmv_intent_lookup cleans up reqp
From: Nathaniel ClarkEnsure there aren't invalid pointers hanging around after ptlrpc_req_finished is called. Signed-off-by: Nathaniel Clark Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4826 Reviewed-on: http://review.whamcloud.com/9841 Reviewed-by: John L. Hammond Reviewed-by: Bob Glossman Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_intent.c |3 +++ 1 files changed, 3 insertions(+), 0 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c index a38d343..d7e165f 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c @@ -464,6 +464,9 @@ static int lmv_intent_lookup(struct obd_export *exp, return PTR_ERR(tgt); ptlrpc_req_finished(*reqp); + it->it_request = NULL; + *reqp = NULL; + CDEBUG(D_INODE, "For migrating dir, try target dir "DFID"\n", PFID(>lsm_md_oinfo[1].lmo_fid)); -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 09/80] staging: lustre: lmv: change handling of lmv striping information
From: wang diThe lmv_[un]pack_md function are used to calculate the size of the data used to represent the LMV striping data. The original code was straight forward in its calculate with lmv_get_easize since only one type of data format could exist. We want to be able to support different version of this data in the future so this patch moves to generating the size of the data using the stripe count and which LMV_MAGIC_* version. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h |4 - drivers/staging/lustre/lustre/include/lustre_lmv.h | 15 +- drivers/staging/lustre/lustre/lmv/lmv_internal.h |7 + drivers/staging/lustre/lustre/lmv/lmv_obd.c| 245 +++- 4 files changed, 198 insertions(+), 73 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 5f31724..0ad6605 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -2482,10 +2482,6 @@ struct lmv_desc { struct obd_uuid ld_uuid; }; -#define MEA_MAGIC_LAST_CHAR 0xb2221ca1 -#define MEA_MAGIC_ALL_CHARS 0xb222a11c -#define MEA_MAGIC_HASH_SEGMENT 0xb222a11b - /* lmv structures */ #define LMV_MAGIC_V1 0x0CD10CD0 /* normal stripe lmv magic */ #define LMV_USER_MAGIC 0x0CD20CD0 /* default lmv magic*/ diff --git a/drivers/staging/lustre/lustre/include/lustre_lmv.h b/drivers/staging/lustre/lustre/include/lustre_lmv.h index 0620c8c..784d67b 100644 --- a/drivers/staging/lustre/lustre/include/lustre_lmv.h +++ b/drivers/staging/lustre/lustre/include/lustre_lmv.h @@ -41,12 +41,15 @@ struct lmv_oinfo { }; struct lmv_stripe_md { - __u32 mea_magic; - __u32 mea_count; - __u32 mea_master; - __u32 mea_padding; - charmea_pool_name[LOV_MAXPOOLNAME]; - struct lu_fid mea_ids[0]; + __u32 lsm_md_magic; + __u32 lsm_md_stripe_count; + __u32 lsm_md_master_mdt_index; + __u32 lsm_md_hash_type; + __u32 lsm_md_layout_version; + __u32 lsm_md_default_count; + __u32 lsm_md_default_index; + charlsm_md_pool_name[LOV_MAXPOOLNAME]; + struct lmv_oinfo lsm_md_oinfo[0]; }; union lmv_mds_md; diff --git a/drivers/staging/lustre/lustre/lmv/lmv_internal.h b/drivers/staging/lustre/lustre/lmv/lmv_internal.h index ab01560..90a9786 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_internal.h +++ b/drivers/staging/lustre/lustre/lmv/lmv_internal.h @@ -94,6 +94,13 @@ lmv_find_target(struct lmv_obd *lmv, const struct lu_fid *fid) return lmv_get_target(lmv, mds); } +static inline int lmv_stripe_md_size(int stripe_count) +{ + struct lmv_stripe_md *lsm; + + return sizeof(*lsm) + stripe_count * sizeof(lsm->lsm_md_oinfo[0]); +} + struct lmv_tgt_desc *lmv_locate_mds(struct lmv_obd *lmv, struct md_op_data *op_data, struct lu_fid *fid); diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index 8e83263..1ba5900 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -2376,105 +2376,224 @@ static int lmv_set_info_async(const struct lu_env *env, struct obd_export *exp, return -EINVAL; } -static int lmv_packmd(struct obd_export *exp, struct lov_mds_md **lmmp, - struct lov_stripe_md *lsm) +static int lmv_pack_md_v1(const struct lmv_stripe_md *lsm, + struct lmv_mds_md_v1 *lmm1) { - struct obd_device*obd = class_exp2obd(exp); - struct lmv_obd *lmv = >u.lmv; - struct lmv_stripe_md *meap; - struct lmv_stripe_md *lsmp; - int mea_size; - int i; + int cplen; + int i; + + lmm1->lmv_magic = cpu_to_le32(lsm->lsm_md_magic); + lmm1->lmv_stripe_count = cpu_to_le32(lsm->lsm_md_stripe_count); + lmm1->lmv_master_mdt_index = cpu_to_le32(lsm->lsm_md_master_mdt_index); + lmm1->lmv_hash_type = cpu_to_le32(lsm->lsm_md_hash_type); + cplen = strlcpy(lmm1->lmv_pool_name, lsm->lsm_md_pool_name, + sizeof(lmm1->lmv_pool_name)); + if (cplen >= sizeof(lmm1->lmv_pool_name)) + return -E2BIG; + + for (i = 0; i < lsm->lsm_md_stripe_count; i++) + fid_cpu_to_le(>lmv_stripe_fids[i], + >lsm_md_oinfo[i].lmo_fid); +
[PATCH 32/80] staging: lustre: lov: handle the case of stripe size is not power 2
From: Jinshan XiongCalculate the end of current stripe correctly when the stripe size is not power 2. Signed-off-by: Jinshan Xiong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4860 Reviewed-on: http://review.whamcloud.com/9882 Reviewed-by: Bobi Jam Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lov/lov_page.c | 11 +-- 1 files changed, 9 insertions(+), 2 deletions(-) diff --git a/drivers/staging/lustre/lustre/lov/lov_page.c b/drivers/staging/lustre/lustre/lov/lov_page.c index c17026f..45b5ae9 100644 --- a/drivers/staging/lustre/lustre/lov/lov_page.c +++ b/drivers/staging/lustre/lustre/lov/lov_page.c @@ -65,7 +65,9 @@ static int lov_raid0_page_is_under_lock(const struct lu_env *env, pgoff_t index = *max_index; unsigned int pps; /* pages per stripe */ - CDEBUG(D_READA, "*max_index = %lu, nr = %d\n", index, r0->lo_nr); + CDEBUG(D_READA, DFID "*max_index = %lu, nr = %d\n", + PFID(lu_object_fid(lov2lu(loo))), index, r0->lo_nr); + if (index == 0) /* the page is not covered by any lock */ return 0; @@ -80,7 +82,12 @@ static int lov_raid0_page_is_under_lock(const struct lu_env *env, /* calculate the end of current stripe */ pps = loo->lo_lsm->lsm_stripe_size >> PAGE_SHIFT; - index = ((slice->cpl_index + pps) & ~(pps - 1)) - 1; + index = slice->cpl_index + pps - slice->cpl_index % pps - 1; + + CDEBUG(D_READA, DFID "*max_index = %lu, index = %lu, pps = %u, stripe_size = %u, stripe no = %u, page index = %lu\n", + PFID(lu_object_fid(lov2lu(loo))), *max_index, index, pps, + loo->lo_lsm->lsm_stripe_size, lov_page_stripe(slice->cpl_page), + slice->cpl_index); /* never exceed the end of the stripe */ *max_index = min_t(pgoff_t, *max_index, index); -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 31/80] staging: lustre: llite: avoid a deadlock in page write
From: Jinshan XiongFor a partial page write, it will have to issue a READ RPC firstly to get a full uptodate page. If another page is already locked by this thread it can easily cause deadlock. Signed-off-by: Jinshan Xiong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4873 Reviewed-on: http://review.whamcloud.com/9928 Reviewed-by: Bobi Jam Reviewed-by: Niu Yawei Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/rw26.c |5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c index d98c7ac..c14a1b6 100644 --- a/drivers/staging/lustre/lustre/llite/rw26.c +++ b/drivers/staging/lustre/lustre/llite/rw26.c @@ -506,8 +506,9 @@ static int ll_write_begin(struct file *file, struct address_space *mapping, env = lcc->lcc_env; io = lcc->lcc_io; - /* To avoid deadlock, try to lock page first. */ - vmpage = grab_cache_page_nowait(mapping, index); + if (likely(to == PAGE_SIZE)) /* LU-4873 */ + /* To avoid deadlock, try to lock page first. */ + vmpage = grab_cache_page_nowait(mapping, index); if (unlikely(!vmpage || PageDirty(vmpage) || PageWriteback(vmpage))) { struct vvp_io *vio = vvp_env_io(env); struct cl_page_list *plist = >u.write.vui_queue; -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 26/80] staging: lustre: use bool for several function in lustre_idl.h/lustre_fid.h
From: John L. HammondChange the return type of several predicate functions from int to bool. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675 Reviewed-on: http://review.whamcloud.com/8641 Reviewed-by: wangdi Reviewed-by: Fan Yong Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h | 72 ++-- drivers/staging/lustre/lustre/include/lustre_fid.h |4 +- 2 files changed, 38 insertions(+), 38 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index c932e20..d3a9db9 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -196,12 +196,12 @@ static inline unsigned fld_range_type(const struct lu_seq_range *range) return range->lsr_flags & LU_SEQ_RANGE_MASK; } -static inline int fld_range_is_ost(const struct lu_seq_range *range) +static inline bool fld_range_is_ost(const struct lu_seq_range *range) { return fld_range_type(range) == LU_SEQ_RANGE_OST; } -static inline int fld_range_is_mdt(const struct lu_seq_range *range) +static inline bool fld_range_is_mdt(const struct lu_seq_range *range) { return fld_range_type(range) == LU_SEQ_RANGE_MDT; } @@ -260,23 +260,23 @@ static inline void range_init(struct lu_seq_range *range) * check if given seq id \a s is within given range \a r */ -static inline int range_within(const struct lu_seq_range *range, - __u64 s) +static inline bool range_within(const struct lu_seq_range *range, + __u64 s) { return s >= range->lsr_start && s < range->lsr_end; } -static inline int range_is_sane(const struct lu_seq_range *range) +static inline bool range_is_sane(const struct lu_seq_range *range) { return (range->lsr_end >= range->lsr_start); } -static inline int range_is_zero(const struct lu_seq_range *range) +static inline bool range_is_zero(const struct lu_seq_range *range) { return (range->lsr_start == 0 && range->lsr_end == 0); } -static inline int range_is_exhausted(const