Re: [PATCH 1/2] Convert target drivers to use sbitmap
Hi, Matthew Wilcox writes: > From: Matthew Wilcox > > The sbitmap and the percpu_ida perform essentially the same task, > allocating tags for commands. Since the sbitmap is more used than > the percpu_ida, convert the percpu_ida users to the sbitmap API. > > Signed-off-by: Matthew Wilcox > --- [...] > drivers/usb/gadget/function/f_tcm.c | 8 +++--- for drivers/usb/gadget/function/f_tcm.c Acked-by: Felipe Balbi -- balbi signature.asc Description: PGP signature
Re: [RFC PATCH V2] scsi: ufs: Add specific callback for setting DMA mask
Ping !!!
On Thu, Mar 8, 2018 at 4:33 PM, Alim Akhtar wrote:
> Currently DMA mask for UFS HCI is set by reading CAP register's
> [64AS] bit. Some HCI controller like Exynos support 36-bit bus address.
> This works perfectly fine with DMA mask set as 64 in case there is no
> IOMMU attached to HCI.
> In case if HCI is behind an IOMMU, setting DMA mask as 64 bit won't
> work as HCI has only 36bit addressing and SMMU has created mapping of
> 64 bit and as the device truncates the address, its mapping will not
> be found by iommu.
> To resolve such issues, let the variant driver sets its own DMA mask.
>
> Signed-off-by: Alim Akhtar
> ---
> drivers/scsi/ufs/ufshcd.c | 3 +++
> drivers/scsi/ufs/ufshcd.h | 2 ++
> 2 files changed, 5 insertions(+)
>
> I am not sure if there are other ways available to handle such cases.
> The IOMMU I am talking about is arm-smmu and it DT binding does not
> give much idea about handling such cases.
> Have tested this patch with HCI controller with IOMMU attached.
>
> Changes Since V1:
> - Fixed build issue as reported by Kbuild test robot.
>
> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> index a355d98..9a1374e 100644
> --- a/drivers/scsi/ufs/ufshcd.c
> +++ b/drivers/scsi/ufs/ufshcd.c
> @@ -7781,6 +7781,9 @@ EXPORT_SYMBOL_GPL(ufshcd_dealloc_host);
> */
> static int ufshcd_set_dma_mask(struct ufs_hba *hba)
> {
> + if (hba->vops && hba->vops->set_dma_mask)
> + return hba->vops->set_dma_mask(hba);
> +
> if (hba->capabilities & MASK_64_ADDRESSING_SUPPORT) {
> if (!dma_set_mask_and_coherent(hba->dev, DMA_BIT_MASK(64)))
> return 0;
> diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
> index 1332e54..89c6dae 100644
> --- a/drivers/scsi/ufs/ufshcd.h
> +++ b/drivers/scsi/ufs/ufshcd.h
> @@ -297,6 +297,7 @@ struct ufs_pwr_mode_info {
> * @resume: called during host controller PM callback
> * @dbg_register_dump: used to dump controller debug information
> * @phy_initialization: used to initialize phys
> + * @set_dma_mask: used to set variant specific DMA mask
> */
> struct ufs_hba_variant_ops {
> const char *name;
> @@ -325,6 +326,7 @@ struct ufs_hba_variant_ops {
> int (*resume)(struct ufs_hba *, enum ufs_pm_op);
> void(*dbg_register_dump)(struct ufs_hba *hba);
> int (*phy_initialization)(struct ufs_hba *);
> + int (*set_dma_mask)(struct ufs_hba *hba);
> };
>
> /* clock gating state */
> --
> 2.7.4
>
--
Regards,
Alim
Re: [PATCH 0/4] Add required changes to ufshcd to support exynos ufs hci
Hi All, Any thought on this patch set? On Sun, May 6, 2018 at 3:44 PM, Alim Akhtar wrote: > Hi All > > These patches are part of a larger patch series [1] which attempts upstreaming > EXYNOS UFS driver support. There was not much activities after v5 of that > series. In between I saw there were other teams in Samsung tried upstreaming > the same, but that has not really gone anywhere. > I have taken this task again and here is another attempt to upstream > exynos-ufs > driver support. I have divided the patches into two series, one which adds > required infra in the ufshcd core needed by exynos-ufs driver and other part > will have actual exynos-ufs driver. Splitting this has a advantage of less > reviewing over head. > > I am floating these as a new patch set. > > [1] https://www.spinics.net/lists/linux-scsi/msg90292.html > > These patches are based on mainline v4.17-rc3. > > Alim Akhtar (4): > scsi: ufs: add quirk to fix mishandling utrlclr/utmrlclr > scsi: ufs: add quirk not to allow reset of interrupt aggregation > scsi: ufs: add quirk to enable host controller without hce > scsi: ufs: make ufshcd_config_pwr_mode of non-static func > > drivers/scsi/ufs/ufshcd.c | 104 > ++ > drivers/scsi/ufs/ufshcd.h | 18 > 2 files changed, 114 insertions(+), 8 deletions(-) > > -- > 2.7.4 > -- Regards, Alim
drivers/scsi/libiscsi.c : Fix iscsi_data_xmit function design defects cause a time-out
problem description :
Target ： created a raid1 array use 8 disks,
Initiator :
The iscsi negotiation result :
ImmediateData=Yes
InitialR2T=Yes
FirstBurstLength=65536
MaxBurstLength=262144
MaxRecvDataSegmentLength=262144
In Initiator, we perform a iozone test (iozone -a -i 0 -i 2 -n 64g -g
64g -y 64K -q 64K -f /mnt/iot) on it,
and When running to random tests, the target report: "Unable to recover
from DataOut timeout while in ERL=0",
and the Initiator report : "connection1:0: detected conn error (1020)"
problem analysis :
Because FirstBurstLength=65536, In a 64k random write test，
A large amount of data is sent to the target through the Initiator
"conn->cmdqueue" queue,
If a merge request occurs at this time（The probability of merging in random
tests is very low),
and it will lead to have some data sent to the target through the
"conn->requeue" queue.