struct lu_seq_range *range) +static inline bool range_is_exhausted(const struct lu_seq_range *range) { return range_space(range) == 0; @@ -437,69 +437,69 @@ enum dot_lustre_oid { FID_OID_DOT_LUSTRE_OBF = 2UL, }; -static inline int fid_seq_is_mdt0(__u64 seq) +static inline bool fid_seq_is_mdt0(__u64 seq) { return (seq == FID_SEQ_OST_MDT0); } -static inline int fid_seq_is_mdt(__u64 seq) +static inline bool fid_seq_is_mdt(__u64 seq) { return seq == FID_SEQ_OST_MDT0 || seq >= FID_SEQ_NORMAL; }; -static inline int fid_seq_is_echo(__u64 seq) +static inline bool fid_seq_is_echo(__u64 seq) { return (seq == FID_SEQ_ECHO); } -static inline int fid_is_echo(const struct lu_fid *fid) +static inline bool fid_is_echo(const struct lu_fid *fid) { return fid_seq_is_echo(fid_seq(fid)); } -static inline int fid_seq_is_llog(__u64 seq) +static inline bool fid_seq_is_llog(__u64 seq) { return (seq == FID_SEQ_LLOG); } -static inline int fid_is_llog(const struct lu_fid *fid) +static inline bool fid_is_llog(const struct lu_fid *fid) { /* file with OID == 0 is not llog but contains last oid */ return fid_seq_is_llog(fid_seq(fid)) && fid_oid(fid) > 0; } -static inline int fid_seq_is_rsvd(__u64 seq) +static inline bool fid_seq_is_rsvd(__u64 seq) { return (seq > FID_SEQ_OST_MDT0 && seq <= FID_SEQ_RSVD); }; -static inline int fid_seq_is_special(__u64 seq) +static inline bool fid_seq_is_special(__u64 seq) { return seq == FID_SEQ_SPECIAL; }; -static inline int fid_seq_is_local_file(__u64 seq) +static inline bool fid_seq_is_local_file(__u64 seq) { return seq == FID_SEQ_LOCAL_FILE || seq == FID_SEQ_LOCAL_NAME; }; -static inline int fid_seq_is_root(__u64 seq) +static inline bool fid_seq_is_root(__u64 seq) { return seq == FID_SEQ_ROOT; } -static inline int fid_seq_is_dot(__u64 seq) +static inline bool fid_seq_is_dot(__u64 seq) { return seq == FID_SEQ_DOT_LUSTRE; } -static inline int fid_seq_is_default(__u64 seq) +static inline bool fid_seq_is_default(__u64 seq) { return seq == FID_SEQ_LOV_DEFAULT; } -static inline int fid_is_mdt0(const struct lu_fid *fid) +static inline bool fid_is_mdt0(const struct lu_fid *fid) { return fid_seq_is_mdt0(fid_seq(fid)); } @@ -516,12 +516,12 @@ static inline void lu_root_fid(struct lu_fid *fid) * \param fid the fid to be tested. * \return true if the fid is a igif; otherwise false. */ -static inline int fid_seq_is_igif(__u64 seq) +static inline bool
[PATCH 20/80] staging: lustre: llite: enable clients to inject error for lfsck
From: Fan YongThis enables the client to inject an error by altering the parent FID in order to test if the server side file system checker behaves properly. Signed-off-by: Fan Yong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3951 Reviewed-on: http://review.whamcloud.com/7667 Reviewed-by: Andreas Dilger Reviewed-by: Alex Zhuravlev Signed-off-by: James Simmons --- .../staging/lustre/lustre/include/obd_support.h|1 + drivers/staging/lustre/lustre/llite/vvp_req.c |2 ++ 2 files changed, 3 insertions(+), 0 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/obd_support.h b/drivers/staging/lustre/lustre/include/obd_support.h index 845e64a..71bf844 100644 --- a/drivers/staging/lustre/lustre/include/obd_support.h +++ b/drivers/staging/lustre/lustre/include/obd_support.h @@ -474,6 +474,7 @@ extern char obd_jobid_var[]; #define OBD_FAIL_LFSCK_CRASH 0x160a #define OBD_FAIL_LFSCK_NO_AUTO 0x160b #define OBD_FAIL_LFSCK_NO_DOUBLESCAN 0x160c +#define OBD_FAIL_LFSCK_INVALID_PFID0x1619 /* UPDATE */ #define OBD_FAIL_UPDATE_OBJ_NET0x1700 diff --git a/drivers/staging/lustre/lustre/llite/vvp_req.c b/drivers/staging/lustre/lustre/llite/vvp_req.c index 9fe9d6c..0567a15 100644 --- a/drivers/staging/lustre/lustre/llite/vvp_req.c +++ b/drivers/staging/lustre/lustre/llite/vvp_req.c @@ -83,6 +83,8 @@ static void vvp_req_attr_set(const struct lu_env *env, } obdo_from_inode(oa, inode, valid_flags & flags); obdo_set_parent_fid(oa, _i2info(inode)->lli_fid); + if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_INVALID_PFID)) + oa->o_parent_oid++; memcpy(attr->cra_jobid, ll_i2info(inode)->lli_jobid, JOBSTATS_JOBID_SIZE); } -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 46/80] staging: lustre: llite: Change readdir BRW metrics
From: Jinshan XiongTo simplify the code, change the metrics from bytes to pages. Signed-off-by: Andreas Dilger Signed-off-by: Jinshan Xiong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5034 Reviewed-on: http://review.whamcloud.com/10275 Reviewed-by: wangdi Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/dir.c |2 +- .../staging/lustre/lustre/llite/llite_internal.h |2 +- drivers/staging/lustre/lustre/llite/llite_lib.c|4 ++-- 3 files changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index 47fbcd2..924b5df 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -148,7 +148,7 @@ static int ll_dir_filler(void *_hash, struct page *page0) struct page **page_pool; struct page *page; struct lu_dirpage *dp; - int max_pages = ll_i2sbi(inode)->ll_md_brw_size >> PAGE_SHIFT; + int max_pages = ll_i2sbi(inode)->ll_md_brw_pages; int nrdpgs = 0; /* number of pages read actually */ int npages; int i; diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h index 500b5ec..3d7fa9a 100644 --- a/drivers/staging/lustre/lustre/llite/llite_internal.h +++ b/drivers/staging/lustre/lustre/llite/llite_internal.h @@ -474,7 +474,7 @@ struct ll_sb_info { unsigned int ll_namelen; struct file_operations *ll_fop; - unsigned int ll_md_brw_size; /* used by readdir */ + unsigned int ll_md_brw_pages; /* readdir pages per RPC */ struct lu_site *ll_site; struct cl_device *ll_cl; diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c index 0a28925..ac59cd6 100644 --- a/drivers/staging/lustre/lustre/llite/llite_lib.c +++ b/drivers/staging/lustre/lustre/llite/llite_lib.c @@ -319,9 +319,9 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt, sbi->ll_flags |= LL_SBI_64BIT_HASH; if (data->ocd_connect_flags & OBD_CONNECT_BRW_SIZE) - sbi->ll_md_brw_size = data->ocd_brw_size; + sbi->ll_md_brw_pages = data->ocd_brw_size >> PAGE_SHIFT; else - sbi->ll_md_brw_size = PAGE_SIZE; + sbi->ll_md_brw_pages = 1; if (data->ocd_connect_flags & OBD_CONNECT_LAYOUTLOCK) sbi->ll_flags |= LL_SBI_LAYOUT_LOCK; -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 36/80] staging: lustre: reorder LOV_MAGIC_* definition
From: Fan YongSince all the LOV_MAGIC_* definitions have the same postfix values break that value out into its own definition. With this we can check whether the magic's postfix match the LOV_MAGIC_MAGIC or not: if yes, then it is quite possible that the clients has encountered an newer LOV magic. This extra information can let us handle those cases more gracefully. Signed-off-by: Fan Yong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4941 Reviewed-on: http://review.whamcloud.com/10045 Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h | 20 +++- 1 files changed, 15 insertions(+), 5 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index d3a9db9..3444add 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -1478,11 +1478,21 @@ enum obdo_flags { OBD_FL_LOCAL_MASK = 0xF000, }; -#define LOV_MAGIC_V1 0x0BD10BD0 -#define LOV_MAGIC LOV_MAGIC_V1 -#define LOV_MAGIC_JOIN_V1 0x0BD20BD0 -#define LOV_MAGIC_V3 0x0BD30BD0 -#define LOV_MAGIC_MIGRATE 0x0BD40BD0 +/* + * All LOV EA magics should have the same postfix, if some new version + * Lustre instroduces new LOV EA magic, then when down-grade to an old + * Lustre, even though the old version system does not recognizes such + * new magic, it still can distinguish the corrupted cases by checking + * the magic's postfix. + */ +#define LOV_MAGIC_MAGIC 0x0BD0 +#define LOV_MAGIC_MASK 0x + +#define LOV_MAGIC_V1 (0x0BD1 | LOV_MAGIC_MAGIC) +#define LOV_MAGIC_JOIN_V1 (0x0BD2 | LOV_MAGIC_MAGIC) +#define LOV_MAGIC_V3 (0x0BD3 | LOV_MAGIC_MAGIC) +#define LOV_MAGIC_MIGRATE (0x0BD4 | LOV_MAGIC_MAGIC) +#define LOV_MAGIC LOV_MAGIC_V1 /* * magic for fully defined striping -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 34/80] staging: lustre: lmv: rename request to preq in lmv_getattr_name()
From: John L. HammondRename request to preq. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4826 Reviewed-on: http://review.whamcloud.com/9863 Reviewed-by: Nathaniel Clark Reviewed-by: Bob Glossman Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/lmv/lmv_obd.c | 16 +++- 1 files changed, 7 insertions(+), 9 deletions(-) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c index f1b8ae9..d07fd17 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c @@ -1883,7 +1883,7 @@ lmv_enqueue(struct obd_export *exp, struct ldlm_enqueue_info *einfo, static int lmv_getattr_name(struct obd_export *exp, struct md_op_data *op_data, -struct ptlrpc_request **request) +struct ptlrpc_request **preq) { struct ptlrpc_request *req = NULL; struct obd_device *obd = exp->exp_obd; @@ -1904,13 +1904,11 @@ lmv_getattr_name(struct obd_export *exp, struct md_op_data *op_data, op_data->op_namelen, op_data->op_name, PFID(_data->op_fid1), tgt->ltd_idx); - rc = md_getattr_name(tgt->ltd_exp, op_data, request); + rc = md_getattr_name(tgt->ltd_exp, op_data, preq); if (rc != 0) return rc; - body = req_capsule_server_get(&(*request)->rq_pill, - _MDT_BODY); - + body = req_capsule_server_get(&(*preq)->rq_pill, _MDT_BODY); if (body->valid & OBD_MD_MDS) { struct lu_fid rid = body->fid1; @@ -1919,8 +1917,8 @@ lmv_getattr_name(struct obd_export *exp, struct md_op_data *op_data, tgt = lmv_find_target(lmv, ); if (IS_ERR(tgt)) { - ptlrpc_req_finished(*request); - *request = NULL; + ptlrpc_req_finished(*preq); + *preq = NULL; return PTR_ERR(tgt); } @@ -1929,8 +1927,8 @@ lmv_getattr_name(struct obd_export *exp, struct md_op_data *op_data, op_data->op_namelen = 0; op_data->op_name = NULL; rc = md_getattr_name(tgt->ltd_exp, op_data, ); - ptlrpc_req_finished(*request); - *request = req; + ptlrpc_req_finished(*preq); + *preq = req; } return rc; -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 05/80] staging: lustre: llite: change remove parameter to bool