Because A large amount of data is sent to the target through the
"conn->cmdqueue" queue,
from the design of the iscsi_data_xmit function,
the "While (! List_empty (& conn->cmdqueue))" will always be true,
and ultimately make the "conn->requeue" queue can not be processed in time,
so the target will can not get the report, and cause a time-out.
and also, we can open the kernel debug and Extend the timeout, we can get
dmesg ( 64k random write test and occurs a merge request ) :
17:34:41 kylinOS kernel: Got SCSI Command, ITT: 0x0015,
CmdSN: 0x8f500e00, ExpXferLen: 131072, Length: 65536, CID: 0
..
17:34:41 kylinOS kernel: Built R2T, ITT: 0x0015, TTT:
0x000ead3e, StatSN: 0x111c0410, R2TSN: 0x, Offset: 65536, DDTL: 65536,
CID: 0
17:34:41 kylinOS kernel: ret: 48, sent data: 48
17:34:41 kylinOS kernel: Starting DataOUT timer for ITT:
0x0015 on CID: 0, timeout:203.
17:34:41 kylinOS kernel: Updated MaxCmdSN to 0x000e50c4
...
17:34:47 kylinOS kernel: Got DataOut ITT: 0x0015, TTT:
0x3ead0e00, DataSN: 0x, Offset: 65536, Length: 65536, CID: 0
17:34:47 kylinOS kernel: Updated DataOUT timer for ITT:
0x0015, timeout:203.
..
17:34:47 kylinOS kernel: Stopped DataOUT Timer for ITT:
0x0015
as you see, the merge request(128K) "ITT: 0x0015", Delay 6S to complete
So I think the iscsi_data_xmit function need to modify to reduce I/O delay
and unnecessary time-out.
problem solutions ：
We should add a timeout mechanism to the while loop :
diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi/libiscsi.c
index c051694..77108fe 100644
--- a/drivers/scsi/libiscsi.c
+++ b/drivers/scsi/libiscsi.c
@@ -1492,6 +1492,8 @@ EXPORT_SYMBOL_GPL(iscsi_requeue_task);
static int iscsi_data_xmit(struct iscsi_conn *conn)
{
struct iscsi_task *task;
+ unsigned long cmd_timeout = 0;
+ unsigned long req_timeout = 0;
int rc = 0;
spin_lock_bh(&conn->session->frwd_lock);
@@ -1530,8 +1532,11 @@ check_mgmt:
goto done;
}
+check_cmd:
/* process pending command queue */
- while (!list_empty(&conn->cmdqueue)) {
+ if (!cmd_timeout)
+ cmd_timeout = jiffies + HZ/2; /* timeout in 0.5s */
+ while (!list_empty(&conn->cmdqueue) && time_before(jiffies,
cmd_timeout)) {
conn->task = list_entry(conn->cmdqueue.next, struct iscsi_task,
running);
list_del_init(&conn->task->running);
@@ -1562,7 +1567,10 @@ check_mgmt:
goto check_mgmt;
}
- while (!list_empty(&conn->requeue)) {
+check_req:
+ if (!req_timeout)
+ req_timeout = jiffies + HZ/2; /* timeout in 0.5s */
+ while (!list_empty(&conn->requeue) && time_before(jiffies,
req_timeout)) {
/*
* we always do fastlogout - conn stop code will clean up.
*/
@@ -1583,6 +1591,15 @@ check_mgmt:
if (!list_empty(&conn->mgmtqueue))
goto check_mgmt;
}
+
+ /* Check whether there are data needs to be sent */
+ cmd_timeout = 0;
+ req_timeout = 0;
+ if (!list_empty(&conn->cmdqueue))
+ goto check_cmd;
+ if (!list_empty(&conn->requeue))
+ goto check_req;
+
spin_unlock_bh(&conn->session->frwd_lock);
return -ENODATA;
Do you have any idea on this problem?
Looking forward to your answer!
thank you!