From: wang diChange the 3rd parameter remove to a bool for ll_release_page function. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/dir.c | 10 +- .../staging/lustre/lustre/llite/llite_internal.h |2 +- drivers/staging/lustre/lustre/llite/statahead.c| 10 +- 3 files changed, 11 insertions(+), 11 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index 3a800b2..a72b486 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -236,7 +236,7 @@ static int ll_dir_filler(void *_hash, struct page *page0) return rc; } -void ll_release_page(struct inode *inode, struct page *page, int remove) +void ll_release_page(struct inode *inode, struct page *page, bool remove) { kunmap(page); if (remove) { @@ -297,7 +297,7 @@ static struct page *ll_dir_page_locate(struct inode *dir, __u64 *hash, CDEBUG(D_VFSTRACE, "page %lu [%llu %llu], hash %llu\n", offset, *start, *end, *hash); if (*hash > *end) { - ll_release_page(dir, page, 0); + ll_release_page(dir, page, false); page = NULL; } else if (*end != *start && *hash == *end) { /* @@ -463,7 +463,7 @@ out_unlock: return page; fail: - ll_release_page(dir, page, 1); + ll_release_page(dir, page, true); page = ERR_PTR(-EIO); goto out_unlock; } @@ -561,7 +561,7 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, if (done) { pos = hash; - ll_release_page(inode, page, 0); + ll_release_page(inode, page, false); break; } @@ -572,7 +572,7 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, * End of directory reached. */ done = 1; - ll_release_page(inode, page, 0); + ll_release_page(inode, page, false); } else { /* * Normal case: continue to the next diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h index 4b03a64..07b6918 100644 --- a/drivers/staging/lustre/lustre/llite/llite_internal.h +++ b/drivers/staging/lustre/lustre/llite/llite_internal.h @@ -655,7 +655,7 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, int ll_get_mdt_idx(struct inode *inode); struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data, __u64 hash, struct ll_dir_chain *chain); -void ll_release_page(struct inode *inode, struct page *page, int remove); +void ll_release_page(struct inode *inode, struct page *page, bool remove); /* llite/namei.c */ extern const struct inode_operations ll_special_inode_operations; diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c index 2949ff6..6ce7442 100644 --- a/drivers/staging/lustre/lustre/llite/statahead.c +++ b/drivers/staging/lustre/lustre/llite/statahead.c @@ -1142,7 +1142,7 @@ interpret_it: ll_post_statahead(sai); if (unlikely(!thread_is_running(thread))) { - ll_release_page(dir, page, 0); + ll_release_page(dir, page, false); rc = 0; goto out; } @@ -1166,7 +1166,7 @@ interpret_it: if (unlikely( !thread_is_running(thread))) { - ll_release_page(dir, page, 0); + ll_release_page(dir, page, false); rc = 0; goto out; } @@ -1189,7 +1189,7 @@ do_it: /* * End of directory reached. */ - ll_release_page(dir, page, 0); +
[PATCH 21/80] staging: lustre: osc: allow to call brw_commit() multiple times
From: Jinshan XiongSometimes the rq_commit_cb of BRW RPC can be called twice if that RPC has already committed at reply time. This will cause inaccuracy of unstable pages accounting and then assertion. Signed-off-by: Jinshan Xiong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3274 Reviewed-on: http://review.whamcloud.com/8215 Reviewed-by: Prakash Surya Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/osc/osc_cache.c | 19 --- drivers/staging/lustre/lustre/osc/osc_request.c |8 2 files changed, 8 insertions(+), 19 deletions(-) diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c index 53b5d73..683b3c2 100644 --- a/drivers/staging/lustre/lustre/osc/osc_cache.c +++ b/drivers/staging/lustre/lustre/osc/osc_cache.c @@ -1875,11 +1875,6 @@ void osc_dec_unstable_pages(struct ptlrpc_request *req) atomic_sub(page_count, _unstable_pages); LASSERT(atomic_read(_unstable_pages) >= 0); - spin_lock(>rq_lock); - req->rq_committed = 1; - req->rq_unstable = 0; - spin_unlock(>rq_lock); - wake_up_all(>cl_cache->ccc_unstable_waitq); } @@ -1909,27 +1904,21 @@ void osc_inc_unstable_pages(struct ptlrpc_request *req) LASSERT(atomic_read(_unstable_pages) >= 0); atomic_add(page_count, _unstable_pages); - spin_lock(>rq_lock); - /* * If the request has already been committed (i.e. brw_commit * called via rq_commit_cb), we need to undo the unstable page * increments we just performed because rq_commit_cb wont be -* called again. Otherwise, just set the commit callback so the -* unstable page accounting is properly updated when the request -* is committed +* called again. */ - if (req->rq_committed) { + spin_lock(>rq_lock); + if (unlikely(req->rq_committed)) { /* Drop lock before calling osc_dec_unstable_pages */ spin_unlock(>rq_lock); osc_dec_unstable_pages(req); - spin_lock(>rq_lock); } else { req->rq_unstable = 1; - req->rq_commit_cb = osc_dec_unstable_pages; + spin_unlock(>rq_lock); } - - spin_unlock(>rq_lock); } /* this must be called holding the loi list lock to give coverage to exit_cache, diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c index 536b868..a2d948f 100644 --- a/drivers/staging/lustre/lustre/osc/osc_request.c +++ b/drivers/staging/lustre/lustre/osc/osc_request.c @@ -1847,21 +1847,21 @@ static int brw_interpret(const struct lu_env *env, static void brw_commit(struct ptlrpc_request *req) { - spin_lock(>rq_lock); /* * If osc_inc_unstable_pages (via osc_extent_finish) races with * this called via the rq_commit_cb, I need to ensure * osc_dec_unstable_pages is still called. Otherwise unstable * pages may be leaked. */ - if (req->rq_unstable) { + spin_lock(>rq_lock); + if (unlikely(req->rq_unstable)) { + req->rq_unstable = 0; spin_unlock(>rq_lock); osc_dec_unstable_pages(req); - spin_lock(>rq_lock); } else { req->rq_committed = 1; + spin_unlock(>rq_lock); } - spin_unlock(>rq_lock); } /** -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 02/80] staging: lustre: llite: remove comment from ll_dir_read
From: wang diRemove comment about fixing swabbing that is not needed. Signed-off-by: wang di Reviewed-on: http://review.whamcloud.com/7043 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3531 Reviewed-by: John L. Hammond Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/llite/dir.c |4 1 files changed, 0 insertions(+), 4 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c index 82c7f88..d854edd 100644 --- a/drivers/staging/lustre/lustre/llite/dir.c +++ b/drivers/staging/lustre/lustre/llite/dir.c @@ -526,10 +526,6 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, __u64 lhash; __u64 ino; - /* -* XXX: implement correct swabbing here. -*/ - hash = le64_to_cpu(ent->lde_hash); if (hash < pos) /* -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 12/80] staging: lustre: create striped directory
From: wang di1. client send create request to the master MDT, which will allocate FIDs and create slaves. for all of slaves. 2. Client needs to revalidate slaves during intent getattr and open request. 3. lmv_stripe_md will include attributes(size, nlink etc) from all of stripe, which will be protected by UPDATE lock. client needs to merge these attributes when update inode. 4. send create request to the MDT where the file is located, which can help creating master stripe of striped directory. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3529 Reviewed-on: http://review.whamcloud.com/7196 Reviewed-by: Andreas Dilger Reviewed-by: John L. Hammond Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/include/cl_object.h |3 + .../lustre/lustre/include/lustre/lustre_idl.h | 40 +++- .../lustre/lustre/include/lustre/lustre_user.h | 16 +- drivers/staging/lustre/lustre/include/lustre_lib.h |2 + drivers/staging/lustre/lustre/include/lustre_lmv.h | 59 + drivers/staging/lustre/lustre/include/obd.h| 16 +- drivers/staging/lustre/lustre/include/obd_class.h | 19 ++ drivers/staging/lustre/lustre/llite/dir.c | 26 ++- drivers/staging/lustre/lustre/llite/file.c | 40 +++- .../staging/lustre/lustre/llite/llite_internal.h | 12 +- drivers/staging/lustre/lustre/llite/llite_lib.c| 193 +++- drivers/staging/lustre/lustre/llite/llite_nfs.c|7 +- drivers/staging/lustre/lustre/llite/namei.c| 42 +++- drivers/staging/lustre/lustre/lmv/lmv_intent.c | 244 +--- drivers/staging/lustre/lustre/lmv/lmv_internal.h | 32 +++ drivers/staging/lustre/lustre/lmv/lmv_obd.c| 221 +++--- drivers/staging/lustre/lustre/mdc/mdc_locks.c |3 + .../staging/lustre/lustre/ptlrpc/pack_generic.c| 11 + 18 files changed, 880 insertions(+), 106 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h index 3cd4a25..0fa71a5 100644 --- a/drivers/staging/lustre/lustre/include/cl_object.h +++ b/drivers/staging/lustre/lustre/include/cl_object.h @@ -191,6 +191,9 @@ struct cl_attr { * Group identifier for quota purposes. */ gid_t cat_gid; + + /* nlink of the directory */ + __u64 cat_nlink; }; /** diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 0ad6605..a612080 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -1610,6 +1610,7 @@ static inline void lmm_oi_cpu_to_le(struct ost_id *dst_oi, #define XATTR_NAME_LOV "trusted.lov" #define XATTR_NAME_LMA "trusted.lma" #define XATTR_NAME_LMV "trusted.lmv" +#define XATTR_NAME_DEFAULT_LMV "trusted.dmv" #define XATTR_NAME_LINK "trusted.link" #define XATTR_NAME_FID "trusted.fid" #define XATTR_NAME_VERSION "trusted.version" @@ -2472,7 +2473,7 @@ struct lmv_desc { __u32 ld_tgt_count; /* how many MDS's */ __u32 ld_active_tgt_count; /* how many active */ __u32 ld_default_stripe_count; /* how many objects are used */ - __u32 ld_pattern; /* default MEA_MAGIC_* */ + __u32 ld_pattern; /* default hash pattern */ __u64 ld_default_hash_size; __u64 ld_padding_1; /* also fix lustre_swab_lmv_desc */ __u32 ld_padding_2; /* also fix lustre_swab_lmv_desc */ @@ -2486,6 +2487,43 @@ struct lmv_desc { #define LMV_MAGIC_V1 0x0CD10CD0 /* normal stripe lmv magic */ #define LMV_USER_MAGIC 0x0CD20CD0 /* default lmv magic*/ #define LMV_MAGIC LMV_MAGIC_V1 + +enum lmv_hash_type { + LMV_HASH_TYPE_ALL_CHARS = 1, + LMV_HASH_TYPE_FNV_1A_64 = 2, +}; + +#define LMV_HASH_NAME_ALL_CHARS"all_char" +#define LMV_HASH_NAME_FNV_1A_64"fnv_1a_64" + +/** + * The FNV-1a hash algorithm is as follows: + * hash = FNV_offset_basis + * for each octet_of_data to be hashed + * hash = hash XOR octet_of_data + * hash = hash à FNV_prime + * return hash + * http://en.wikipedia.org/wiki/FowlerâNollâVo_hash_function#FNV-1a_hash + * + * http://www.isthe.com/chongo/tech/comp/fnv/index.html#FNV-reference-source + * FNV_prime is 2^40 + 2^8 + 0xb3 = 0x10001b3ULL + **/ +#define LUSTRE_FNV_1A_64_PRIME 0x10001b3ULL +#define LUSTRE_FNV_1A_64_OFFSET_BIAS 0xcbf29ce484222325ULL +static inline __u64 lustre_hash_fnv_1a_64(const void *buf, size_t size) +{ + __u64 hash = LUSTRE_FNV_1A_64_OFFSET_BIAS; + const unsigned char *p = buf; + size_t i; + + for (i = 0; i < size;
[PATCH 24/80] staging: lustre: don't need to const __u64 parameters for lustre_idl.h
From: John L. HammondRemove the const for the __u64 parameters for inline functions in lustre_idl.h. Signed-off-by: John L. Hammond Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675 Reviewed-on: http://review.whamcloud.com/8641 Reviewed-by: wangdi Reviewed-by: Fan Yong Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h | 20 ++-- 1 files changed, 10 insertions(+), 10 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 6853f62..87e79b9 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -442,7 +442,7 @@ static inline int fid_seq_is_mdt0(__u64 seq) return (seq == FID_SEQ_OST_MDT0); } -static inline int fid_seq_is_mdt(const __u64 seq) +static inline int fid_seq_is_mdt(__u64 seq) { return seq == FID_SEQ_OST_MDT0 || seq >= FID_SEQ_NORMAL; }; @@ -468,33 +468,33 @@ static inline int fid_is_llog(const struct lu_fid *fid) return fid_seq_is_llog(fid_seq(fid)) && fid_oid(fid) > 0; } -static inline int fid_seq_is_rsvd(const __u64 seq) +static inline int fid_seq_is_rsvd(__u64 seq) { return (seq > FID_SEQ_OST_MDT0 && seq <= FID_SEQ_RSVD); }; -static inline int fid_seq_is_special(const __u64 seq) +static inline int fid_seq_is_special(__u64 seq) { return seq == FID_SEQ_SPECIAL; }; -static inline int fid_seq_is_local_file(const __u64 seq) +static inline int fid_seq_is_local_file(__u64 seq) { return seq == FID_SEQ_LOCAL_FILE || seq == FID_SEQ_LOCAL_NAME; }; -static inline int fid_seq_is_root(const __u64 seq) +static inline int fid_seq_is_root(__u64 seq) { return seq == FID_SEQ_ROOT; } -static inline int fid_seq_is_dot(const __u64 seq) +static inline int fid_seq_is_dot(__u64 seq) { return seq == FID_SEQ_DOT_LUSTRE; } -static inline int fid_seq_is_default(const __u64 seq) +static inline int fid_seq_is_default(__u64 seq) { return seq == FID_SEQ_LOV_DEFAULT; } @@ -516,7 +516,7 @@ static inline void lu_root_fid(struct lu_fid *fid) * \param fid the fid to be tested. * \return true if the fid is a igif; otherwise false. */ -static inline int fid_seq_is_igif(const __u64 seq) +static inline int fid_seq_is_igif(__u64 seq) { return seq >= FID_SEQ_IGIF && seq <= FID_SEQ_IGIF_MAX; } @@ -531,7 +531,7 @@ static inline int fid_is_igif(const struct lu_fid *fid) * \param fid the fid to be tested. * \return true if the fid is a idif; otherwise false. */ -static inline int fid_seq_is_idif(const __u64 seq) +static inline int fid_seq_is_idif(__u64 seq) { return seq >= FID_SEQ_IDIF && seq <= FID_SEQ_IDIF_MAX; } @@ -546,7 +546,7 @@ static inline int fid_is_local_file(const struct lu_fid *fid) return fid_seq_is_local_file(fid_seq(fid)); } -static inline int fid_seq_is_norm(const __u64 seq) +static inline int fid_seq_is_norm(__u64 seq) { return (seq >= FID_SEQ_NORMAL); } -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 67/80] staging: lustre: ptlrpc: request gets stuck in UNREGISTERING phase
From: Andriy SkulyshExit condition from UNREGISTERING phase is releasing of both reply and bulk buffers. Call ptlrpc_unregister_bulk() if ptlrpc_unregister_reply() wasn't completed in async mode before switching to UNREGISTERING phase. Signed-off-by: Andriy Skulysh Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5259 Xyratex-bug-id: MRP-1960 Reviewed-on: http://review.whamcloud.com/10846 Reviewed-by: Andreas Dilger Reviewed-by: Ann Koehler Reviewed-by: Liang Zhen Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/ptlrpc/client.c |4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c index f2e71b4..bae91bd 100644 --- a/drivers/staging/lustre/lustre/ptlrpc/client.c +++ b/drivers/staging/lustre/lustre/ptlrpc/client.c @@ -1630,8 +1630,10 @@ int ptlrpc_check_set(const struct lu_env *env, struct ptlrpc_request_set *set) req->rq_waiting || req->rq_wait_ctx) { int status; - if (!ptlrpc_unregister_reply(req, 1)) + if (!ptlrpc_unregister_reply(req, 1)) { + ptlrpc_unregister_bulk(req, 1); continue; + } spin_lock(>imp_lock); if (ptlrpc_import_delay_req(imp, req, -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 57/80] staging: lustre: osc: revise unstable pages accounting
From: Jinshan XiongA few changes are made in this patch for unstable pages tracking: 1. Remove kernel NFS unstable pages tracking because it killed performance 2. Track unstable pages as part of LRU cache. Otherwise Lustre can use much more memory than max_cached_mb 3. Remove obd_unstable_pages tracking to avoid using global atomic counter 4. Make unstable pages track optional. Tracking unstable pages is turned off by default, and can be controlled by llite.*.unstable_stats. Signed-off-by: Jinshan Xiong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4841 Reviewed-on: http://review.whamcloud.com/10003 Reviewed-by: Andreas Dilger Reviewed-by: Lai Siyao Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/include/cl_object.h | 35 +++- .../staging/lustre/lustre/include/obd_support.h|1 - drivers/staging/lustre/lustre/llite/lproc_llite.c | 41 - drivers/staging/lustre/lustre/obdclass/class_obd.c |2 - drivers/staging/lustre/lustre/osc/osc_cache.c | 96 +- drivers/staging/lustre/lustre/osc/osc_internal.h |2 +- drivers/staging/lustre/lustre/osc/osc_page.c | 208 +--- drivers/staging/lustre/lustre/osc/osc_request.c| 13 +- 8 files changed, 253 insertions(+), 145 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h index d269b32..ec6cf7c 100644 --- a/drivers/staging/lustre/lustre/include/cl_object.h +++ b/drivers/staging/lustre/lustre/include/cl_object.h @@ -1039,23 +1039,32 @@ do { \ }\ } while (0) -static inline int __page_in_use(const struct cl_page *page, int refc) -{ - if (page->cp_type == CPT_CACHEABLE) - ++refc; - LASSERT(atomic_read(>cp_ref) > 0); - return (atomic_read(>cp_ref) > refc); -} - -#define cl_page_in_use(pg) __page_in_use(pg, 1) -#define cl_page_in_use_noref(pg) __page_in_use(pg, 0) - static inline struct page *cl_page_vmpage(struct cl_page *page) { LASSERT(page->cp_vmpage); return page->cp_vmpage; } +/** + * Check if a cl_page is in use. + * + * Client cache holds a refcount, this refcount will be dropped when + * the page is taken out of cache, see vvp_page_delete(). + */ +static inline bool __page_in_use(const struct cl_page *page, int refc) +{ + return (atomic_read(>cp_ref) > refc + 1); +} + +/** + * Caller itself holds a refcount of cl_page. + */ +#define cl_page_in_use(pg) __page_in_use(pg, 1) +/** + * Caller doesn't hold a refcount. + */ +#define cl_page_in_use_noref(pg) __page_in_use(pg, 0) + /** @} cl_page */ /** \addtogroup cl_lock cl_lock @@ -2331,6 +2340,10 @@ struct cl_client_cache { */ spinlock_t ccc_lru_lock; /** +* Set if unstable check is enabled +*/ + unsigned intccc_unstable_check:1; + /** * # of unstable pages for this mount point */ atomic_tccc_unstable_nr; diff --git a/drivers/staging/lustre/lustre/include/obd_support.h b/drivers/staging/lustre/lustre/include/obd_support.h index 26fdff6..a11fff1 100644 --- a/drivers/staging/lustre/lustre/include/obd_support.h +++ b/drivers/staging/lustre/lustre/include/obd_support.h @@ -54,7 +54,6 @@ extern int at_early_margin; extern int at_extra; extern unsigned int obd_sync_filter; extern unsigned int obd_max_dirty_pages; -extern atomic_t obd_unstable_pages; extern atomic_t obd_dirty_pages; extern atomic_t obd_dirty_transit_pages; extern char obd_jobid_var[]; diff --git a/drivers/staging/lustre/lustre/llite/lproc_llite.c b/drivers/staging/lustre/lustre/llite/lproc_llite.c index 2f1f389..5f8e78d 100644 --- a/drivers/staging/lustre/lustre/llite/lproc_llite.c +++ b/drivers/staging/lustre/lustre/llite/lproc_llite.c @@ -828,10 +828,45 @@ static ssize_t unstable_stats_show(struct kobject *kobj, pages = atomic_read(>ccc_unstable_nr); mb = (pages * PAGE_SIZE) >> 20; - return sprintf(buf, "unstable_pages: %8d\n" - "unstable_mb:%8d\n", pages, mb); + return sprintf(buf, "unstable_check: %8d\n" + "unstable_pages: %8d\n" + "unstable_mb:%8d\n", + cache->ccc_unstable_check, pages, mb); } -LUSTRE_RO_ATTR(unstable_stats); + +static ssize_t unstable_stats_store(struct kobject *kobj, + struct attribute *attr, + const char *buffer, + size_t count) +{ + struct ll_sb_info *sbi = container_of(kobj, struct
[PATCH 64/80] staging: lustre: ptlrpc: add OBD_CONNECT_UNLINK_CLOSE flag