[PATCH v4] target: transport should handle st FM/EOM/ILI reads
When a tape drive is exported via LIO using the pscsi module, a read that
requests more bytes per block than the tape can supply returns an empty
buffer. This is because the pscsi pass-through target module sees the
"ILI" illegal length bit set and thinks there is no reason to return
the data.
This is a long-standing transport issue, since it assumes that no data
need be returned under a check condition, which isn't always the case
for tape.
Add in a check for tape reads with the ILI, EOM, or FM bits set,
with a sense code of NO_SENSE, treating such cases as if the read
succeeded. The layered tape driver then "does the right thing" when
it gets such a response.
Changes from v3:
- cleaned up comment
- Added residual handling
Changes from v2:
- Cleaned up subject line and bug text formatting
- Removed unneeded inner braces
- Removed ugly goto
- Also updated the "queue full" path to handle this case
Changes from RFC:
- Moved ugly code from transport to pscsi module
- Added checking EOM and FM bits, as well as ILI
- fixed malformed patch
- Clarified description a bit
Signed-off-by: Lee Duncan
Signed-off-by: Bodo Stroesser )
Reviewed-by: Hannes Reinecke
---
drivers/target/target_core_pscsi.c | 26 ++--
drivers/target/target_core_transport.c | 43 +-
include/target/target_core_base.h | 1 +
3 files changed, 62 insertions(+), 8 deletions(-)
diff --git a/drivers/target/target_core_pscsi.c
b/drivers/target/target_core_pscsi.c
index 0d99b242e82e..f31215b1d009 100644
--- a/drivers/target/target_core_pscsi.c
+++ b/drivers/target/target_core_pscsi.c
@@ -689,8 +689,29 @@ static void pscsi_complete_cmd(struct se_cmd *cmd, u8
scsi_status,
}
after_mode_select:
- if (scsi_status == SAM_STAT_CHECK_CONDITION)
+ if (scsi_status == SAM_STAT_CHECK_CONDITION) {
transport_copy_sense_to_cmd(cmd, req_sense);
+
+ /*
+* check for TAPE device reads with
+* FM/EOM/ILI set, so that we can get data
+* back despite framework assumption that a
+* check condition means there is no data
+*/
+ if (sd->type == TYPE_TAPE &&
+ cmd->data_direction == DMA_FROM_DEVICE) {
+ /*
+* is sense data valid, fixed format,
+* and have FM, EOM, or ILI set?
+*/
+ if (req_sense[0] == 0xf0 && /* valid, fixed format
*/
+ req_sense[2] & 0xe0 && /* FM, EOM, or ILI */
+ (req_sense[2] & 0xf) == 0) { /* key==NO_SENSE */
+ pr_debug("Tape FM/EOM/ILI status detected.
Treat as normal read.\n");
+ cmd->se_cmd_flags |= SCF_TREAT_READ_AS_NORMAL;
+ }
+ }
+ }
}
enum {
@@ -1061,7 +1082,8 @@ static void pscsi_req_done(struct request *req,
blk_status_t status)
switch (host_byte(result)) {
case DID_OK:
- target_complete_cmd(cmd, scsi_status);
+ target_complete_cmd_with_length(cmd, scsi_status,
+ cmd->data_length - scsi_req(req)->resid_len);
break;
default:
pr_debug("PSCSI Host Byte exception at cmd: %p CDB:"
diff --git a/drivers/target/target_core_transport.c
b/drivers/target/target_core_transport.c
index 74b646f165d4..791ff9ba2bc6 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target/target_core_transport.c
@@ -779,7 +779,9 @@ EXPORT_SYMBOL(target_complete_cmd);
void target_complete_cmd_with_length(struct se_cmd *cmd, u8 scsi_status, int
length)
{
- if (scsi_status == SAM_STAT_GOOD && length < cmd->data_length) {
+ if ((scsi_status == SAM_STAT_GOOD ||
+cmd->se_cmd_flags & SCF_TREAT_READ_AS_NORMAL) &&
+ length < cmd->data_length) {
if (cmd->se_cmd_flags & SCF_UNDERFLOW_BIT) {
cmd->residual_count += cmd->data_length - length;
} else {
@@ -2084,12 +2086,24 @@ static void transport_complete_qf(struct se_cmd *cmd)
goto queue_status;
}
- if (cmd->se_cmd_flags & SCF_TRANSPORT_TASK_SENSE)
+ /*
+* Check if we need to send a sense buffer from
+* the struct se_cmd in question. We do NOT want
+* to take this path of the IO has been marked as
+* needing to be treated like a "normal read". This
+* is the case if it's a tape read, and either the
+* FM, EOM, or ILI bits are set, but there is no
+* sense data.