From: Lai SiyaoAdd OBD_CONNECT_UNLINK_CLOSE flag for interop, once this is supported, client packs file handle in unlink RPC, and MDT will close file before unlink. Signed-off-by: Lai Siyao Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4367 Reviewed-on: http://review.whamcloud.com/10426 Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h |1 + .../lustre/lustre/obdclass/lprocfs_status.c|2 ++ drivers/staging/lustre/lustre/ptlrpc/wiretest.c|2 ++ 3 files changed, 5 insertions(+), 0 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index a9661c0..4a7ccc8 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -1290,6 +1290,7 @@ void lustre_swab_ptlrpc_body(struct ptlrpc_body *pb); * name in request */ #define OBD_CONNECT_LFSCK 0x40ULL/* support online LFSCK */ +#define OBD_CONNECT_UNLINK_CLOSE 0x100ULL/* close file in unlink */ /* XXX README XXX: * Please DO NOT add flag values here before first ensuring that this same diff --git a/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c b/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c index fbb0851..45e3c4a 100644 --- a/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c +++ b/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c @@ -99,6 +99,8 @@ static const char * const obd_connect_names[] = { "open_by_fid", "lfsck", "unknown", + "unlink_close", + "unknown", NULL }; diff --git a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c index 60d03dd..2c718e0 100644 --- a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c +++ b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c @@ -1073,6 +1073,8 @@ void lustre_assert_wire_constants(void) "found 0x%.16llxULL\n", OBD_CONNECT_OPEN_BY_FID); LASSERTF(OBD_CONNECT_LFSCK == 0x40ULL, "found 0x%.16llxULL\n", OBD_CONNECT_LFSCK); + LASSERTF(OBD_CONNECT_UNLINK_CLOSE == 0x100ULL, "found 0x%.16llxULL\n", +OBD_CONNECT_UNLINK_CLOSE); LASSERTF(OBD_CKSUM_CRC32 == 0x0001UL, "found 0x%.8xUL\n", (unsigned)OBD_CKSUM_CRC32); LASSERTF(OBD_CKSUM_ADLER == 0x0002UL, "found 0x%.8xUL\n", -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 66/80] staging: lustre: lmv: try all stripes for unknown hash functions
From: wang diFor unknown hash type, LMV should try all stripes to locate the name entry. But it will only for lookup and unlink, i.e. we can only list and unlink entries under striped dir with unknown hash type. Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4921 Reviewed-on: http://review.whamcloud.com/10041 Reviewed-by: John L. Hammond Reviewed-by: Andreas Dilger Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_user.h |1 + .../staging/lustre/lustre/include/obd_support.h|3 + drivers/staging/lustre/lustre/lmv/lmv_intent.c | 70 +++--- drivers/staging/lustre/lustre/lmv/lmv_internal.h | 12 ++ drivers/staging/lustre/lustre/lmv/lmv_obd.c| 144 5 files changed, 182 insertions(+), 48 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h index 9e38ed3..52cd585 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h @@ -383,6 +383,7 @@ struct lmv_user_mds_data { }; enum lmv_hash_type { + LMV_HASH_TYPE_UNKNOWN = 0,/* 0 is reserved for testing purpose */ LMV_HASH_TYPE_ALL_CHARS = 1, LMV_HASH_TYPE_FNV_1A_64 = 2, }; diff --git a/drivers/staging/lustre/lustre/include/obd_support.h b/drivers/staging/lustre/lustre/include/obd_support.h index a11fff1..f747bca 100644 --- a/drivers/staging/lustre/lustre/include/obd_support.h +++ b/drivers/staging/lustre/lustre/include/obd_support.h @@ -483,6 +483,9 @@ extern char obd_jobid_var[]; #define OBD_FAIL_UPDATE_OBJ_NET0x1700 #define OBD_FAIL_UPDATE_OBJ_NET_REP0x1701 +/* LMV */ +#define OBD_FAIL_UNKNOWN_LMV_STRIPE0x1901 + /* Assign references to moved code to reduce code changes */ #define OBD_FAIL_PRECHECK(id) CFS_FAIL_PRECHECK(id) #define OBD_FAIL_CHECK(id) CFS_FAIL_CHECK(id) diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c index cde1d7b..0559445 100644 --- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c +++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c @@ -402,10 +402,28 @@ static int lmv_intent_lookup(struct obd_export *exp, struct mdt_body *body; int rc = 0; + /* +* If it returns ERR_PTR(-EBADFD) then it is an unknown hash type +* it will try all stripes to locate the object +*/ tgt = lmv_locate_mds(lmv, op_data, _data->op_fid1); - if (IS_ERR(tgt)) + if (IS_ERR(tgt) && (PTR_ERR(tgt) != -EBADFD)) return PTR_ERR(tgt); + /* +* Both migrating dir and unknown hash dir need to try +* all of sub-stripes +*/ + if (lsm && !lmv_is_known_hash_type(lsm)) { + struct lmv_oinfo *oinfo = >lsm_md_oinfo[0]; + + op_data->op_fid1 = oinfo->lmo_fid; + op_data->op_mds = oinfo->lmo_mds; + tgt = lmv_get_target(lmv, oinfo->lmo_mds, NULL); + if (IS_ERR(tgt)) + return PTR_ERR(tgt); + } + if (!fid_is_sane(_data->op_fid2)) fid_zero(_data->op_fid2); @@ -435,27 +453,39 @@ static int lmv_intent_lookup(struct obd_export *exp, } return rc; } else if (it_disposition(it, DISP_LOOKUP_NEG) && lsm && - lsm->lsm_md_magic & LMV_HASH_FLAG_MIGRATION) { + lmv_need_try_all_stripes(lsm)) { /* -* For migrating directory, if it can not find the child in -* the source directory(master stripe), try the targeting -* directory(stripe 1) +* For migrating and unknown hash type directory, it will +* try to target the entry on other stripes */ - tgt = lmv_find_target(lmv, >lsm_md_oinfo[1].lmo_fid); - if (IS_ERR(tgt)) - return PTR_ERR(tgt); - - ptlrpc_req_finished(*reqp); - it->it_request = NULL; - *reqp = NULL; - - CDEBUG(D_INODE, "For migrating dir, try target dir "DFID"\n", - PFID(>lsm_md_oinfo[1].lmo_fid)); - - op_data->op_fid1 = lsm->lsm_md_oinfo[1].lmo_fid; - it->it_disposition &= ~DISP_ENQ_COMPLETE; - rc = md_intent_lock(tgt->ltd_exp, op_data, lmm, lmmsize, it, - flags, reqp, cb_blocking, extra_lock_flags); + int stripe_index; + + for (stripe_index = 1; +stripe_index < lsm->lsm_md_stripe_count && +it_disposition(it, DISP_LOOKUP_NEG);
[PATCH 79/80] staging: lustre: lnet: lock improvement for ko2iblnd
From: Liang Zhenkiblnd_check_sends() takes conn::ibc_lock at the begin and release this lock at the end, this is inefficient because most use-case needs to explicitly release ibc_lock before caling this function. This patches changes it to kiblnd_check_sends_locked() and avoid unnecessary lock dances. Signed-off-by: Liang Zhen Signed-off-by: Doug Oucharek Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7099 Reviewed-on: http://review.whamcloud.com/20322 Reviewed-by: Amir Shehata Reviewed-by: Dmitry Eremin Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c | 46 1 files changed, 18 insertions(+), 28 deletions(-) diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c index 6cd78ea..6d1b14a 100644 --- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c +++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c @@ -38,7 +38,6 @@ static void kiblnd_peer_alive(struct kib_peer *peer); static void kiblnd_peer_connect_failed(struct kib_peer *peer, int active, int error); -static void kiblnd_check_sends(struct kib_conn *conn); static void kiblnd_init_tx_msg(lnet_ni_t *ni, struct kib_tx *tx, int type, int body_nob); static int kiblnd_init_rdma(struct kib_conn *conn, struct kib_tx *tx, int type, @@ -46,6 +45,7 @@ static int kiblnd_init_rdma(struct kib_conn *conn, struct kib_tx *tx, int type, static void kiblnd_queue_tx_locked(struct kib_tx *tx, struct kib_conn *conn); static void kiblnd_queue_tx(struct kib_tx *tx, struct kib_conn *conn); static void kiblnd_unmap_tx(lnet_ni_t *ni, struct kib_tx *tx); +static void kiblnd_check_sends_locked(struct kib_conn *conn); static void kiblnd_tx_done(lnet_ni_t *ni, struct kib_tx *tx) @@ -211,9 +211,9 @@ kiblnd_post_rx(struct kib_rx *rx, int credit) conn->ibc_outstanding_credits++; else conn->ibc_reserved_credits++; + kiblnd_check_sends_locked(conn); spin_unlock(>ibc_lock); - kiblnd_check_sends(conn); out: kiblnd_conn_decref(conn); return rc; @@ -344,8 +344,8 @@ kiblnd_handle_rx(struct kib_rx *rx) !IBLND_OOB_CAPABLE(conn->ibc_version)) /* v1 only */ conn->ibc_outstanding_credits++; + kiblnd_check_sends_locked(conn); spin_unlock(>ibc_lock); - kiblnd_check_sends(conn); } switch (msg->ibm_type) { @@ -800,7 +800,7 @@ kiblnd_post_tx_locked(struct kib_conn *conn, struct kib_tx *tx, int credit) conn->ibc_noops_posted == IBLND_OOB_MSGS(ver { /* * OK to drop when posted enough NOOPs, since -* kiblnd_check_sends will queue NOOP again when +* kiblnd_check_sends_locked will queue NOOP again when * posted NOOPs complete */ spin_unlock(>ibc_lock); @@ -905,7 +905,7 @@ kiblnd_post_tx_locked(struct kib_conn *conn, struct kib_tx *tx, int credit) } static void -kiblnd_check_sends(struct kib_conn *conn) +kiblnd_check_sends_locked(struct kib_conn *conn) { int ver = conn->ibc_version; lnet_ni_t *ni = conn->ibc_peer->ibp_ni; @@ -918,8 +918,6 @@ kiblnd_check_sends(struct kib_conn *conn) return; } - spin_lock(>ibc_lock); - LASSERT(conn->ibc_nsends_posted <= kiblnd_concurrent_sends(ver, ni)); LASSERT(!IBLND_OOB_CAPABLE(ver) || conn->ibc_noops_posted <= IBLND_OOB_MSGS(ver)); @@ -969,8 +967,6 @@ kiblnd_check_sends(struct kib_conn *conn) if (kiblnd_post_tx_locked(conn, tx, credit)) break; } - - spin_unlock(>ibc_lock); } static void @@ -1016,16 +1012,11 @@ kiblnd_tx_complete(struct kib_tx *tx, int status) if (idle) list_del(>tx_list); - kiblnd_conn_addref(conn); /* 1 ref for me */ - + kiblnd_check_sends_locked(conn); spin_unlock(>ibc_lock); if (idle) kiblnd_tx_done(conn->ibc_peer->ibp_ni, tx); - - kiblnd_check_sends(conn); - - kiblnd_conn_decref(conn); /* ...until here */ } static void @@ -1204,9 +1195,8 @@ kiblnd_queue_tx(struct kib_tx *tx, struct kib_conn *conn) { spin_lock(>ibc_lock); kiblnd_queue_tx_locked(tx, conn); + kiblnd_check_sends_locked(conn); spin_unlock(>ibc_lock); - - kiblnd_check_sends(conn); } static int kiblnd_resolve_addr(struct rdma_cm_id *cmid, @@ -2183,14 +2173,11 @@ kiblnd_connreq_done(struct kib_conn *conn, int status) return; } - /** -* refcount taken