+*/
+ if (!(cmd->se_cmd_flags & SCF_TREAT_READ_AS_NORMAL) &&
+ cmd->se_cmd_flags & SCF_TRANSPORT_TASK_SENSE)
goto queue_status;
switch (cmd->data_direction) {
case DMA_FROM_DEVICE:
- if (c
[Bug 199703] HPSA blocking boot on HP smart Array P400
https://bugzilla.kernel.org/show_bug.cgi?id=199703 --- Comment #5 from Roberto M. (roby_program...@fastwebnet.it) --- Downloaded from kernel.org 4.13.16 kernel, build and tested it works uname -a Linux Server-Ubuntu-01 4.13.16 #1 SMP Mon May 14 19:51:55 CEST 2018 x86_64 x86_64 x86_64 GNU/Linux Downloaded from kernel.org 4.13.1 kernel, build and tested, it doesn't boot, same result as picture I am not a programmer, but I think bug was introduced in 4.14.1 version, I can test patch or other kernels if necessary... I am not able to check changelog file to see what is modified -- You are receiving this mail because: You are the assignee for the bug.
Re: [PATCH 1/2] Convert target drivers to use sbitmap
On 5/15/18 10:11 AM, Jens Axboe wrote: > On 5/15/18 10:00 AM, Matthew Wilcox wrote: >> From: Matthew Wilcox >> >> The sbitmap and the percpu_ida perform essentially the same task, >> allocating tags for commands. Since the sbitmap is more used than >> the percpu_ida, convert the percpu_ida users to the sbitmap API. > > It should also be the same performance as percpu_ida in light use, and > performs much better at > 50% utilization of the tag space. I think > that's better justification than "more used than". Had to search long and hard for the perf numbers I did for percpu_ida on higher utilization, but here it is: https://lkml.org/lkml/2014/4/22/553 -- Jens Axboe
Re: [PATCH 1/2] Convert target drivers to use sbitmap
On 5/15/18 10:00 AM, Matthew Wilcox wrote:
> From: Matthew Wilcox
>
> The sbitmap and the percpu_ida perform essentially the same task,
> allocating tags for commands. Since the sbitmap is more used than
> the percpu_ida, convert the percpu_ida users to the sbitmap API.
It should also be the same performance as percpu_ida in light use, and
performs much better at > 50% utilization of the tag space. I think
that's better justification than "more used than".
> diff --git a/drivers/target/iscsi/iscsi_target_util.c
> b/drivers/target/iscsi/iscsi_target_util.c
> index 4435bf374d2d..28bcffae609f 100644
> --- a/drivers/target/iscsi/iscsi_target_util.c
> +++ b/drivers/target/iscsi/iscsi_target_util.c
> @@ -17,7 +17,7 @@
>
> **/
>
> #include
> -#include
> +#include
> #include /* ipv6_addr_equal() */
> #include
> #include
> @@ -147,6 +147,28 @@ void iscsit_free_r2ts_from_list(struct iscsi_cmd *cmd)
> spin_unlock_bh(&cmd->r2t_lock);
> }
>
> +int iscsit_wait_for_tag(struct se_session *se_sess, int state, int *cpup)
> +{
> + int tag = -1;
> + DEFINE_WAIT(wait);
> + struct sbq_wait_state *ws;
> +
> + if (state == TASK_RUNNING)
> + return tag;
> +
> + ws = &se_sess->sess_tag_pool.ws[0];
> + for (;;) {
> + prepare_to_wait_exclusive(&ws->wait, &wait, state);
> + if (signal_pending_state(state, current))
> + break;
> + schedule();
> + tag = sbitmap_queue_get(&se_sess->sess_tag_pool, cpup);
> + }
> +
> + finish_wait(&ws->wait, &wait);
> + return tag;
> +}
Seems like that should be:
ws = &se_sess->sess_tag_pool.ws[0];
for (;;) {
prepare_to_wait_exclusive(&ws->wait, &wait, state);
if (signal_pending_state(state, current))
break;
tag = sbitmap_queue_get(&se_sess->sess_tag_pool, cpup);
if (tag != -1)
break;
schedule();
}
finish_wait(&ws->wait, &wait);
return tag;
> /*
> * May be called from software interrupt (timer) context for allocating
> * iSCSI NopINs.
> @@ -155,9 +177,11 @@ struct iscsi_cmd *iscsit_allocate_cmd(struct iscsi_conn
> *conn, int state)
> {
> struct iscsi_cmd *cmd;
> struct se_session *se_sess = conn->sess->se_sess;
> - int size, tag;
> + int size, tag, cpu;
>
> - tag = percpu_ida_alloc(&se_sess->sess_tag_pool, state);
> + tag = sbitmap_queue_get(&se_sess->sess_tag_pool, &cpu);
> + if (tag < 0)
> + tag = iscsit_wait_for_tag(se_sess, state, &cpu);
> if (tag < 0)
> return NULL;
Might make sense to just roll the whole thing into iscsi_get_tag(), that
would be cleaner.
sbitmap should provide a helper for that, but we can do that cleanup
later. That would encapsulate things like the per-cpu caching hint too,
for instance.
Rest looks fine to me.
--
Jens Axboe
[PATCH 1/2] Convert target drivers to use sbitmap
From: Matthew Wilcox
The sbitmap and the percpu_ida perform essentially the same task,
allocating tags for commands. Since the sbitmap is more used than
the percpu_ida, convert the percpu_ida users to the sbitmap API.