[PATCH 65/80] staging: lustre: llog: keep llog ctxt indices constant
From: Mikhail PershinThe llog context id table cannot be shrunk easily because that will cause index shifting and incompatibility between old client and new server and vice versa. Patch moves llog_ctxt_id table to the lustre_idl.h because this is wire protocol data, these values are added to the wirecheck. Signed-off-by: Mikhail Pershin Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5218 Reviewed-on: http://review.whamcloud.com/10758 Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../lustre/lustre/include/lustre/lustre_idl.h | 24 +++- drivers/staging/lustre/lustre/include/obd.h| 21 - drivers/staging/lustre/lustre/ptlrpc/wiretest.c| 13 ++ 3 files changed, 36 insertions(+), 22 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h index 4a7ccc8..05fe359 100644 --- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h +++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h @@ -2936,7 +2936,29 @@ enum obd_cmd { }; #define OBD_FIRST_OPC OBD_PING -/* catalog of log objects */ +/** + * llog contexts indices. + * + * There is compatibility problem with indexes below, they are not + * continuous and must keep their numbers for compatibility needs. + * See LU-5218 for details. + */ +enum llog_ctxt_id { + LLOG_CONFIG_ORIG_CTXT = 0, + LLOG_CONFIG_REPL_CTXT = 1, + LLOG_MDS_OST_ORIG_CTXT = 2, + LLOG_MDS_OST_REPL_CTXT = 3, /* kept just to avoid re-assignment */ + LLOG_SIZE_ORIG_CTXT = 4, + LLOG_SIZE_REPL_CTXT = 5, + LLOG_TEST_ORIG_CTXT = 8, + LLOG_TEST_REPL_CTXT = 9, /* kept just to avoid re-assignment */ + LLOG_CHANGELOG_ORIG_CTXT = 12, /**< changelog generation on mdd */ + LLOG_CHANGELOG_REPL_CTXT = 13, /**< changelog access on clients */ + /* for multiple changelog consumers */ + LLOG_CHANGELOG_USER_ORIG_CTXT = 14, + LLOG_AGENT_ORIG_CTXT = 15, /**< agent requests generation on cdt */ + LLOG_MAX_CTXTS +}; /** Identifier for a single log object */ struct llog_logid { diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h index b7bdd07..e7e03be 100644 --- a/drivers/staging/lustre/lustre/include/obd.h +++ b/drivers/staging/lustre/lustre/include/obd.h @@ -172,27 +172,6 @@ struct brw_page { u32 flag; }; -/* llog contexts */ -enum llog_ctxt_id { - LLOG_CONFIG_ORIG_CTXT = 0, - LLOG_CONFIG_REPL_CTXT, - LLOG_MDS_OST_ORIG_CTXT, - LLOG_MDS_OST_REPL_CTXT, - LLOG_SIZE_ORIG_CTXT, - LLOG_SIZE_REPL_CTXT, - LLOG_RD1_ORIG_CTXT, - LLOG_RD1_REPL_CTXT, - LLOG_TEST_ORIG_CTXT, - LLOG_TEST_REPL_CTXT, - LLOG_LOVEA_ORIG_CTXT, - LLOG_LOVEA_REPL_CTXT, - LLOG_CHANGELOG_ORIG_CTXT, /**< changelog generation on mdd */ - LLOG_CHANGELOG_REPL_CTXT, /**< changelog access on clients */ - LLOG_CHANGELOG_USER_ORIG_CTXT, /**< for multiple changelog consumers */ - LLOG_AGENT_ORIG_CTXT, /**< agent requests generation on cdt */ - LLOG_MAX_CTXTS -}; - struct timeout_item { enum timeout_event ti_event; unsigned longti_timeout; diff --git a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c index 2c718e0..31d3326 100644 --- a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c +++ b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c @@ -3483,6 +3483,19 @@ void lustre_assert_wire_constants(void) CLASSERT(LLOG_ORIGIN_HANDLE_DESTROY == 509); CLASSERT(LLOG_FIRST_OPC == 501); CLASSERT(LLOG_LAST_OPC == 510); + CLASSERT(LLOG_CONFIG_ORIG_CTXT == 0); + CLASSERT(LLOG_CONFIG_REPL_CTXT == 1); + CLASSERT(LLOG_MDS_OST_ORIG_CTXT == 2); + CLASSERT(LLOG_MDS_OST_REPL_CTXT == 3); + CLASSERT(LLOG_SIZE_ORIG_CTXT == 4); + CLASSERT(LLOG_SIZE_REPL_CTXT == 5); + CLASSERT(LLOG_TEST_ORIG_CTXT == 8); + CLASSERT(LLOG_TEST_REPL_CTXT == 9); + CLASSERT(LLOG_CHANGELOG_ORIG_CTXT == 12); + CLASSERT(LLOG_CHANGELOG_REPL_CTXT == 13); + CLASSERT(LLOG_CHANGELOG_USER_ORIG_CTXT == 14); + CLASSERT(LLOG_AGENT_ORIG_CTXT == 15); + CLASSERT(LLOG_MAX_CTXTS == 16); /* Checks for struct llogd_conn_body */ LASSERTF((int)sizeof(struct llogd_conn_body) == 40, "found %lld\n", -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 58/80] staging: lustre: mdc: always use D_INFO for debug info when mdc_put_rpc_lock fails
From: wang diAlso use D_INFO no matter what the error returned from mdc_put_rpc_lock. Signed-off-by: Andreas Dilger Signed-off-by: wang di Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4973 Reviewed-on: http://review.whamcloud.com/10150 Reviewed-by: Mike Pershin Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/mdc/mdc_locks.c |5 ++--- 1 files changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/staging/lustre/lustre/mdc/mdc_locks.c b/drivers/staging/lustre/lustre/mdc/mdc_locks.c index 551f3d9..3291201 100644 --- a/drivers/staging/lustre/lustre/mdc/mdc_locks.c +++ b/drivers/staging/lustre/lustre/mdc/mdc_locks.c @@ -841,9 +841,8 @@ resend: mdc_put_rpc_lock(obddev->u.cli.cl_rpc_lock, it); if (rc < 0) { - CDEBUG_LIMIT((rc == -EACCES || rc == -EIDRM) ? D_INFO : D_ERROR, -"%s: ldlm_cli_enqueue failed: rc = %d\n", -obddev->obd_name, rc); + CDEBUG(D_INFO, "%s: ldlm_cli_enqueue failed: rc = %d\n", + obddev->obd_name, rc); mdc_clear_replay_flag(req, rc); ptlrpc_req_finished(req); -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 69/80] staging: lustre: osc: Automatically increase the max_dirty_mb
From: Hongchao ZhangWhen RPC size or the max RPCs in flight is increased, the actual limit might be max_dirty_mb. This patch automatically increases the max_dirty_mb value at connection time and when the related values are tuned manually by proc file system. this patch also changes the unit of "cl_dirty" and "cl_dirty_max" in client_obd from byte to page. Signed-off-by: Li Xi Signed-off-by: Hongchao Zhang Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4933 Reviewed-on: http://review.whamcloud.com/10446 Reviewed-by: Jinshan Xiong Reviewed-by: Andreas Dilger Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/include/obd.h | 28 +- drivers/staging/lustre/lustre/ldlm/ldlm_lib.c | 12 +--- drivers/staging/lustre/lustre/osc/lproc_osc.c | 10 --- drivers/staging/lustre/lustre/osc/osc_cache.c | 28 +- drivers/staging/lustre/lustre/osc/osc_request.c | 34 +- drivers/staging/lustre/lustre/ptlrpc/import.c |1 + 6 files changed, 74 insertions(+), 39 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h index e7e03be..e91f65a 100644 --- a/drivers/staging/lustre/lustre/include/obd.h +++ b/drivers/staging/lustre/lustre/include/obd.h @@ -222,8 +222,8 @@ struct client_obd { struct sptlrpc_flavorcl_flvr_mgc; /* fixed flavor of mgc->mgs */ /* the grant values are protected by loi_list_lock below */ - long cl_dirty; /* all _dirty_ in bytes */ - long cl_dirty_max; /* allowed w/o rpc */ + long cl_dirty_pages;/* all _dirty_ in pahges */ + long cl_dirty_max_pages;/* allowed w/o rpc */ long cl_dirty_transit; /* dirty synchronous */ long cl_avail_grant; /* bytes of credit for ost */ long cl_lost_grant;/* lost credits (trunc) */ @@ -1225,4 +1225,28 @@ static inline int cli_brw_size(struct obd_device *obd) return obd->u.cli.cl_max_pages_per_rpc << PAGE_SHIFT; } +/* + * when RPC size or the max RPCs in flight is increased, the max dirty pages + * of the client should be increased accordingly to avoid sending fragmented + * RPCs over the network when the client runs out of the maximum dirty space + * when so many RPCs are being generated. + */ +static inline void client_adjust_max_dirty(struct client_obd *cli) +{ + /* initializing */ + if (cli->cl_dirty_max_pages <= 0) + cli->cl_dirty_max_pages = + (OSC_MAX_DIRTY_DEFAULT * 1024 * 1024) >> PAGE_SHIFT; + else { + long dirty_max = cli->cl_max_rpcs_in_flight * +cli->cl_max_pages_per_rpc; + + if (dirty_max > cli->cl_dirty_max_pages) + cli->cl_dirty_max_pages = dirty_max; + } + + if (cli->cl_dirty_max_pages > totalram_pages / 8) + cli->cl_dirty_max_pages = totalram_pages / 8; +} + #endif /* __OBD_H */ diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c index ee40006..3c98ce2 100644 --- a/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c +++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c @@ -299,12 +299,14 @@ int client_obd_setup(struct obd_device *obddev, struct lustre_cfg *lcfg) min_t(unsigned int, LUSTRE_CFG_BUFLEN(lcfg, 2), sizeof(server_uuid))); - cli->cl_dirty = 0; + cli->cl_dirty_pages = 0; cli->cl_avail_grant = 0; - /* FIXME: Should limit this for the sum of all cl_dirty_max. */ - cli->cl_dirty_max = OSC_MAX_DIRTY_DEFAULT * 1024 * 1024; - if (cli->cl_dirty_max >> PAGE_SHIFT > totalram_pages / 8) - cli->cl_dirty_max = totalram_pages << (PAGE_SHIFT - 3); + /* FIXME: Should limit this for the sum of all cl_dirty_max_pages. */ + /* +* cl_dirty_max_pages may be changed at connect time in +* ptlrpc_connect_interpret(). +*/ + client_adjust_max_dirty(cli); INIT_LIST_HEAD(>cl_cache_waiters); INIT_LIST_HEAD(>cl_loi_ready_list); INIT_LIST_HEAD(>cl_loi_hp_ready_list); diff --git a/drivers/staging/lustre/lustre/osc/lproc_osc.c b/drivers/staging/lustre/lustre/osc/lproc_osc.c index 7e83d39..9172b78 100644 --- a/drivers/staging/lustre/lustre/osc/lproc_osc.c +++ b/drivers/staging/lustre/lustre/osc/lproc_osc.c @@ -119,6 +119,7 @@ static ssize_t max_rpcs_in_flight_store(struct kobject *kobj, spin_lock(>cl_loi_list_lock); cli->cl_max_rpcs_in_flight = val; + client_adjust_max_dirty(cli); spin_unlock(>cl_loi_list_lock); return count; @@ -136,10 +137,10 @@ static ssize_t
[PATCH 78/80] staging: lustre: lnet: make connection more stable with packet loss
From: Alexander BoykoIB network may lose last connection handshake packet. This problem isn't Lustre specific and described at https://oss.oracle.com/pipermail/rds-devel/2007-December/000271.html for example. Solution is to make conection established if any packet is received for it. Signed-off-by: Alexander Boyko Signed-off-by: Alexey Lyashkov Seagate-bug-id: MRP-2883 Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8303 Reviewed-on: http://review.whamcloud.com/20874 Reviewed-by: Doug Oucharek Reviewed-by: Alexander Zarochentsev Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c |6 ++ 1 files changed, 6 insertions(+), 0 deletions(-) diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c index 19c90fc..6cd78ea 100644 --- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c +++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c @@ -3419,6 +3419,12 @@ kiblnd_qp_event(struct ib_event *event, void *arg) case IB_EVENT_COMM_EST: CDEBUG(D_NET, "%s established\n", libcfs_nid2str(conn->ibc_peer->ibp_nid)); + /* +* We received a packet but connection isn't established +* probably handshake packet was lost, so free to +* force make connection established +*/ + rdma_notify(conn->ibc_cmid, IB_EVENT_COMM_EST); return; default: -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 60/80] staging: lustre: ldlm: improve ldlm_lock_create() return value
From: Emoly Liuldlm_lock_create() and ldlm_resource_get() always return NULL as error reporting and "NULL" is interpretted as ENOMEM incorrectly sometimes. This patch fixes this problem by using ERR_PTR() rather than NULL. Signed-off-by: Emoly Liu Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4524 Reviewed-on: http://review.whamcloud.com/9004 Reviewed-by: Bobi Jam Reviewed-by: Andreas Dilger Reviewed-by: John L. Hammond Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/ldlm/ldlm_flock.c|4 +- drivers/staging/lustre/lustre/ldlm/ldlm_lock.c | 28 drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 13 +++- drivers/staging/lustre/lustre/ldlm/ldlm_resource.c | 25 + drivers/staging/lustre/lustre/mdc/mdc_locks.c |2 +- drivers/staging/lustre/lustre/mdc/mdc_reint.c |2 +- drivers/staging/lustre/lustre/osc/osc_request.c|2 +- 7 files changed, 34 insertions(+), 42 deletions(-) diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c index 65e8e14..61d649f 100644 --- a/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c +++ b/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c @@ -339,10 +339,10 @@ reprocess: lock->l_granted_mode, _cbs, NULL, 0, LVB_T_NONE); lock_res_and_lock(req); - if (!new2) { + if (IS_ERR(new2)) { ldlm_flock_destroy(req, lock->l_granted_mode, *flags); - *err = -ENOLCK; + *err = PTR_ERR(new2); return LDLM_ITER_STOP; } goto reprocess; diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c index 1a0fce1..a91cdb4 100644 --- a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c +++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c @@ -481,8 +481,8 @@ int ldlm_lock_change_resource(struct ldlm_namespace *ns, struct ldlm_lock *lock, unlock_res_and_lock(lock); newres = ldlm_resource_get(ns, NULL, new_resid, type, 1); - if (!newres) - return -ENOMEM; + if (IS_ERR(newres)) + return PTR_ERR(newres); lu_ref_add(>lr_reference, "lock", lock); /* @@ -1227,7 +1227,7 @@ enum ldlm_mode ldlm_lock_match(struct ldlm_namespace *ns, __u64 flags, } res = ldlm_resource_get(ns, NULL, res_id, type, 0); - if (!res) { + if (IS_ERR(res)) { LASSERT(!old_lock); return 0; } @@ -1475,15 +1475,15 @@ struct ldlm_lock *ldlm_lock_create(struct ldlm_namespace *ns, { struct ldlm_lock *lock; struct ldlm_resource *res; + int rc; res = ldlm_resource_get(ns, NULL, res_id, type, 1); - if (!res) - return NULL; + if (IS_ERR(res)) + return ERR_CAST(res); lock = ldlm_lock_new(res); - if (!lock) - return NULL; + return ERR_PTR(-ENOMEM); lock->l_req_mode = mode; lock->l_ast_data = data; @@ -1497,27 +1497,33 @@ struct ldlm_lock *ldlm_lock_create(struct ldlm_namespace *ns, lock->l_tree_node = NULL; /* if this is the extent lock, allocate the interval tree node */ if (type == LDLM_EXTENT) { - if (!ldlm_interval_alloc(lock)) + if (!ldlm_interval_alloc(lock)) { + rc = -ENOMEM; goto out; + } } if (lvb_len) { lock->l_lvb_len = lvb_len; lock->l_lvb_data = kzalloc(lvb_len, GFP_NOFS); - if (!lock->l_lvb_data) + if (!lock->l_lvb_data) { + rc = -ENOMEM; goto out; + } } lock->l_lvb_type = lvb_type; - if (OBD_FAIL_CHECK(OBD_FAIL_LDLM_NEW_LOCK)) + if (OBD_FAIL_CHECK(OBD_FAIL_LDLM_NEW_LOCK)) { + rc = -ENOENT; goto out; + } return lock; out: ldlm_lock_destroy(lock); LDLM_LOCK_RELEASE(lock); - return NULL; + return ERR_PTR(rc); } /** diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c index 984a460..048214c 100644 --- a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c +++ b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c @@ -694,8 +694,8 @@ int ldlm_cli_enqueue(struct obd_export *exp, struct ptlrpc_request **reqp, lock =
[PATCH 80/80] staging: lustre: lnet: Stop Infinite CON RACE Condition
From: Doug OucharekIn current code, when a CON RACE occurs, the passive side will let the node with the higher NID value win the race. We have a field case where a node can have a "stuck" connection which never goes away and is the trigger of a never-ending loop of re-connections. This patch introduces a counter to how many times a connection in a connecting state has been the cause of a CON RACE rejection. After 20 times (constant MAX_CONN_RACES_BEFORE_ABORT), we assume the connection is stuck and let the other side (with lower NID) win. Signed-off-by: Doug Oucharek Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7646 Reviewed-on: http://review.whamcloud.com/19430 Reviewed-by: Amir Shehata Reviewed-by: Andreas Dilger Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- .../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h|2 + .../staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c | 24 --- 2 files changed, 22 insertions(+), 4 deletions(-) diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h index 078a0c3..fbc4f68 100644 --- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h +++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h @@ -582,6 +582,8 @@ struct kib_peer { unsigned short ibp_connecting; /* reconnect this peer later */ unsigned short ibp_reconnecting:1; + /* counter of how many times we triggered a conn race */ + unsigned char ibp_races; /* # consecutive reconnection attempts to this peer */ unsigned intibp_reconnected; /* errno on closing this peer */ diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c index 6d1b14a..430ff85 100644 --- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c +++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c @@ -36,6 +36,8 @@ #include "o2iblnd.h" +#define MAX_CONN_RACES_BEFORE_ABORT 20 + static void kiblnd_peer_alive(struct kib_peer *peer); static void kiblnd_peer_connect_failed(struct kib_peer *peer, int active, int error); static void kiblnd_init_tx_msg(lnet_ni_t *ni, struct kib_tx *tx, @@ -2405,23 +2407,37 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob) goto failed; } - /* tie-break connection race in favour of the higher NID */ + /* +* Tie-break connection race in favour of the higher NID. +* If we keep running into a race condition multiple times, +* we have to assume that the connection attempt with the +* higher NID is stuck in a connecting state and will never +* recover. As such, we pass through this if-block and let +* the lower NID connection win so we can move forward. +*/ if (peer2->ibp_connecting && - nid < ni->ni_nid) { + nid < ni->ni_nid && peer2->ibp_races < + MAX_CONN_RACES_BEFORE_ABORT) { + peer2->ibp_races++; write_unlock_irqrestore(g_lock, flags); - CWARN("Conn race %s\n", libcfs_nid2str(peer2->ibp_nid)); + CDEBUG(D_NET, "Conn race %s\n", + libcfs_nid2str(peer2->ibp_nid)); kiblnd_peer_decref(peer); rej.ibr_why = IBLND_REJECT_CONN_RACE; goto failed; } - + if (peer2->ibp_races >= MAX_CONN_RACES_BEFORE_ABORT) + CNETERR("Conn race %s: unresolved after %d attempts, letting lower NID win\n", + libcfs_nid2str(peer2->ibp_nid), + MAX_CONN_RACES_BEFORE_ABORT); /** * passive connection is allowed even this peer is waiting for * reconnection. */ peer2->ibp_reconnecting = 0; + peer2->ibp_races = 0; peer2->ibp_accepting++; kiblnd_peer_addref(peer2); -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 71/80] staging: lustre: llite: remove assert for acl refcount
The purpose of this asssert to was to ensure lustre was properly managing its posix_acl access. This test is invalid due to the VFS layer also taking references on the posix_acl. In reality their is no simple way to detect this class of mistakes. Signed-off-by: James Simmons--- drivers/staging/lustre/lustre/llite/llite_lib.c |1 - 1 files changed, 0 insertions(+), 1 deletions(-) diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c index da00fbd..64c8a2b 100644 --- a/drivers/staging/lustre/lustre/llite/llite_lib.c +++ b/drivers/staging/lustre/lustre/llite/llite_lib.c @@ -1247,7 +1247,6 @@ void ll_clear_inode(struct inode *inode) #ifdef CONFIG_FS_POSIX_ACL if (lli->lli_posix_acl) { - LASSERT(atomic_read(>lli_posix_acl->a_refcount) == 1); posix_acl_release(lli->lli_posix_acl); lli->lli_posix_acl = NULL; } -- 1.7.1 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 56/80] staging: lustre: clio: Reduce memory overhead of per-page allocation