Signed-off-by: Matthew Wilcox
---
drivers/scsi/qla2xxx/qla_target.c| 16 ++-
drivers/target/iscsi/iscsi_target_util.c | 34 +---
drivers/target/sbp/sbp_target.c | 8 +++---
drivers/target/target_core_transport.c | 5 ++--
drivers/target/tcm_fc/tfc_cmd.c | 11
drivers/usb/gadget/function/f_tcm.c | 8 +++---
drivers/vhost/scsi.c | 9 ---
drivers/xen/xen-scsiback.c | 8 +++---
include/target/iscsi/iscsi_target_core.h | 1 +
include/target/target_core_base.h| 5 ++--
10 files changed, 73 insertions(+), 32 deletions(-)
diff --git a/drivers/scsi/qla2xxx/qla_target.c
b/drivers/scsi/qla2xxx/qla_target.c
index 025dc2d3f3de..cdf671c2af61 100644
--- a/drivers/scsi/qla2xxx/qla_target.c
+++ b/drivers/scsi/qla2xxx/qla_target.c
@@ -3719,7 +3719,8 @@ void qlt_free_cmd(struct qla_tgt_cmd *cmd)
return;
}
cmd->jiffies_at_free = get_jiffies_64();
- percpu_ida_free(&sess->se_sess->sess_tag_pool, cmd->se_cmd.map_tag);
+ sbitmap_queue_clear(&sess->se_sess->sess_tag_pool, cmd->se_cmd.map_tag,
+ cmd->se_cmd.map_cpu);
}
EXPORT_SYMBOL(qlt_free_cmd);
@@ -4084,7 +4085,8 @@ static void __qlt_do_work(struct qla_tgt_cmd *cmd)
qlt_send_term_exchange(qpair, NULL, &cmd->atio, 1, 0);
qlt_decr_num_pend_cmds(vha);
- percpu_ida_free(&sess->se_sess->sess_tag_pool, cmd->se_cmd.map_tag);
+ sbitmap_queue_clear(&sess->se_sess->sess_tag_pool, cmd->se_cmd.map_tag,
+ cmd->se_cmd.map_cpu);
spin_unlock_irqrestore(qpair->qp_lock_ptr, flags);
spin_lock_irqsave(&ha->tgt.sess_lock, flags);
@@ -4215,9 +4217,9 @@ static struct qla_tgt_cmd *qlt_get_tag(scsi_qla_host_t
*vha,
{
struct se_session *se_sess = sess->se_sess;
struct qla_tgt_cmd *cmd;
- int tag;
+ int tag, cpu;
- tag = percpu_ida_alloc(&se_sess->sess_tag_pool, TASK_RUNNING);
+ tag = sbitmap_queue_get(&se_sess->sess_tag_pool, &cpu);
if (tag < 0)
return NULL;
@@ -4230,6 +4232,7 @@ static struct qla_tgt_cmd *qlt_get_tag(scsi_qla_host_t
*vha,
qlt_incr_num_pend_cmds(vha);
cmd->vha = vha;
cmd->se_cmd.map_tag = tag;
+ cmd->se_cmd.map_cpu = cpu;
cmd->sess = sess;
cmd->loop_id = sess->loop_id;
cmd->conf_compl_supported = sess->conf_compl_supported;
@@ -5212,7 +5215,7 @@ qlt_alloc_qfull_cmd(struct scsi_qla_host *vha,
struct fc_port *sess;
struct se_session *se_sess;
struct qla_tgt_cmd *cmd;
- int tag;
+ int tag, cpu;
unsigned long flags;
if (unlikely(tgt->tgt_stop)) {
@@ -5244,7 +5247,7 @@ qlt_alloc_qfull_cmd(struct scsi_qla_host *vha,
se_sess = sess->se_sess;
- tag = percpu_ida_alloc(&se_sess->sess_tag_pool, TASK_RUNNING);
+ tag = sbitmap_queue_get(&se_sess->sess_tag_pool, &cpu);
if (tag < 0)
return;
@@ -5275,6 +5278,7 @@ qlt_alloc_qfull_cmd(struct scsi_qla_host *vha,
cmd->reset_count = ha->base_qpair->chip_reset;
cmd->q_full = 1;
cmd->qpair = ha->base_qpair;
+ cmd->se_cmd.map_cpu = cpu;
if (qfull) {
cmd->q_full = 1;
diff --git a/drivers/target/iscsi/iscsi_target_util.c
b/drivers/target/iscsi/iscsi_target_util.c
index 4435bf374d2d..28bcffae609f 100644
--- a/drivers/target/iscsi/iscsi_target_util.c
+++ b/drivers/target/iscsi/iscsi_target_util.c
@@ -17,7 +17,7 @@
**/
#include
-#include
+#include
#include /* ipv6_addr_equal() */
#include
#include
@@ -147,6 +147,28 @@ void iscsit_free_r2ts_from_list(struct iscsi_cmd *cmd)
spin_unlock_bh(&cmd->r2t_lock);
}
+int iscsit_wait_for_tag(struct se_session *se_sess, int state, int *cpup)
+{
+ int tag = -1;
+ DEFINE_WAIT(wait);
+ struct sbq_wait_state *ws;
+
+ if (state == TASK_RUNNING)
+ return tag;
+
+ ws = &se_sess->sess_tag_pool.ws[0];
+ for (;;) {
+ prepare_to_wait_exclusive(&ws->wait, &wait, state);
+ if (signal_pending_state(state, current))
+ break;
+ schedule();
+ tag = sbitmap_queue_get(&se_sess->sess_tag_pool, cpup);
+ }
+
+ finish_wait(&ws->wait, &wait);
+ return tag;
+}
+
/*
* May be called from software interrupt (timer) context for allocating
* iSCSI NopINs.
@@ -155,9 +177,11 @@ struct iscsi_cmd *iscsit_allocate_cmd(struct iscsi_conn
*conn, int state)
{
struct iscsi_cmd *cmd;
struct se_session *se_sess = conn->sess->se_sess;
-
[PATCH 0/2] Use sbitmap instead of percpu_ida
From: Matthew Wilcox This is a pretty rough-and-ready conversion of the target drivers from using percpu_ida to sbitmap. It compiles; I don't have a target setup, so it's completely untested. I haven't tried to do anything particularly clever here, so it's possible that, for example, the wait queue in iscsi_target_util could be more clever, like the block layer uses multiple wait queues to avoid pingpongs. Or maybe we could figure out a way to not store the CPU that the ID was allocated on, or perhaps the options I specified to sbitmap_queue_init() are suboptimal. Patch 2 isn't interesting; it just deletes the implementation. Patch 1 will be where all the action is. Matthew Wilcox (2): Convert target drivers to use sbitmap Remove percpu_ida drivers/scsi/qla2xxx/qla_target.c| 16 +- drivers/target/iscsi/iscsi_target_util.c | 34 +- drivers/target/sbp/sbp_target.c | 8 +- drivers/target/target_core_transport.c | 5 +- drivers/target/tcm_fc/tfc_cmd.c | 11 +- drivers/usb/gadget/function/f_tcm.c | 8 +- drivers/vhost/scsi.c | 9 +- drivers/xen/xen-scsiback.c | 8 +- include/linux/percpu_ida.h | 83 - include/target/iscsi/iscsi_target_core.h | 1 + include/target/target_core_base.h| 5 +- lib/Makefile | 2 +- lib/percpu_ida.c | 391 --- 13 files changed, 74 insertions(+), 507 deletions(-) delete mode 100644 include/linux/percpu_ida.h delete mode 100644 lib/percpu_ida.c -- 2.17.0
[PATCH 2/2] Remove percpu_ida
From: Matthew Wilcox
With its one user gone, remove the library code.
Signed-off-by: Matthew Wilcox
---
include/linux/percpu_ida.h | 83
lib/Makefile | 2 +-
lib/percpu_ida.c | 391 -
3 files changed, 1 insertion(+), 475 deletions(-)
delete mode 100644 include/linux/percpu_ida.h
delete mode 100644 lib/percpu_ida.c
diff --git a/include/linux/percpu_ida.h b/include/linux/percpu_ida.h
deleted file mode 100644
index 07d78e4653bc..
--- a/include/linux/percpu_ida.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0 */
-#ifndef __PERCPU_IDA_H__
-#define __PERCPU_IDA_H__
-
-#include
-#include
-#include
-#include
-#include
-#include
-#include
-
-struct percpu_ida_cpu;
-
-struct percpu_ida {
- /*
-* number of tags available to be allocated, as passed to
-* percpu_ida_init()
-*/
- unsignednr_tags;
- unsignedpercpu_max_size;
- unsignedpercpu_batch_size;
-
- struct percpu_ida_cpu __percpu *tag_cpu;
-
- /*
-* Bitmap of cpus that (may) have tags on their percpu freelists:
-* steal_tags() uses this to decide when to steal tags, and which cpus
-* to try stealing from.
-*
-* It's ok for a freelist to be empty when its bit is set - steal_tags()
-* will just keep looking - but the bitmap _must_ be set whenever a
-* percpu freelist does have tags.
-*/
- cpumask_t cpus_have_tags;
-
- struct {
- spinlock_t lock;
- /*
-* When we go to steal tags from another cpu (see steal_tags()),
-* we want to pick a cpu at random. Cycling through them every
-* time we steal is a bit easier and more or less equivalent:
-*/
- unsignedcpu_last_stolen;
-
- /* For sleeping on allocation failure */
- wait_queue_head_t wait;
-
- /*
-* Global freelist - it's a stack where nr_free points to the
-* top
-*/
- unsignednr_free;
- unsigned*freelist;
- } cacheline_aligned_in_smp;
-};
-
-/*
- * Number of tags we move between the percpu freelist and the global freelist
at
- * a time
- */
-#define IDA_DEFAULT_PCPU_BATCH_MOVE32U
-/* Max size of percpu freelist, */
-#define IDA_DEFAULT_PCPU_SIZE ((IDA_DEFAULT_PCPU_BATCH_MOVE * 3) / 2)
-
-int percpu_ida_alloc(struct percpu_ida *pool, int state);
-void percpu_ida_free(struct percpu_ida *pool, unsigned tag);
-
-void percpu_ida_destroy(struct percpu_ida *pool);
-int __percpu_ida_init(struct percpu_ida *pool, unsigned long nr_tags,
- unsigned long max_size, unsigned long batch_size);
-static inline int percpu_ida_init(struct percpu_ida *pool, unsigned long
nr_tags)
-{
- return __percpu_ida_init(pool, nr_tags, IDA_DEFAULT_PCPU_SIZE,
- IDA_DEFAULT_PCPU_BATCH_MOVE);
-}
-
-typedef int (*percpu_ida_cb)(unsigned, void *);
-int percpu_ida_for_each_free(struct percpu_ida *pool, percpu_ida_cb fn,
- void *data);
-
-unsigned percpu_ida_free_tags(struct percpu_ida *pool, int cpu);
-#endif /* __PERCPU_IDA_H__ */
diff --git a/lib/Makefile b/lib/Makefile
index ce20696d5a92..7626dece1d27 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -39,7 +39,7 @@ obj-y += bcd.o div64.o sort.o parser.o debug_locks.o
random32.o \
bust_spinlocks.o kasprintf.o bitmap.o scatterlist.o \
gcd.o lcm.o list_sort.o uuid.o flex_array.o iov_iter.o clz_ctz.o \
bsearch.o find_bit.o llist.o memweight.o kfifo.o \
-percpu-refcount.o percpu_ida.o rhashtable.o reciprocal_div.o \
+percpu-refcount.o rhashtable.o reciprocal_div.o \
once.o refcount.o usercopy.o errseq.o bucket_locks.o
obj-$(CONFIG_STRING_SELFTEST) += test_string.o
obj-y += string_helpers.o
diff --git a/lib/percpu_ida.c b/lib/percpu_ida.c
deleted file mode 100644
index 6016f1deb1f5..
--- a/lib/percpu_ida.c
+++ /dev/null
@@ -1,391 +0,0 @@
-/*
- * Percpu IDA library
- *
- * Copyright (C) 2013 Datera, Inc. Kent Overstreet
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation; either version 2, or (at
- * your option) any later version.
- *
- * 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 for more details.
- */
-
-#include
-#include
-#include
-#include
-#include
-#include
-#include
-#include
-#include
-#include
-#include
-#include
-#include
-
-struct
[GIT PULL] SCSI fixes for 4.17-rc5
Two small driver fixes: aacraid to fix an unknown IU type on task
management functions which causes a firmware fault and vmw_pvscsi to
change a return code to retry the operation instead of causing an
immediate error
The patch is available here:
git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi.git scsi-fixes
The short changelog is:
Dave Carroll (1):
scsi: aacraid: Correct hba_send to include iu_type
Jim Gill (1):
scsi: vmw-pvscsi: return DID_BUS_BUSY for adapter-initated aborts
And the diffstat:
drivers/scsi/aacraid/commsup.c | 8
drivers/scsi/vmw_pvscsi.c | 2 +-
2 files changed, 5 insertions(+), 5 deletions(-)
With full diff below.
James
---
diff --git a/drivers/scsi/aacraid/commsup.c
b/drivers/scsi/aacraid/commsup.c
index 0156c9623c35..d62ddd63f4fe 100644
--- a/drivers/scsi/aacraid/commsup.c
+++ b/drivers/scsi/aacraid/commsup.c
@@ -724,6 +724,8 @@ int aac_hba_send(u8 command, struct fib *fibptr,
fib_callback callback,
int wait;
unsigned long flags = 0;
unsigned long mflags = 0;
+ struct aac_hba_cmd_req *hbacmd = (struct aac_hba_cmd_req *)
+ fibptr->hw_fib_va;
fibptr->flags = (FIB_CONTEXT_FLAG |
FIB_CONTEXT_FLAG_NATIVE_HBA);
if (callback) {
@@ -734,11 +736,9 @@ int aac_hba_send(u8 command, struct fib *fibptr,
fib_callback callback,
wait = 1;
- if (command == HBA_IU_TYPE_SCSI_CMD_REQ) {
- struct aac_hba_cmd_req *hbacmd =
- (struct aac_hba_cmd_req *)fibptr->hw_fib_va;
+ hbacmd->iu_type = command;
- hbacmd->iu_type = command;
+ if (command == HBA_IU_TYPE_SCSI_CMD_REQ) {
/* bit1 of request_id must be 0 */
hbacmd->request_id =
cpu_to_le32u32)(fibptr - dev->fibs)) << 2)
+ 1);
diff --git a/drivers/scsi/vmw_pvscsi.c b/drivers/scsi/vmw_pvscsi.c
index c374e3b5c678..777e5f1e52d1 100644
--- a/drivers/scsi/vmw_pvscsi.c
+++ b/drivers/scsi/vmw_pvscsi.c
@@ -609,7 +609,7 @@ static void pvscsi_complete_request(struct
pvscsi_adapter *adapter,
break;
case BTSTAT_ABORTQUEUE:
- cmd->result = (DID_ABORT << 16);
+ cmd->result = (DID_BUS_BUSY << 16);
break;
case BTSTAT_SCSIPARITY:
Re: [PATCH 11/40] ipv6/flowlabel: simplify pid namespace lookup
On Sat, May 05, 2018 at 07:37:33AM -0500, Eric W. Biederman wrote:
> Christoph Hellwig writes:
>
> > The shole seq_file sequence already operates under a single RCU lock pair,
> > so move the pid namespace lookup into it, and stop grabbing a reference
> > and remove all kinds of boilerplate code.
>
> This is wrong.
>
> Move task_active_pid_ns(current) from open to seq_start actually means
> that the results if you pass this proc file between callers the results
> will change. So this breaks file descriptor passing.
>
> Open is a bad place to access current. In the middle of read/write is
> broken.
>
>
> In this particular instance looking up the pid namespace with
> task_active_pid_ns was a personal brain fart. What the code should be
> doing (with an appropriate helper) is:
>
> struct pid_namespace *pid_ns = inode->i_sb->s_fs_info;
>
> Because each mount of proc is bound to a pid namespace. Looking up the
> pid namespace from the super_block is a much better way to go.
What do you have in mind for the helper? For now I've thrown it in
opencoded into my working tree, but I'd be glad to add a helper.
struct pid_namespace *proc_pid_namespace(struct inode *inode)
{
// maybe warn on for s_magic not on procfs??
return inode->i_sb->s_fs_info;
}
?
Re: [PATCH 04/40] proc: introduce proc_create_seq{,_data}
On Mon, Apr 30, 2018 at 02:19:25PM +0100, David Howells wrote:
> Christoph Hellwig wrote:
>
> > +
> > +struct proc_dir_entry *proc_create_seq_data(const char *name, umode_t mode,
> > + struct proc_dir_entry *parent, const struct seq_operations *ops,
> > + void *data)
> > +{
> > ...
> > +EXPORT_SYMBOL(proc_create_seq_data);
>
> Please add documentation comments to exported functions when you add them.
None of the base functions are document, and we really want people to not
use procfs for new code anyway. But if I get some consensus from the
maintainers and the list I can throw in another patch to document
all proc_create* variants.
Re: [PATCH 34/40] atm: simplify procfs code
On Sat, May 05, 2018 at 07:51:18AM -0500, Eric W. Biederman wrote: > Christoph Hellwig writes: > > > Use remove_proc_subtree to remove the whole subtree on cleanup, and > > unwind the registration loop into individual calls. Switch to use > > proc_create_seq where applicable. > > Can you please explain why you are removing the error handling when > you are unwinding the registration loop? Because there is no point in handling these errors. The code work perfectly fine without procfs, or without given proc files and the removal works just fine if they don't exist either. This is a very common patter in various parts of the kernel already. I'll document it better in the changelog.
Re: simplify procfs code for seq_file instances V2
On Sun, May 06, 2018 at 08:19:49PM +0300, Alexey Dobriyan wrote: > On Wed, Apr 25, 2018 at 05:47:47PM +0200, Christoph Hellwig wrote: > > Changes since V1: > > - open code proc_create_data to avoid setting not fully initialized > >entries live > > - use unsigned int for state_size > > Need this to maintain sizeof(struct proc_dir_entry): I'll fold your changes into the relevant patches. > Otherwise ACK fs/proc/ part. I'll take this as a formal ACK-ed by for all patches touching procfs. If I was wrong please scream.
Re: [PATCH 06/40] proc: introduce proc_create_single{,_data}
On Thu, Apr 26, 2018 at 11:45:50AM +1000, Finn Thain wrote:
> >
> > -/*
> > - * /proc/nubus stuff
> > - */
> > -
>
> I don't think that the introduction of proc_create_single{,_data} alters
> the value of that comment. That comment and similar comments in the same
> file do have a purpose, which is to keep separate the /proc/nubus
> implementation is kept separate from the /proc/bus/nubus/devices
> implementation and so on.
Added back.
Hello
Dear Friend, I am Mr. Joel TOURE,a banker by profession and I work with one of the largest bank here as the Group Managing Director/Chief Executive Officer. Meanwhile,I'm seriously in need of a partner then I decide to contact you believing that you can partner with me as I have opportunity to transfer overdue funds worth Four Million Five Hundred Thousand US dollars (US$4,500.000.00) out from my bank and i need a serious and Trustworthy partner who can provide a reliable bank account where the money can be transfer . So, if you interested and agree with me then contact me here ( elume...@aol.fr ) and i will give you full details. Mr. Joel TOURE +22675329619