From: Jinshan XiongA page in clio used to occupy 584 bytes, which will use size-1024 slab cache. This patch reduces the per-page overhead to 512 bytes so it can use size-512 instead. Signed-off-by: Jinshan Xiong Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4793 Reviewed-on: http://review.whamcloud.com/10070 Reviewed-by: Andreas Dilger Reviewed-by: Bobi Jam Signed-off-by: James Simmons --- drivers/staging/lustre/lustre/include/cl_object.h | 37 +--- drivers/staging/lustre/lustre/llite/vvp_internal.h |6 ++-- .../staging/lustre/lustre/lov/lov_cl_internal.h|4 +- drivers/staging/lustre/lustre/lov/lov_io.c |6 +-- drivers/staging/lustre/lustre/lov/lov_page.c |1 + drivers/staging/lustre/lustre/obdclass/cl_io.c | 10 +- drivers/staging/lustre/lustre/obdclass/cl_page.c | 12 +- drivers/staging/lustre/lustre/osc/osc_internal.h |1 - drivers/staging/lustre/lustre/osc/osc_io.c |7 +++- drivers/staging/lustre/lustre/osc/osc_request.c|6 --- 10 files changed, 26 insertions(+), 64 deletions(-) diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h index 0fa71a5..d269b32 100644 --- a/drivers/staging/lustre/lustre/include/cl_object.h +++ b/drivers/staging/lustre/lustre/include/cl_object.h @@ -690,17 +690,6 @@ enum cl_page_type { }; /** - * Flags maintained for every cl_page. - */ -enum cl_page_flags { - /** -* Set when pagein completes. Used for debugging (read completes at -* most once for a page). -*/ - CPF_READ_COMPLETED = 1 << 0 -}; - -/** * Fields are protected by the lock on struct page, except for atomics and * immutables. * @@ -712,26 +701,23 @@ enum cl_page_flags { struct cl_page { /** Reference counter. */ atomic_t cp_ref; + /** Transfer error. */ + int cp_error; /** An object this page is a part of. Immutable after creation. */ struct cl_object*cp_obj; - /** List of slices. Immutable after creation. */ - struct list_head cp_layers; /** vmpage */ struct page *cp_vmpage; + /** Linkage of pages within group. Pages must be owned */ + struct list_head cp_batch; + /** List of slices. Immutable after creation. */ + struct list_head cp_layers; + /** Linkage of pages within cl_req. */ + struct list_head cp_flight; /** * Page state. This field is const to avoid accidental update, it is * modified only internally within cl_page.c. Protected by a VM lock. */ const enum cl_page_state cp_state; - /** Linkage of pages within group. Protected by cl_page::cp_mutex. */ - struct list_headcp_batch; - /** Mutex serializing membership of a page in a batch. */ - struct mutexcp_mutex; - /** Linkage of pages within cl_req. */ - struct list_head cp_flight; - /** Transfer error. */ - int cp_error; - /** * Page type. Only CPT_TRANSIENT is used so far. Immutable after * creation. @@ -744,10 +730,6 @@ struct cl_page { */ struct cl_io*cp_owner; /** -* Debug information, the task is owning the page. -*/ - struct task_struct *cp_task; - /** * Owning IO request in cl_page_state::CPS_PAGEOUT and * cl_page_state::CPS_PAGEIN states. This field is maintained only in * the top-level pages. Protected by a VM lock. @@ -759,8 +741,6 @@ struct cl_page { struct lu_ref_link cp_obj_ref; /** Link to a queue, for debugging. */ struct lu_ref_link cp_queue_ref; - /** Per-page flags from enum cl_page_flags. Protected by a VM lock. */ - unsigned cp_flags; /** Assigned if doing a sync_io */ struct cl_sync_io *cp_sync_io; }; @@ -2200,6 +2180,7 @@ static inline void cl_object_page_init(struct cl_object *clob, int size) { clob->co_slice_off = cl_object_header(clob)->coh_page_bufsize; cl_object_header(clob)->coh_page_bufsize += cfs_size_round(size); + WARN_ON(cl_object_header(clob)->coh_page_bufsize > 512); } static inline void *cl_object_page_slice(struct cl_object *clob, diff --git a/drivers/staging/lustre/lustre/llite/vvp_internal.h b/drivers/staging/lustre/lustre/llite/vvp_internal.h index 79fc428..99437b8 100644 --- a/drivers/staging/lustre/lustre/llite/vvp_internal.h +++ b/drivers/staging/lustre/lustre/llite/vvp_internal.h @@ -247,9 +247,9 @@ struct vvp_object { */ struct vvp_page { struct cl_page_slice vpg_cl; - int
Re: [PATCH v2] android: binder: fix dangling pointer comparison
On Mon, Aug 15, 2016 at 7:58 AM, Greg Kroah-Hartmanwrote: > On Thu, Jun 16, 2016 at 12:45:33AM +0200, Jann Horn wrote: >> If /dev/binder is opened and the opener process then e.g. calls execve, >> proc->vma_vm_mm will still point to the location of the now-freed >> mm_struct. If the process then calls ioctl(binder_fd, ...), the dangling >> proc->vma_vm_mm pointer will be compared to current->mm. >> >> Let the binder take a reference to the mm_struct to avoid this. >> >> v2: use the right refcounter >> >> Fixes: a906d6931f3ccaf7de805643190765ddd7378e27 > > Nit, the proper way to do this is: > > Fixes: a906d6931f3c ("android: binder: Sanity check at binder ioctl") > > You can get that by doing: > git show -s --abbrev-commit --abbrev=12 --pretty=format:"%h > (\"%s\")%n" a906d6931f3ccaf7de805643190765ddd7378e27 > > Also, I'm guessing this should go to the stable kernels that include the > above patch? > Has this patch been tested? Does it actually fix anything? Changes to vma_vm_mm are normally protected by the mmap_sem it contains, and is set to NULL in the vma->vm_ops->close call on the binder vma. The vma_vm_mm pointer is only used to check if the mm struct that the binder driver was mmapped into is the same mm struct that the driver just locked (in binder_update_page_range). I don't think change description here is accurate. If a process calls execve, the old vma should be closed (and vma_vm_mm cleared) before the mm_struct that it belongs too is freed. -- Arve Hjønnevåg ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH] Staging: android: ion: ion_heap.c: fix parenthesis alignment
This fixes the checkpatch.pl "Alignment should match open parenthesis" issues in ion_heap.c. Signed-off-by: Ben LeMasurier--- drivers/staging/android/ion/ion_heap.c | 10 +- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/drivers/staging/android/ion/ion_heap.c b/drivers/staging/android/ion/ion_heap.c index ca15a87..4e5c0f1 100644 --- a/drivers/staging/android/ion/ion_heap.c +++ b/drivers/staging/android/ion/ion_heap.c @@ -93,7 +93,7 @@ int ion_heap_map_user(struct ion_heap *heap, struct ion_buffer *buffer, } len = min(len, remainder); ret = remap_pfn_range(vma, addr, page_to_pfn(page), len, - vma->vm_page_prot); + vma->vm_page_prot); if (ret) return ret; addr += len; @@ -116,7 +116,7 @@ static int ion_heap_clear_pages(struct page **pages, int num, pgprot_t pgprot) } static int ion_heap_sglist_zero(struct scatterlist *sgl, unsigned int nents, - pgprot_t pgprot) + pgprot_t pgprot) { int p = 0; int ret = 0; @@ -181,7 +181,7 @@ size_t ion_heap_freelist_size(struct ion_heap *heap) } static size_t _ion_heap_freelist_drain(struct ion_heap *heap, size_t size, - bool skip_pools) + bool skip_pools) { struct ion_buffer *buffer; size_t total_drained = 0; @@ -266,7 +266,7 @@ int ion_heap_init_deferred_free(struct ion_heap *heap) } static unsigned long ion_heap_shrink_count(struct shrinker *shrinker, - struct shrink_control *sc) + struct shrink_control *sc) { struct ion_heap *heap = container_of(shrinker, struct ion_heap, shrinker); @@ -279,7 +279,7 @@ static unsigned long ion_heap_shrink_count(struct shrinker *shrinker, } static unsigned long ion_heap_shrink_scan(struct shrinker *shrinker, - struct shrink_control *sc) + struct shrink_control *sc) { struct ion_heap *heap = container_of(shrinker, struct ion_heap, shrinker); -- 2.9.3 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [PATCH 2/9] staging: most: hdm-usb: fix race between enqueue and most_stop_enqueue
On Tue, Aug 16, 2016 at 01:38:50PM +0200, Christian Gromm wrote: > On Mon, 15 Aug 2016 20:39:38 +0200 > Greg KHwrote: > > > On Mon, Jul 18, 2016 at 05:25:28PM +0200, Christian Gromm wrote: > > > The "broken in pipe" handler of the USB-HDM calls most_stop_enqueue() to > > > stop the MBO traffic before returning all MBOs back to the Mostcore. As > > > the enqueue() call from the Mostcore may run in parallel with the > > > most_stop_enqueue(), the HDM may run into the inconsistent state and > > > crash the kernel. > > > > > > This patch synchronizes enqueue(), most_stop_enqueue() and > > > most_resume_enqueue() with a mutex, hence avoiding the race condition. > > > > > > Signed-off-by: Andrey Shvetsov > > > Signed-off-by: Christian Gromm > > > --- > > > drivers/staging/most/hdm-usb/hdm_usb.c | 3 +- > > > drivers/staging/most/mostcore/core.c | 53 > > > -- > > > 2 files changed, 38 insertions(+), 18 deletions(-) > > > > Doesn't apply to my tree anymore. Can you fix this up and resend it and > > the rest in this series so I can apply them? > > According to git you didn't apply the patchset I sent in on Friday 6/17/16, > did you? > > >>> "[PATCH 0/9] staging: most: fix hdm-usb issues" <<< > > Those patches need to be applied before the latest set. All of the "staging: most:" patches are now applied to my tree, can you sync up with it and resend anything reamaining? thanks, greg k-h ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH v3 0/4] Drivers: hv: balloon: fix WS2012 memory hotplug issues and do some cleanup
Changes since v2: - check for gaps when onlining 'tails' in hv_bring_pgs_online() too [Alex Ng] Changes since v1: - Keep ol_waitevent and wait when kernel memory onlining is disabled [Alex Ng] Crashes with Hyper-V balloon driver are reported with WS2012 (non-R2), hosts I was able to identify two issues which I fix with first two patches of this series. Patches 3 removes wait on ol_waitevent when we have in in-kernel memory onlining, patch 4 gets rid of ha_region_mutex by doing the locking fine-grained with a spinlock. Vitaly Kuznetsov (4): Drivers: hv: balloon: keep track of where ha_region starts Drivers: hv: balloon: account for gaps in hot add regions Drivers: hv: balloon: don't wait for ol_waitevent when memhp_auto_online is enabled Drivers: hv: balloon: replace ha_region_mutex with spinlock drivers/hv/hv_balloon.c | 241 +++- 1 file changed, 156 insertions(+), 85 deletions(-) -- 2.7.4 ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel