svn commit: r361272 - head/contrib/wpa/src/drivers
Author: cy
Date: Wed May 20 04:16:13 2020
New Revision: 361272
URL: https://svnweb.freebsd.org/changeset/base/361272
Log:
Silence the once per second CTRL-EVENT-SCAN-FAILED errors when the WiFi
radio is disabled through the communication device toggle key (also known
as the RF raidio kill button). Only the CTRL-EVENT-DISCONNECTED will be
issued.
Submitted by: avg
Reported by: avg
MFC after:1 week
Modified:
head/contrib/wpa/src/drivers/driver_bsd.c
Modified: head/contrib/wpa/src/drivers/driver_bsd.c
==
--- head/contrib/wpa/src/drivers/driver_bsd.c Wed May 20 02:33:41 2020
(r361271)
+++ head/contrib/wpa/src/drivers/driver_bsd.c Wed May 20 04:16:13 2020
(r361272)
@@ -1358,14 +1358,18 @@ wpa_driver_bsd_event_receive(int sock, void *ctx, void
drv = bsd_get_drvindex(global, ifm->ifm_index);
if (drv == NULL)
return;
- if ((ifm->ifm_flags & IFF_UP) == 0 &&
- (drv->flags & IFF_UP) != 0) {
+ if (((ifm->ifm_flags & IFF_UP) == 0 ||
+ (ifm->ifm_flags & IFF_RUNNING) == 0) &&
+ (drv->flags & IFF_UP) != 0 &&
+ (drv->flags & IFF_RUNNING) != 0) {
wpa_printf(MSG_DEBUG, "RTM_IFINFO: Interface '%s' DOWN",
drv->ifname);
wpa_supplicant_event(drv->ctx, EVENT_INTERFACE_DISABLED,
NULL);
} else if ((ifm->ifm_flags & IFF_UP) != 0 &&
- (drv->flags & IFF_UP) == 0) {
+ (ifm->ifm_flags & IFF_RUNNING) != 0 &&
+ ((drv->flags & IFF_UP) == 0 ||
+ (drv->flags & IFF_RUNNING) == 0)) {
wpa_printf(MSG_DEBUG, "RTM_IFINFO: Interface '%s' UP",
drv->ifname);
wpa_supplicant_event(drv->ctx, EVENT_INTERFACE_ENABLED,
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361271 - in head/sys/powerpc: aim include
Author: jhibbits
Date: Wed May 20 02:33:41 2020
New Revision: 361271
URL: https://svnweb.freebsd.org/changeset/base/361271
Log:
powerpc/radix mmu: No need for delayed TLB invalidation
x86 needs delayed TLB invalidation because invalidation requires an
expensive IPI. PowerPC has had a TLB invalidation instruction since the
POWER1 in 1990, so there's no need to delay anything.
Modified:
head/sys/powerpc/aim/mmu_radix.c
head/sys/powerpc/include/proc.h
Modified: head/sys/powerpc/aim/mmu_radix.c
==
--- head/sys/powerpc/aim/mmu_radix.cWed May 20 01:25:46 2020
(r361270)
+++ head/sys/powerpc/aim/mmu_radix.cWed May 20 02:33:41 2020
(r361271)
@@ -33,7 +33,6 @@ __FBSDID("$FreeBSD$");
#include
#include
#include
-#include
#include
#include
#include
@@ -406,7 +405,6 @@ static u_int64_tKPTphys;/* phys addr of kernel
level
static vm_offset_t qframe = 0;
static struct mtx qframe_mtx;
-static epoch_t pmap_epoch;
void mmu_radix_activate(mmu_t mmu, struct thread *);
void mmu_radix_advise(mmu_t mmu, pmap_t, vm_offset_t, vm_offset_t, int);
@@ -849,43 +847,6 @@ pa_cmp(const void *a, const void *b)
PG_M | PG_A | RPTE_EAA_MASK | PG_V)
-static void
-pmap_epoch_init(void *arg __unused)
-{
- pmap_epoch = epoch_alloc("pmap", EPOCH_PREEMPT | EPOCH_LOCKED);
-}
-SYSINIT(epoch, SI_SUB_EPOCH + 1, SI_ORDER_ANY, pmap_epoch_init, NULL);
-
-static bool
-pmap_not_in_di(void)
-{
-
- return (curthread->td_md.md_invl_gen.gen == 0);
-}
-
-#definePMAP_ASSERT_NOT_IN_DI() \
-KASSERT(pmap_not_in_di(), ("DI already started"))
-
-static void
-pmap_delayed_invl_started(epoch_tracker_t et)
-{
- epoch_enter_preempt(pmap_epoch, et);
- curthread->td_md.md_invl_gen.gen = 1;
-}
-
-static void
-pmap_delayed_invl_finished(epoch_tracker_t et)
-{
- curthread->td_md.md_invl_gen.gen = 0;
- epoch_exit_preempt(pmap_epoch, et);
-}
-
-static void
-pmap_delayed_invl_wait(vm_page_t m __unused)
-{
- epoch_wait_preempt(pmap_epoch);
-}
-
static __inline void
pmap_resident_count_inc(pmap_t pmap, int count)
{
@@ -1370,8 +1331,7 @@ out:
}
static void
-reclaim_pv_chunk_leave_pmap(pmap_t pmap, pmap_t locked_pmap, bool start_di,
- epoch_tracker_t et)
+reclaim_pv_chunk_leave_pmap(pmap_t pmap, pmap_t locked_pmap)
{
if (pmap == NULL)
@@ -1379,8 +1339,6 @@ reclaim_pv_chunk_leave_pmap(pmap_t pmap, pmap_t locked
pmap_invalidate_all(pmap);
if (pmap != locked_pmap)
PMAP_UNLOCK(pmap);
- if (start_di)
- pmap_delayed_invl_finished(et);
}
/*
@@ -1410,8 +1368,6 @@ reclaim_pv_chunk(pmap_t locked_pmap, struct rwlock **l
struct spglist free;
uint64_t inuse;
int bit, field, freed;
- bool start_di;
- struct epoch_tracker et;
PMAP_LOCK_ASSERT(locked_pmap, MA_OWNED);
KASSERT(lockp != NULL, ("reclaim_pv_chunk: lockp is NULL"));
@@ -1423,13 +1379,6 @@ reclaim_pv_chunk(pmap_t locked_pmap, struct rwlock **l
pc_marker = (struct pv_chunk *)_marker_b;
pc_marker_end = (struct pv_chunk *)_marker_end_b;
- /*
-* A delayed invalidation block should already be active if
-* pmap_advise() or pmap_remove() called this function by way
-* of pmap_demote_l3e_locked().
-*/
- start_di = pmap_not_in_di();
-
mtx_lock(_chunks_mutex);
active_reclaims++;
TAILQ_INSERT_HEAD(_chunks, pc_marker, pc_lru);
@@ -1454,21 +1403,16 @@ reclaim_pv_chunk(pmap_t locked_pmap, struct rwlock **l
* corresponding pmap is locked.
*/
if (pmap != next_pmap) {
- reclaim_pv_chunk_leave_pmap(pmap, locked_pmap,
- start_di, );
+ reclaim_pv_chunk_leave_pmap(pmap, locked_pmap);
pmap = next_pmap;
/* Avoid deadlock and lock recursion. */
if (pmap > locked_pmap) {
RELEASE_PV_LIST_LOCK(lockp);
PMAP_LOCK(pmap);
- if (start_di)
- pmap_delayed_invl_started();
mtx_lock(_chunks_mutex);
continue;
} else if (pmap != locked_pmap) {
if (PMAP_TRYLOCK(pmap)) {
- if (start_di)
- pmap_delayed_invl_started();
mtx_lock(_chunks_mutex);
continue;
} else {
@@ -1480,8 +1424,7 @@ reclaim_pv_chunk(pmap_t locked_pmap, struct rwlock **l
continue;
svn commit: r361268 - head/tests/sys/net
Author: trasz
Date: Tue May 19 21:55:29 2020
New Revision: 361268
URL: https://svnweb.freebsd.org/changeset/base/361268
Log:
Make sys.net.if_bridge_test depend on python; sys.net.if_bridge_test.span
requires it.
MFC after:2 weeks
Sponsored by: DARPA
Modified:
head/tests/sys/net/Makefile
Modified: head/tests/sys/net/Makefile
==
--- head/tests/sys/net/Makefile Tue May 19 21:48:45 2020(r361267)
+++ head/tests/sys/net/Makefile Tue May 19 21:55:29 2020(r361268)
@@ -7,6 +7,7 @@ BINDIR= ${TESTSDIR}
ATF_TESTS_C+= if_epair
ATF_TESTS_SH+= if_bridge_test
+TEST_METADATA.if_bridge_test+= required_programs="python"
ATF_TESTS_SH+= if_clone_test
ATF_TESTS_SH+= if_lagg_test
ATF_TESTS_SH+= if_tun_test
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361267 - in head/tests/sys: netinet netinet6
Author: trasz
Date: Tue May 19 21:48:45 2020
New Revision: 361267
URL: https://svnweb.freebsd.org/changeset/base/361267
Log:
Make sys.netinet.output and sys.netinet6.output6 tests depend on python;
they need to be able to run net_receiver.py.
MFC after:2 weeks
Sponsored by: DARPA
Modified:
head/tests/sys/netinet/Makefile
head/tests/sys/netinet6/Makefile
Modified: head/tests/sys/netinet/Makefile
==
--- head/tests/sys/netinet/Makefile Tue May 19 19:53:12 2020
(r361266)
+++ head/tests/sys/netinet/Makefile Tue May 19 21:48:45 2020
(r361267)
@@ -10,6 +10,7 @@ ATF_TESTS_C= ip_reass_test \
socket_afinet
ATF_TESTS_SH= carp fibs_test redirect divert forward output lpm
+TEST_METADATA.output+= required_programs="python"
PROGS= udp_dontroute tcp_user_cookie
Modified: head/tests/sys/netinet6/Makefile
==
--- head/tests/sys/netinet6/MakefileTue May 19 19:53:12 2020
(r361266)
+++ head/tests/sys/netinet6/MakefileTue May 19 21:48:45 2020
(r361267)
@@ -14,6 +14,7 @@ ATF_TESTS_SH= \
forward6 \
output6 \
lpm6
+TEST_METADATA.output6+=required_programs="python"
${PACKAGE}FILES+= exthdr.py
${PACKAGE}FILES+= mld.py
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r347418 - head/sys/net
On 19 May 2020, at 17:02, Andrew Gallatin wrote:
On 2020-05-19 04:21, Kristof Provost wrote:
The if_bnxt driver initialises |.isc_nrxd_max = {INT32_MAX,
INT32_MAX, INT32_MAX},|, so presumably that’s the cause.
I don’t know what a sane value would be though. I’ve defaulted to
4096 (because that’s what some other iflib users seems to do) for
now, and that seems to work. It doesn’t panic and I can get traffic
through it at least:
You seem to be setting the max, not the default, and 4K max
descriptors on a 100g device is going to basically cripple it.
Yeah, I just grabbed whatever number other iflib users used. My
immediate concern was to get it to stop panicking.
How about setting to the next power of 2 below max int so as to keep
with the authors intent?
Makes sense, yes.
If we don't already have a macro, something like (INT32_MAX >> 1) + 1
https://reviews.freebsd.org/D24922
Best regards,
Kristof
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361266 - head/sys/cddl/contrib/opensolaris/common/lz4
Author: tsoome Date: Tue May 19 19:53:12 2020 New Revision: 361266 URL: https://svnweb.freebsd.org/changeset/base/361266 Log: lz4 hash table does not start zeroed illumos issue: https://www.illumos.org/issues/12757 Submitted by: andyf Modified: head/sys/cddl/contrib/opensolaris/common/lz4/lz4.c Modified: head/sys/cddl/contrib/opensolaris/common/lz4/lz4.c == --- head/sys/cddl/contrib/opensolaris/common/lz4/lz4.c Tue May 19 18:41:46 2020(r361265) +++ head/sys/cddl/contrib/opensolaris/common/lz4/lz4.c Tue May 19 19:53:12 2020(r361266) @@ -850,7 +850,7 @@ real_LZ4_compress(const char *source, char *dest, int #if defined(_KERNEL) || defined(_FAKE_KERNEL) void *ctx = kmem_cache_alloc(lz4_ctx_cache, KM_NOSLEEP); #else - void *ctx = malloc(sizeof(struct refTables)); + void *ctx = calloc(1, sizeof(struct refTables)); #endif int result; ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361265 - head
Author: freqlabs Date: Tue May 19 18:41:46 2020 New Revision: 361265 URL: https://svnweb.freebsd.org/changeset/base/361265 Log: Mention new jail(8) command hooks in RELNOTES Reported by: 0mp Approved by: mmacy (mentor) Modified: head/RELNOTES Modified: head/RELNOTES == --- head/RELNOTES Tue May 19 18:38:46 2020(r361264) +++ head/RELNOTES Tue May 19 18:41:46 2020(r361265) @@ -13,6 +13,12 @@ Changes to this file should not be MFCed. r361238: ZFS will now reject read(2) of a dirfd with EISDIR. +r361066: + Add exec.prepare and exec.release hooks for jail(8) and jail.conf(5). + exec.prepare runs before mounts, so can be used to populate new jails. + exec.release runs after unmounts, so can be used to remove ephemeral + jails. + r360920,r360923,r360924,r360927,r360928,r360931,r360933,r360936: Remove support for ARC4, Blowfish, Cast, DES, Triple DES, MD5, MD5-KPDK, MD5-HMAC, SHA1-KPDK, and Skipjack algorithms from ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361264 - head
Author: freqlabs Date: Tue May 19 18:38:46 2020 New Revision: 361264 URL: https://svnweb.freebsd.org/changeset/base/361264 Log: Mention new dd flags in RELNOTES Approved by: mmacy (mentor) Sponsored by: iXsystems, Inc. Modified: head/RELNOTES Modified: head/RELNOTES == --- head/RELNOTES Tue May 19 18:35:08 2020(r361263) +++ head/RELNOTES Tue May 19 18:38:46 2020(r361264) @@ -87,6 +87,10 @@ r352304: ntpd is no longer by default locked in memory. rlimit memlock 32 or rlimit memlock 0 can be used to restore this behaviour. +r351770,r352920,r352922,r352923: + dd(1) now supports conv=fsync, conv=fdatasync, oflag=fsync, oflag=sync, + and iflag=fullblock flags, compatible with illumos and GNU. + r351522: Add kernel-side support for in-kernel Transport Layer Security (KTLS). KTLS permits using sendfile(2) over sockets using ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361263 - in head/sys: netinet sys
Author: markj
Date: Tue May 19 18:35:08 2020
New Revision: 361263
URL: https://svnweb.freebsd.org/changeset/base/361263
Log:
Define a module version for accept filter modules.
Otherwise accept filters compiled into the kernel do not preempt
preloaded accept filter modules. Then, the preloaded file registers its
accept filter module before the kernel, and the kernel's attempt fails
since duplicate accept filter list entries are not permitted. This
causes the preloaded file's module to be released, since
module_register_init() does a lookup by name, so the preloaded file is
unloaded, and the accept filter's callback points to random memory since
preload_delete_name() unmaps the file on x86 as of r336505.
Add a new ACCEPT_FILTER_DEFINE macro which wraps the accept filter and
module definitions, and ensures that a module version is defined.
PR: 245870
Reported by: Thomas von Dein
MFC after:2 weeks
Sponsored by: The FreeBSD Foundation
Modified:
head/sys/netinet/accf_data.c
head/sys/netinet/accf_dns.c
head/sys/netinet/accf_http.c
head/sys/sys/socketvar.h
Modified: head/sys/netinet/accf_data.c
==
--- head/sys/netinet/accf_data.cTue May 19 18:34:50 2020
(r361262)
+++ head/sys/netinet/accf_data.cTue May 19 18:35:08 2020
(r361263)
@@ -42,20 +42,7 @@ __FBSDID("$FreeBSD$");
static int sohasdata(struct socket *so, void *arg, int waitflag);
-static struct accept_filter accf_data_filter = {
- "dataready",
- sohasdata,
- NULL,
- NULL
-};
-
-static moduledata_t accf_data_mod = {
- "accf_data",
- accept_filt_generic_mod_event,
- _data_filter
-};
-
-DECLARE_MODULE(accf_data, accf_data_mod, SI_SUB_DRIVERS, SI_ORDER_MIDDLE);
+ACCEPT_FILTER_DEFINE(accf_data, "dataready", sohasdata, NULL, NULL, 1);
static int
sohasdata(struct socket *so, void *arg, int waitflag)
Modified: head/sys/netinet/accf_dns.c
==
--- head/sys/netinet/accf_dns.c Tue May 19 18:34:50 2020(r361262)
+++ head/sys/netinet/accf_dns.c Tue May 19 18:35:08 2020(r361263)
@@ -41,6 +41,8 @@
/* check for full DNS request */
static int sohasdns(struct socket *so, void *arg, int waitflag);
+ACCEPT_FILTER_DEFINE(accf_dns, "dnsready", sohasdns, NULL, NULL, 1);
+
struct packet {
struct mbuf *m; /* Current mbuf. */
struct mbuf *n; /* nextpkt mbuf. */
@@ -55,21 +57,6 @@ struct packet {
/* check we can skip over various parts of DNS request */
static int skippacket(struct sockbuf *sb);
-
-static struct accept_filter accf_dns_filter = {
- "dnsready",
- sohasdns,
- NULL,
- NULL
-};
-
-static moduledata_t accf_dns_mod = {
- "accf_dns",
- accept_filt_generic_mod_event,
- _dns_filter
-};
-
-DECLARE_MODULE(accf_dns, accf_dns_mod, SI_SUB_DRIVERS, SI_ORDER_MIDDLE);
static int
sohasdns(struct socket *so, void *arg, int waitflag)
Modified: head/sys/netinet/accf_http.c
==
--- head/sys/netinet/accf_http.cTue May 19 18:34:50 2020
(r361262)
+++ head/sys/netinet/accf_http.cTue May 19 18:35:08 2020
(r361263)
@@ -54,29 +54,16 @@ static int mbufstrncmp(struct mbuf *m, struct mbuf *np
/* socketbuffer is full */
static int sbfull(struct sockbuf *sb);
-static struct accept_filter accf_http_filter = {
- "httpready",
- sohashttpget,
- NULL,
- NULL
-};
+ACCEPT_FILTER_DEFINE(accf_http, "httpready", sohashttpget, NULL, NULL, 1);
-static moduledata_t accf_http_mod = {
- "accf_http",
- accept_filt_generic_mod_event,
- _http_filter
-};
-
-DECLARE_MODULE(accf_http, accf_http_mod, SI_SUB_DRIVERS, SI_ORDER_MIDDLE);
-
static int parse_http_version = 1;
static SYSCTL_NODE(_net_inet_accf, OID_AUTO, http,
CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
"HTTP accept filter");
SYSCTL_INT(_net_inet_accf_http, OID_AUTO, parsehttpversion, CTLFLAG_RW,
-_http_version, 1,
-"Parse http version so that non 1.x requests work");
+_http_version, 1,
+"Parse http version so that non 1.x requests work");
#ifdef ACCF_HTTP_DEBUG
#define DPRINT(fmt, args...) \
Modified: head/sys/sys/socketvar.h
==
--- head/sys/sys/socketvar.hTue May 19 18:34:50 2020(r361262)
+++ head/sys/sys/socketvar.hTue May 19 18:35:08 2020(r361263)
@@ -335,6 +335,22 @@ struct accept_filter {
SLIST_ENTRY(accept_filter) accf_next;
};
+#defineACCEPT_FILTER_DEFINE(modname, filtname, cb, create, destroy,
ver) \
+ static struct accept_filter modname##_filter = {\
+ .accf_name = filtname,
svn commit: r361262 - head/sys/kern
Author: markj
Date: Tue May 19 18:34:50 2020
New Revision: 361262
URL: https://svnweb.freebsd.org/changeset/base/361262
Log:
Use the symbolic name for "modmetadata_set".
MFC after:1 week
Sponsored by: The FreeBSD Foundation
Modified:
head/sys/kern/kern_linker.c
Modified: head/sys/kern/kern_linker.c
==
--- head/sys/kern/kern_linker.c Tue May 19 16:28:20 2020(r361261)
+++ head/sys/kern/kern_linker.c Tue May 19 18:34:50 2020(r361262)
@@ -369,8 +369,7 @@ linker_file_register_modules(linker_file_t lf)
sx_assert(_sx, SA_XLOCKED);
- if (linker_file_lookup_set(lf, "modmetadata_set", ,
- , NULL) != 0) {
+ if (linker_file_lookup_set(lf, MDT_SETNAME, , , NULL) != 0) {
/*
* This fallback should be unnecessary, but if we get booted
* from boot2 instead of loader and we are missing our
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r361238 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
On 5/19/20 8:34 AM, Rodney W. Grimes wrote: > The EISDIR behavior is what your changing, independent of file system(s) > you have done so far. The fact the behavior is now different between > UFS and ZFS is sic, IMHO. Oh please. NFS already fails with ESIDIR and has failed to read() directories for 30 years: https://svnweb.freebsd.org/csrg?view=revision=41905 We haven't had riots in the streets over that. If you actually looked at what ZFS was returning for read(), it was useless. -- John Baldwin ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361261 - in head/sys/dev/cxgbe: . iw_cxgbe
Author: np
Date: Tue May 19 16:28:20 2020
New Revision: 361261
URL: https://svnweb.freebsd.org/changeset/base/361261
Log:
cxgbe/iw_cxgbe: Add an async callback to notify iw_cxgbe in case of a
fatal error.
Submitted by: Krishnamraju Eraparaju @ Chelsio
MFC after:2 weeks
Sponsored by: Chelsio Communications
Modified:
head/sys/dev/cxgbe/adapter.h
head/sys/dev/cxgbe/iw_cxgbe/cm.c
head/sys/dev/cxgbe/iw_cxgbe/device.c
head/sys/dev/cxgbe/offload.h
head/sys/dev/cxgbe/t4_main.c
Modified: head/sys/dev/cxgbe/adapter.h
==
--- head/sys/dev/cxgbe/adapter.hTue May 19 16:06:03 2020
(r361260)
+++ head/sys/dev/cxgbe/adapter.hTue May 19 16:28:20 2020
(r361261)
@@ -830,6 +830,7 @@ struct adapter {
int sc_do_rxcopy;
struct taskqueue *tq[MAX_NCHAN];/* General purpose taskqueues */
+ struct task async_event_task;
struct port_info *port[MAX_NPORTS];
uint8_t chan_map[MAX_NCHAN];/* channel -> port */
Modified: head/sys/dev/cxgbe/iw_cxgbe/cm.c
==
--- head/sys/dev/cxgbe/iw_cxgbe/cm.cTue May 19 16:06:03 2020
(r361260)
+++ head/sys/dev/cxgbe/iw_cxgbe/cm.cTue May 19 16:28:20 2020
(r361261)
@@ -1085,7 +1085,7 @@ c4iw_so_upcall(struct socket *so, void *arg, int waitf
* Wake up any threads waiting in rdma_init()/rdma_fini(),
* with locks held.
*/
- if (so->so_error)
+ if (so->so_error || (ep->com.dev->rdev.flags & T4_FATAL_ERROR))
c4iw_wake_up(>com.wr_wait, -ECONNRESET);
add_ep_to_req_list(ep, C4IW_EVENT_SOCKET);
@@ -2700,6 +2700,11 @@ c4iw_create_listen(struct iw_cm_id *cm_id, int backlog
CTR3(KTR_IW_CXGBE, "%s: cm_id %p, backlog %s", __func__, cm_id,
backlog);
+ if (c4iw_fatal_error(>rdev)) {
+ CTR2(KTR_IW_CXGBE, "%s: cm_id %p, fatal error", __func__,
+ cm_id);
+ return -EIO;
+ }
lep = alloc_ep(sizeof(*lep), GFP_KERNEL);
lep->com.cm_id = cm_id;
ref_cm_id(>com);
@@ -2800,7 +2805,6 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt,
{
int ret = 0;
int close = 0;
- int fatal = 0;
struct c4iw_rdev *rdev;
@@ -2809,12 +2813,14 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt,
rdev = >com.dev->rdev;
if (c4iw_fatal_error(rdev)) {
-
- CTR2(KTR_IW_CXGBE, "%s:ced1 %p", __func__, ep);
- fatal = 1;
+ CTR3(KTR_IW_CXGBE, "%s:ced1 fatal error %p %s", __func__, ep,
+ states[ep->com.state]);
+ if (ep->com.state != DEAD) {
+ send_abort(ep);
+ ep->com.state = DEAD;
+ }
close_complete_upcall(ep, -ECONNRESET);
- send_abort(ep);
- ep->com.state = DEAD;
+ return ECONNRESET;
}
CTR3(KTR_IW_CXGBE, "%s:ced2 %p %s", __func__, ep,
states[ep->com.state]);
@@ -2877,9 +2883,7 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt,
CTR2(KTR_IW_CXGBE, "%s:ced4 %p", __func__, ep);
set_bit(EP_DISC_ABORT, >com.history);
close_complete_upcall(ep, -ECONNRESET);
- ret = send_abort(ep);
- if (ret)
- fatal = 1;
+ send_abort(ep);
} else {
CTR2(KTR_IW_CXGBE, "%s:ced5 %p", __func__, ep);
@@ -2889,33 +2893,28 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt,
ep->com.state = MORIBUND;
CURVNET_SET(ep->com.so->so_vnet);
- sodisconnect(ep->com.so);
+ ret = sodisconnect(ep->com.so);
CURVNET_RESTORE();
- }
-
- }
-
- if (fatal) {
- set_bit(EP_DISC_FAIL, >com.history);
- if (!abrupt) {
- STOP_EP_TIMER(ep);
- close_complete_upcall(ep, -EIO);
- }
- if (ep->com.qp) {
- struct c4iw_qp_attributes attrs = {0};
-
- attrs.next_state = C4IW_QP_STATE_ERROR;
- ret = c4iw_modify_qp(ep->com.dev, ep->com.qp,
- C4IW_QP_ATTR_NEXT_STATE,
- , 1);
if (ret) {
- CTR2(KTR_IW_CXGBE, "%s:ced7 %p", __func__, ep);
- printf("%s - qp <- error failed!\n", __func__);
+ CTR2(KTR_IW_CXGBE, "%s:ced6 %p",
svn commit: r361260 - head/usr.sbin/ntp/ntpd
Author: delphij Date: Tue May 19 16:06:03 2020 New Revision: 361260 URL: https://svnweb.freebsd.org/changeset/base/361260 Log: Update leap-seconds to leap-seconds.3676924800. Obtained from:ftp://ftp.nist.gov/pub/time/leap-seconds.3676924800 MFC after:3 days Modified: head/usr.sbin/ntp/ntpd/leap-seconds Modified: head/usr.sbin/ntp/ntpd/leap-seconds == --- head/usr.sbin/ntp/ntpd/leap-seconds Tue May 19 16:04:27 2020 (r361259) +++ head/usr.sbin/ntp/ntpd/leap-seconds Tue May 19 16:06:03 2020 (r361260) @@ -204,10 +204,10 @@ # current -- the update time stamp, the data and the name of the file # will not change. # -# Updated through IERS Bulletin C58 -# File expires on: 28 June 2020 +# Updated through IERS Bulletin C59 +# File expires on: 28 December 2020 # -#@ 3802291200 +#@ 3818102400 # 2272060800 10 # 1 Jan 1972 2287785600 11 # 1 Jul 1972 @@ -252,4 +252,4 @@ # the hash line is also ignored in the # computation. # -#h f28827d2 f263b6c3 ec0f19eb a3e0dbf0 97f3fa30 +#h a1c168ae 27c79a7d 9dddcfc3 bcfe616b 2e2c44ea ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361259 - in head/sys/arm64: arm64 include
Author: andrew
Date: Tue May 19 16:04:27 2020
New Revision: 361259
URL: https://svnweb.freebsd.org/changeset/base/361259
Log:
Stop performing a full icache sync when the DIC and IDC flags are set
The DIC and IDC bits in the CTR_EL0 register signal to the kernel when it
can relax the instruction cache synchronisation operations. The IDC bit
means we can relax cleaning the data cache to the point of unification
while the DIC bit means we don't need to invalidate the instruction cache
for data coherence. In both cases an appropriate barrier is still needed.
For now only implement the case where both bits are set, as is the case
on the Neoverse-N1 as used in the Amazon AWS Graviton 2 CPU. Note that
this behaviour is a optional on the N1 so we may later need to implement
only one or the other bit being set.
There is a tunable to disable each flag on boot.
Testing on a 4 core Graviton 2 instance found a significant improvement
in sys and real time when running "make buildkernel -j4", with no
significant difference in user time.
Reviewed by: markj
Sponsored by: Innovate UK
Differential Revision:https://reviews.freebsd.org/D24853
Modified:
head/sys/arm64/arm64/cpufunc_asm.S
head/sys/arm64/arm64/identcpu.c
head/sys/arm64/include/cpufunc.h
Modified: head/sys/arm64/arm64/cpufunc_asm.S
==
--- head/sys/arm64/arm64/cpufunc_asm.S Tue May 19 15:27:20 2020
(r361258)
+++ head/sys/arm64/arm64/cpufunc_asm.S Tue May 19 16:04:27 2020
(r361259)
@@ -133,9 +133,20 @@ ENTRY(arm64_dcache_inv_range)
END(arm64_dcache_inv_range)
/*
- * void arm64_icache_sync_range(vm_offset_t, vm_size_t)
+ * void arm64_dic_idc_icache_sync_range(vm_offset_t, vm_size_t)
+ * When the CTR_EL0.IDC bit is set cleaning to PoU becomes a dsb.
+ * When the CTR_EL0.DIC bit is set icache invalidation becomes an isb.
*/
-ENTRY(arm64_icache_sync_range)
+ENTRY(arm64_dic_idc_icache_sync_range)
+ dsb ishst
+ isb
+ ret
+END(arm64_dic_idc_icache_sync_range)
+
+/*
+ * void arm64_aliasing_icache_sync_range(vm_offset_t, vm_size_t)
+ */
+ENTRY(arm64_aliasing_icache_sync_range)
/*
* XXX Temporary solution - I-cache flush should be range based for
* PIPT cache or IALLUIS for VIVT or VIPT caches
@@ -146,7 +157,7 @@ ENTRY(arm64_icache_sync_range)
dsb ish
isb
ret
-END(arm64_icache_sync_range)
+END(arm64_aliasing_icache_sync_range)
/*
* int arm64_icache_sync_range_checked(vm_offset_t, vm_size_t)
Modified: head/sys/arm64/arm64/identcpu.c
==
--- head/sys/arm64/arm64/identcpu.c Tue May 19 15:27:20 2020
(r361258)
+++ head/sys/arm64/arm64/identcpu.c Tue May 19 16:04:27 2020
(r361259)
@@ -56,6 +56,24 @@ char machine[] = "arm64";
extern int adaptive_machine_arch;
#endif
+static SYSCTL_NODE(_machdep, OID_AUTO, cache, CTLFLAG_RD | CTLFLAG_MPSAFE, 0,
+"Cache management tuning");
+
+static int allow_dic = 1;
+SYSCTL_INT(_machdep_cache, OID_AUTO, allow_dic, CTLFLAG_RDTUN, _dic, 0,
+"Allow optimizations based on the DIC cache bit");
+
+static int allow_idc = 1;
+SYSCTL_INT(_machdep_cache, OID_AUTO, allow_idc, CTLFLAG_RDTUN, _idc, 0,
+"Allow optimizations based on the IDC cache bit");
+
+/*
+ * The default implementation of I-cache sync assumes we have an
+ * aliasing cache until we know otherwise.
+ */
+void (*arm64_icache_sync_range)(vm_offset_t, vm_size_t) =
+_aliasing_icache_sync_range;
+
static int
sysctl_hw_machine(SYSCTL_HANDLER_ARGS)
{
@@ -977,6 +995,7 @@ identify_cpu_sysinit(void *dummy __unused)
{
int cpu;
u_long hwcap;
+ bool dic, idc;
/* Create a user visible cpu description with safe values */
memset(_cpu_desc, 0, sizeof(user_cpu_desc));
@@ -985,6 +1004,8 @@ identify_cpu_sysinit(void *dummy __unused)
ID_AA64PFR0_FP_NONE | ID_AA64PFR0_EL1_64 | ID_AA64PFR0_EL0_64;
user_cpu_desc.id_aa64dfr0 = ID_AA64DFR0_DebugVer_8;
+ dic = (allow_dic != 0);
+ idc = (allow_idc != 0);
CPU_FOREACH(cpu) {
print_cpu_features(cpu);
hwcap = parse_cpu_features_hwcap(cpu);
@@ -993,6 +1014,17 @@ identify_cpu_sysinit(void *dummy __unused)
else
elf_hwcap &= hwcap;
update_user_regs(cpu);
+
+ if (CTR_DIC_VAL(cpu_desc[cpu].ctr) == 0)
+ dic = false;
+ if (CTR_IDC_VAL(cpu_desc[cpu].ctr) == 0)
+ idc = false;
+ }
+
+ if (dic && idc) {
+ arm64_icache_sync_range = _dic_idc_icache_sync_range;
+ if (bootverbose)
+ printf("Enabling DIC & IDC ICache sync\n");
}
if ((elf_hwcap & HWCAP_ATOMICS) != 0) {
Modified:
Re: svn commit: r361238 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
On Tue, May 19, 2020 at 10:34 AM Rodney W. Grimes wrote: > > > On Tue, May 19, 2020 at 10:27 AM Rodney W. Grimes > > wrote: > > > > > > > On Tue, May 19, 2020 at 10:23 AM Rodney W. Grimes > > > > wrote: > > > > > > > > > > > Author: kevans > > > > > > Date: Tue May 19 02:41:05 2020 > > > > > > New Revision: 361238 > > > > > > URL: https://svnweb.freebsd.org/changeset/base/361238 > > > > > > > > > > > > Log: > > > > > > zfs: reject read(2) of a dirfd with EISDIR > > > > > > > > > > > > This is independent of the recently-discussed global change, > > > > > > which is still > > > > > > in review/discussion stage. > > > > > > > > > > > > This is effectively a measure for consistency in the ZFS world, > > > > > > where > > > > > > FreeBSD was the only platform (as far as I could find) that > > > > > > allowed this. > > > > > > What ZFS exposes is decidedly not useful for any real purposes, to > > > > > > paraphrase (hopefully faithfully) jhb's findings when exploring > > > > > > this: > > > > > > > > > > > > The size of a directory in ZFS is the number of directory entries > > > > > > within. > > > > > > When reading a directory, you would instead get the leading part > > > > > > of its raw > > > > > > contents; the amount you get being dictated by the "size," i.e. > > > > > > number of > > > > > > directory entries. There's decidedly (luckily) no stack > > > > > > disclosure happening > > > > > > here, though the behavior is bizarre and almost certainly a > > > > > > historical > > > > > > accident. > > > > > > > > > > > > This change has already been upstreamed to OpenZFS. > > > > > > > > > > Until the grep -d skip issue is addressed I object to this change as > > > > > it is going to cause people who do grep with wildcards to see lots > > > > > of errors that before where pretty much either silent (no match > > > > > occured) > > > > > or spit out a "binary file foo matches." > > > > > > > > > > > > > That seems preferable to grepping random bytes that don't particularly > > > > contain any strings? They'd never see "binary file foo matches" in > > > > this case. > > > > > > The difference is you rarely get a hit, and now your gauranteed to > > > get a hit on every single directory making grep * very noisy, where > > > it was often silent or nearly silent before. > > > > > > > As you noted in the review for the larger change, -d skip is a good > > option for the people that don't like this. It probably makes sense as > > a default, but then we'd be diverging from the other popular grep that > > defaults to -d read and spews out EISDIR more often than not. > > Yet another thing I hate about Linux, thank you for adding it to FreeBSD :-) > > > > > > > > > This isn't exactly divergent from the behavior they'd see with ZFS > > > > anywhere else. > > > > > > It is extremly divergent from 42 years of behavior. > > > > > > > I don't think ZFS has been implemented on FreeBSD for 42 years, and I > > don't find this grep argument compelling enough to restore peoples' > > ability to read the raw znode of a directory. > > The EISDIR behavior is what your changing, independent of file system(s) > you have done so far. The fact the behavior is now different between > UFS and ZFS is sic, IMHO. EISDIR in read(2) denotes that a filesystem does not support reading a directory, this isn't a new kind of error. In particular, ZFS traditionally does NOT support reading a directory like this. The behavior between UFS and ZFS should have always been different, this is correction of a historical *accident*. ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r361238 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
> On Tue, May 19, 2020 at 10:27 AM Rodney W. Grimes > wrote: > > > > > On Tue, May 19, 2020 at 10:23 AM Rodney W. Grimes > > > wrote: > > > > > > > > > Author: kevans > > > > > Date: Tue May 19 02:41:05 2020 > > > > > New Revision: 361238 > > > > > URL: https://svnweb.freebsd.org/changeset/base/361238 > > > > > > > > > > Log: > > > > > zfs: reject read(2) of a dirfd with EISDIR > > > > > > > > > > This is independent of the recently-discussed global change, which > > > > > is still > > > > > in review/discussion stage. > > > > > > > > > > This is effectively a measure for consistency in the ZFS world, > > > > > where > > > > > FreeBSD was the only platform (as far as I could find) that allowed > > > > > this. > > > > > What ZFS exposes is decidedly not useful for any real purposes, to > > > > > paraphrase (hopefully faithfully) jhb's findings when exploring > > > > > this: > > > > > > > > > > The size of a directory in ZFS is the number of directory entries > > > > > within. > > > > > When reading a directory, you would instead get the leading part of > > > > > its raw > > > > > contents; the amount you get being dictated by the "size," i.e. > > > > > number of > > > > > directory entries. There's decidedly (luckily) no stack disclosure > > > > > happening > > > > > here, though the behavior is bizarre and almost certainly a > > > > > historical > > > > > accident. > > > > > > > > > > This change has already been upstreamed to OpenZFS. > > > > > > > > Until the grep -d skip issue is addressed I object to this change as > > > > it is going to cause people who do grep with wildcards to see lots > > > > of errors that before where pretty much either silent (no match occured) > > > > or spit out a "binary file foo matches." > > > > > > > > > > That seems preferable to grepping random bytes that don't particularly > > > contain any strings? They'd never see "binary file foo matches" in > > > this case. > > > > The difference is you rarely get a hit, and now your gauranteed to > > get a hit on every single directory making grep * very noisy, where > > it was often silent or nearly silent before. > > > > As you noted in the review for the larger change, -d skip is a good > option for the people that don't like this. It probably makes sense as > a default, but then we'd be diverging from the other popular grep that > defaults to -d read and spews out EISDIR more often than not. Yet another thing I hate about Linux, thank you for adding it to FreeBSD :-) > > > > > > This isn't exactly divergent from the behavior they'd see with ZFS > > > anywhere else. > > > > It is extremly divergent from 42 years of behavior. > > > > I don't think ZFS has been implemented on FreeBSD for 42 years, and I > don't find this grep argument compelling enough to restore peoples' > ability to read the raw znode of a directory. The EISDIR behavior is what your changing, independent of file system(s) you have done so far. The fact the behavior is now different between UFS and ZFS is sic, IMHO. -- Rod Grimes rgri...@freebsd.org ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r361238 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
On Tue, May 19, 2020 at 10:27 AM Rodney W. Grimes wrote: > > > On Tue, May 19, 2020 at 10:23 AM Rodney W. Grimes > > wrote: > > > > > > > Author: kevans > > > > Date: Tue May 19 02:41:05 2020 > > > > New Revision: 361238 > > > > URL: https://svnweb.freebsd.org/changeset/base/361238 > > > > > > > > Log: > > > > zfs: reject read(2) of a dirfd with EISDIR > > > > > > > > This is independent of the recently-discussed global change, which is > > > > still > > > > in review/discussion stage. > > > > > > > > This is effectively a measure for consistency in the ZFS world, where > > > > FreeBSD was the only platform (as far as I could find) that allowed > > > > this. > > > > What ZFS exposes is decidedly not useful for any real purposes, to > > > > paraphrase (hopefully faithfully) jhb's findings when exploring this: > > > > > > > > The size of a directory in ZFS is the number of directory entries > > > > within. > > > > When reading a directory, you would instead get the leading part of > > > > its raw > > > > contents; the amount you get being dictated by the "size," i.e. > > > > number of > > > > directory entries. There's decidedly (luckily) no stack disclosure > > > > happening > > > > here, though the behavior is bizarre and almost certainly a historical > > > > accident. > > > > > > > > This change has already been upstreamed to OpenZFS. > > > > > > Until the grep -d skip issue is addressed I object to this change as > > > it is going to cause people who do grep with wildcards to see lots > > > of errors that before where pretty much either silent (no match occured) > > > or spit out a "binary file foo matches." > > > > > > > That seems preferable to grepping random bytes that don't particularly > > contain any strings? They'd never see "binary file foo matches" in > > this case. > > The difference is you rarely get a hit, and now your gauranteed to > get a hit on every single directory making grep * very noisy, where > it was often silent or nearly silent before. > As you noted in the review for the larger change, -d skip is a good option for the people that don't like this. It probably makes sense as a default, but then we'd be diverging from the other popular grep that defaults to -d read and spews out EISDIR more often than not. > > > > This isn't exactly divergent from the behavior they'd see with ZFS > > anywhere else. > > It is extremly divergent from 42 years of behavior. > I don't think ZFS has been implemented on FreeBSD for 42 years, and I don't find this grep argument compelling enough to restore peoples' ability to read the raw znode of a directory. ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r361238 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
> > On Tue, May 19, 2020 at 10:23 AM Rodney W. Grimes > > wrote: > > > > > > > Author: kevans > > > > Date: Tue May 19 02:41:05 2020 > > > > New Revision: 361238 > > > > URL: https://svnweb.freebsd.org/changeset/base/361238 > > > > > > > > Log: > > > > zfs: reject read(2) of a dirfd with EISDIR > > > > > > > > This is independent of the recently-discussed global change, which is > > > > still > > > > in review/discussion stage. > > > > > > > > This is effectively a measure for consistency in the ZFS world, where > > > > FreeBSD was the only platform (as far as I could find) that allowed > > > > this. > > > > What ZFS exposes is decidedly not useful for any real purposes, to > > > > paraphrase (hopefully faithfully) jhb's findings when exploring this: > > > > > > > > The size of a directory in ZFS is the number of directory entries > > > > within. > > > > When reading a directory, you would instead get the leading part of > > > > its raw > > > > contents; the amount you get being dictated by the "size," i.e. > > > > number of > > > > directory entries. There's decidedly (luckily) no stack disclosure > > > > happening > > > > here, though the behavior is bizarre and almost certainly a historical > > > > accident. > > > > > > > > This change has already been upstreamed to OpenZFS. > > > > > > Until the grep -d skip issue is addressed I object to this change as > > > it is going to cause people who do grep with wildcards to see lots > > > of errors that before where pretty much either silent (no match occured) > > > or spit out a "binary file foo matches." > > > > > > > That seems preferable to grepping random bytes that don't particularly > > contain any strings? They'd never see "binary file foo matches" in > > this case. > > The difference is you rarely get a hit, and now your gauranteed to > get a hit on every single directory making grep * very noisy, where > it was often silent or nearly silent before. Please, go try this and see if you can see why I am asserting what I am: (on one of your patched systems) cd /etc grep foo * grep -d skip foo * #This makes it closer to old behavior, not exact as binary matches shall be skipped. The first command isg going to return an error for every directory, the second command, closer to historical behavior, is going to be nearly silent. You could run the second command on a pre patched system, ufs or zfs wont matter much. > > This isn't exactly divergent from the behavior they'd see with ZFS > > anywhere else. > > It is extremly divergent from 42 years of behavior. > > > Thanks, > > > > Kyle Evans > > > > -- > Rod Grimes rgri...@freebsd.org > -- Rod Grimes rgri...@freebsd.org ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r361238 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
> On Tue, May 19, 2020 at 10:23 AM Rodney W. Grimes > wrote: > > > > > Author: kevans > > > Date: Tue May 19 02:41:05 2020 > > > New Revision: 361238 > > > URL: https://svnweb.freebsd.org/changeset/base/361238 > > > > > > Log: > > > zfs: reject read(2) of a dirfd with EISDIR > > > > > > This is independent of the recently-discussed global change, which is > > > still > > > in review/discussion stage. > > > > > > This is effectively a measure for consistency in the ZFS world, where > > > FreeBSD was the only platform (as far as I could find) that allowed > > > this. > > > What ZFS exposes is decidedly not useful for any real purposes, to > > > paraphrase (hopefully faithfully) jhb's findings when exploring this: > > > > > > The size of a directory in ZFS is the number of directory entries > > > within. > > > When reading a directory, you would instead get the leading part of its > > > raw > > > contents; the amount you get being dictated by the "size," i.e. number > > > of > > > directory entries. There's decidedly (luckily) no stack disclosure > > > happening > > > here, though the behavior is bizarre and almost certainly a historical > > > accident. > > > > > > This change has already been upstreamed to OpenZFS. > > > > Until the grep -d skip issue is addressed I object to this change as > > it is going to cause people who do grep with wildcards to see lots > > of errors that before where pretty much either silent (no match occured) > > or spit out a "binary file foo matches." > > > > That seems preferable to grepping random bytes that don't particularly > contain any strings? They'd never see "binary file foo matches" in > this case. The difference is you rarely get a hit, and now your gauranteed to get a hit on every single directory making grep * very noisy, where it was often silent or nearly silent before. > > This isn't exactly divergent from the behavior they'd see with ZFS > anywhere else. It is extremly divergent from 42 years of behavior. > Thanks, > > Kyle Evans > -- Rod Grimes rgri...@freebsd.org ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361258 - head/sys/arm64/arm64
Author: andrew
Date: Tue May 19 15:27:20 2020
New Revision: 361258
URL: https://svnweb.freebsd.org/changeset/base/361258
Log:
Create MSI/MSI-X isrcs as needed in the GICv3 ITS driver
Previously we would create an isrc for each MSI/MSI-X interrupt. This
causes issues for other interrupt sources in the system, e.g. a GPIO
driver, as they may be unable to allocate interrupts. This works around
this by allocating the isrc only when needed.
Reported by: alisa...@amazon.com
Reviewed by: mmel
Sponsored by: Innovaate UK
Differential Revision:https://reviews.freebsd.org/D24876
Modified:
head/sys/arm64/arm64/gicv3_its.c
Modified: head/sys/arm64/arm64/gicv3_its.c
==
--- head/sys/arm64/arm64/gicv3_its.cTue May 19 15:19:39 2020
(r361257)
+++ head/sys/arm64/arm64/gicv3_its.cTue May 19 15:27:20 2020
(r361258)
@@ -228,6 +228,7 @@ struct gicv3_its_irqsrc {
u_int gi_id;
u_int gi_lpi;
struct its_dev *gi_its_dev;
+ TAILQ_ENTRY(gicv3_its_irqsrc) gi_link;
};
struct gicv3_its_softc {
@@ -254,12 +255,14 @@ struct gicv3_its_softc {
size_t sc_its_cmd_next_idx;
vmem_t *sc_irq_alloc;
- struct gicv3_its_irqsrc *sc_irqs;
+ struct gicv3_its_irqsrc **sc_irqs;
u_int sc_irq_base;
u_int sc_irq_length;
+ u_int sc_irq_count;
struct mtx sc_its_dev_lock;
TAILQ_HEAD(its_dev_list, its_dev) sc_its_dev_list;
+ TAILQ_HEAD(free_irqs, gicv3_its_irqsrc) sc_free_irqs;
#defineITS_FLAGS_CMDQ_FLUSH0x0001
#defineITS_FLAGS_LPI_CONF_FLUSH0x0002
@@ -800,7 +803,6 @@ static int
gicv3_its_attach(device_t dev)
{
struct gicv3_its_softc *sc;
- const char *name;
uint32_t iidr;
int domain, err, i, rid;
@@ -875,6 +877,7 @@ gicv3_its_attach(device_t dev)
its_init_cpu(dev, sc);
TAILQ_INIT(>sc_its_dev_list);
+ TAILQ_INIT(>sc_free_irqs);
/*
* Create the vmem object to allocate INTRNG IRQs from. We try to
@@ -887,13 +890,6 @@ gicv3_its_attach(device_t dev)
sc->sc_irqs = malloc(sizeof(*sc->sc_irqs) * sc->sc_irq_length,
M_GICV3_ITS, M_WAITOK | M_ZERO);
- name = device_get_nameunit(dev);
- for (i = 0; i < sc->sc_irq_length; i++) {
- sc->sc_irqs[i].gi_id = -1;
- sc->sc_irqs[i].gi_lpi = i + sc->sc_irq_base - GIC_FIRST_LPI;
- err = intr_isrc_register(>sc_irqs[i].gi_isrc, dev, 0,
- "%s,%u", name, i);
- }
/* For GIC-500 install tracking sysctls. */
if ((iidr & (GITS_IIDR_PRODUCT_MASK | GITS_IIDR_IMPLEMENTOR_MASK)) ==
@@ -975,7 +971,7 @@ gicv3_its_intr(void *arg, uintptr_t irq)
struct trapframe *tf;
irq -= sc->sc_irq_base;
- girq = >sc_irqs[irq];
+ girq = sc->sc_irqs[irq];
if (girq == NULL)
panic("gicv3_its_intr: Invalid interrupt %ld",
irq + sc->sc_irq_base);
@@ -1201,6 +1197,53 @@ its_device_release(device_t dev, struct its_dev *its_d
free(its_dev, M_GICV3_ITS);
}
+static struct gicv3_its_irqsrc *
+gicv3_its_alloc_irqsrc(device_t dev, struct gicv3_its_softc *sc, u_int irq)
+{
+ struct gicv3_its_irqsrc *girq = NULL;
+
+ KASSERT(sc->sc_irqs[irq] == NULL,
+ ("%s: Interrupt %u already allocated", __func__, irq));
+ mtx_lock_spin(>sc_its_dev_lock);
+ if (!TAILQ_EMPTY(>sc_free_irqs)) {
+ girq = TAILQ_FIRST(>sc_free_irqs);
+ TAILQ_REMOVE(>sc_free_irqs, girq, gi_link);
+ }
+ mtx_unlock_spin(>sc_its_dev_lock);
+ if (girq == NULL) {
+ girq = malloc(sizeof(*girq), M_GICV3_ITS,
+ M_NOWAIT | M_ZERO);
+ if (girq == NULL)
+ return (NULL);
+ girq->gi_id = -1;
+ if (intr_isrc_register(>gi_isrc, dev, 0,
+ "%s,%u", device_get_nameunit(dev), irq) != 0) {
+ free(girq, M_GICV3_ITS);
+ return (NULL);
+ }
+ }
+ girq->gi_lpi = irq + sc->sc_irq_base - GIC_FIRST_LPI;
+ sc->sc_irqs[irq] = girq;
+
+ return (girq);
+}
+
+static void
+gicv3_its_release_irqsrc(struct gicv3_its_softc *sc,
+struct gicv3_its_irqsrc *girq)
+{
+ u_int irq;
+
+ mtx_assert(>sc_its_dev_lock, MA_OWNED);
+
+ irq = girq->gi_lpi + GIC_FIRST_LPI - sc->sc_irq_base;
+ sc->sc_irqs[irq] = NULL;
+
+ girq->gi_id = -1;
+ girq->gi_its_dev = NULL;
+ TAILQ_INSERT_TAIL(>sc_free_irqs, girq, gi_link);
+}
+
static int
gicv3_its_alloc_msi(device_t dev, device_t child, int count, int maxcount,
device_t *pic, struct intr_irqsrc **srcs)
@@ -1220,12 +1263,35 @@ gicv3_its_alloc_msi(device_t dev, device_t child, int
Re: svn commit: r361238 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
On Tue, May 19, 2020 at 10:23 AM Rodney W. Grimes wrote: > > > Author: kevans > > Date: Tue May 19 02:41:05 2020 > > New Revision: 361238 > > URL: https://svnweb.freebsd.org/changeset/base/361238 > > > > Log: > > zfs: reject read(2) of a dirfd with EISDIR > > > > This is independent of the recently-discussed global change, which is > > still > > in review/discussion stage. > > > > This is effectively a measure for consistency in the ZFS world, where > > FreeBSD was the only platform (as far as I could find) that allowed this. > > What ZFS exposes is decidedly not useful for any real purposes, to > > paraphrase (hopefully faithfully) jhb's findings when exploring this: > > > > The size of a directory in ZFS is the number of directory entries within. > > When reading a directory, you would instead get the leading part of its > > raw > > contents; the amount you get being dictated by the "size," i.e. number of > > directory entries. There's decidedly (luckily) no stack disclosure > > happening > > here, though the behavior is bizarre and almost certainly a historical > > accident. > > > > This change has already been upstreamed to OpenZFS. > > Until the grep -d skip issue is addressed I object to this change as > it is going to cause people who do grep with wildcards to see lots > of errors that before where pretty much either silent (no match occured) > or spit out a "binary file foo matches." > That seems preferable to grepping random bytes that don't particularly contain any strings? They'd never see "binary file foo matches" in this case. This isn't exactly divergent from the behavior they'd see with ZFS anywhere else. Thanks, Kyle Evans ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r361238 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
> Author: kevans
> Date: Tue May 19 02:41:05 2020
> New Revision: 361238
> URL: https://svnweb.freebsd.org/changeset/base/361238
>
> Log:
> zfs: reject read(2) of a dirfd with EISDIR
>
> This is independent of the recently-discussed global change, which is still
> in review/discussion stage.
>
> This is effectively a measure for consistency in the ZFS world, where
> FreeBSD was the only platform (as far as I could find) that allowed this.
> What ZFS exposes is decidedly not useful for any real purposes, to
> paraphrase (hopefully faithfully) jhb's findings when exploring this:
>
> The size of a directory in ZFS is the number of directory entries within.
> When reading a directory, you would instead get the leading part of its raw
> contents; the amount you get being dictated by the "size," i.e. number of
> directory entries. There's decidedly (luckily) no stack disclosure happening
> here, though the behavior is bizarre and almost certainly a historical
> accident.
>
> This change has already been upstreamed to OpenZFS.
Until the grep -d skip issue is addressed I object to this change as
it is going to cause people who do grep with wildcards to see lots
of errors that before where pretty much either silent (no match occured)
or spit out a "binary file foo matches."
>
> MFC after: 1 week
Please no.
> Modified:
> head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
>
> Modified: head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
> ==
> --- head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c Tue May
> 19 02:07:08 2020(r361237)
> +++ head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c Tue May
> 19 02:41:05 2020(r361238)
> @@ -646,6 +646,12 @@ zfs_read(vnode_t *vp, uio_t *uio, int ioflag, cred_t *
> ZFS_ENTER(zfsvfs);
> ZFS_VERIFY_ZP(zp);
>
> + /* We don't copy out anything useful for directories. */
> + if (vp->v_type == VDIR) {
> + ZFS_EXIT(zfsvfs);
> + return (SET_ERROR(EISDIR));
> + }
> +
> if (zp->z_pflags & ZFS_AV_QUARANTINED) {
> ZFS_EXIT(zfsvfs);
> return (SET_ERROR(EACCES));
>
--
Rod Grimes rgri...@freebsd.org
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361257 - head/usr.sbin/bsdinstall/scripts
Author: kevans Date: Tue May 19 15:19:39 2020 New Revision: 361257 URL: https://svnweb.freebsd.org/changeset/base/361257 Log: bsdinstall: do a `certctl rehash` upon installation of configuration If certctl is installed on the system we're configuring, do a certctl rehash. Note that certctl may not be present if the world we've installed was built either WITHOUT_OPENSSL or WITHOUT_CAROOT. In this scenario, we don't currently see if the host has a certctl as this may be an indication that the system *shouldn't* have certs installed into /etc/ssl. Reviewed by: allanjude, dteske MFC after:3 days Differential Revision:https://reviews.freebsd.org/D24640 Modified: head/usr.sbin/bsdinstall/scripts/config Modified: head/usr.sbin/bsdinstall/scripts/config == --- head/usr.sbin/bsdinstall/scripts/config Tue May 19 14:42:09 2020 (r361256) +++ head/usr.sbin/bsdinstall/scripts/config Tue May 19 15:19:39 2020 (r361257) @@ -55,6 +55,9 @@ cp $BSDINSTALL_TMPBOOT/* $BSDINSTALL_CHROOT/boot # Set up other things from installed config chroot $BSDINSTALL_CHROOT /usr/bin/newaliases > /dev/null 2>&1 +if [ -x $BSDINSTALL_CHROOT/usr/sbin/certctl ]; then + chroot $BSDINSTALL_CHROOT /usr/sbin/certctl rehash +fi exit 0 ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r347418 - head/sys/net
On 2020-05-19 04:21, Kristof Provost wrote:
The if_bnxt driver initialises |.isc_nrxd_max = {INT32_MAX, INT32_MAX,
INT32_MAX},|, so presumably that’s the cause.
I don’t know what a sane value would be though. I’ve defaulted to 4096
(because that’s what some other iflib users seems to do) for now, and
that seems to work. It doesn’t panic and I can get traffic through it at
least:
You seem to be setting the max, not the default, and 4K max descriptors
on a 100g device is going to basically cripple it.
How about setting to the next power of 2 below max int so as to keep
with the authors intent?
If we don't already have a macro, something like (INT32_MAX >> 1) + 1
Drew
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361254 - head/sys/netgraph/bluetooth/include
Author: takawata
Date: Tue May 19 13:58:52 2020
New Revision: 361254
URL: https://svnweb.freebsd.org/changeset/base/361254
Log:
Fix Typo in ng_hci_le_connection_complete_ep struct.
PR: 246538
Submitted by: Marc Veldman
Modified:
head/sys/netgraph/bluetooth/include/ng_hci.h
Modified: head/sys/netgraph/bluetooth/include/ng_hci.h
==
--- head/sys/netgraph/bluetooth/include/ng_hci.hTue May 19 13:54:33
2020(r361253)
+++ head/sys/netgraph/bluetooth/include/ng_hci.hTue May 19 13:58:52
2020(r361254)
@@ -1964,7 +1964,7 @@ typedef struct {
u_int16_t interval;
u_int8_tlatency;
u_int16_t supervision_timeout;
- u_int8_tmaster_clock_accracy;
+ u_int8_tmaster_clock_accuracy;
} __attribute__ ((packed)) ng_hci_le_connection_complete_ep;
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361253 - head
Author: markj Date: Tue May 19 13:54:33 2020 New Revision: 361253 URL: https://svnweb.freebsd.org/changeset/base/361253 Log: Belatedly add a RELNOTES entry for r347532. Modified: head/RELNOTES Modified: head/RELNOTES == --- head/RELNOTES Tue May 19 13:52:34 2020(r361252) +++ head/RELNOTES Tue May 19 13:54:33 2020(r361253) @@ -174,4 +174,10 @@ r347908-r347923: ed(4), ep(4), ex(4), fe(4), pcn(4), sf(4), sn(4), tl(4), tx(4), txp(4), vx(4), wb(4), xe(4). +r347532: + Wired page accounting has been split into kernel wirings and user + wirings (e.g., by mlock(2)). Kernel wirings no long count towards + the global limit, which is renamed to vm.max_user_wired. bhyve -S + allocates user-wired memory and is now subject to that limit. + $FreeBSD$ ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361251 - head/usr.sbin/bluetooth/hccontrol
Author: takawata
Date: Tue May 19 13:44:32 2020
New Revision: 361251
URL: https://svnweb.freebsd.org/changeset/base/361251
Log:
Add LE Whitelist commands to hccontrol
PR: 214555
Submitted by Marc Veldman
Modified:
head/usr.sbin/bluetooth/hccontrol/hccontrol.8
head/usr.sbin/bluetooth/hccontrol/le.c
Modified: head/usr.sbin/bluetooth/hccontrol/hccontrol.8
==
--- head/usr.sbin/bluetooth/hccontrol/hccontrol.8 Tue May 19 12:16:44
2020(r361250)
+++ head/usr.sbin/bluetooth/hccontrol/hccontrol.8 Tue May 19 13:44:32
2020(r361251)
@@ -157,6 +157,10 @@ are:
.It Cm LE_Read_Supported_States
.It Cm LE_Read_Buffer_Size
.It Cm LE Scan
+.It Cm LE_Read_White_List_Size
+.It Cm LE_Clear_White_List
+.It Cm LE_Add_Device_To_White_List
+.It Cm LE_Remove_Device_From_White_List
.El
.Pp
The currently supported node commands in
Modified: head/usr.sbin/bluetooth/hccontrol/le.c
==
--- head/usr.sbin/bluetooth/hccontrol/le.c Tue May 19 12:16:44 2020
(r361250)
+++ head/usr.sbin/bluetooth/hccontrol/le.c Tue May 19 13:44:32 2020
(r361251)
@@ -63,6 +63,10 @@ static int le_set_advertising_param(int s, int argc, c
static int le_read_advertising_channel_tx_power(int s, int argc, char *argv[]);
static int le_scan(int s, int argc, char *argv[]);
static void handle_le_event(ng_hci_event_pkt_t* e, bool verbose);
+static int le_read_white_list_size(int s, int argc, char *argv[]);
+static int le_clear_white_list(int s, int argc, char *argv[]);
+static int le_add_device_to_white_list(int s, int argc, char *argv[]);
+static int le_remove_device_from_white_list(int s, int argc, char *argv[]);
static int
le_set_scan_param(int s, int argc, char *argv[])
@@ -762,6 +766,173 @@ static void handle_le_event(ng_hci_event_pkt_t* e, boo
}
}
+static int
+le_read_white_list_size(int s, int argc, char *argv[])
+{
+ ng_hci_le_read_white_list_size_rp rp;
+ int n;
+
+ n = sizeof(rp);
+
+ if (hci_simple_request(s, NG_HCI_OPCODE(NG_HCI_OGF_LE,
+ NG_HCI_OCF_LE_READ_WHITE_LIST_SIZE),
+ (void *), ) == ERROR)
+ return (ERROR);
+
+ if (rp.status != 0x00) {
+ fprintf(stdout, "Status: %s [%#02x]\n",
+ hci_status2str(rp.status), rp.status);
+ return (FAILED);
+ }
+
+fprintf(stdout, "White list size: %d\n",
+ (uint8_t)rp.white_list_size);
+
+ return (OK);
+}
+
+static int
+le_clear_white_list(int s, int argc, char *argv[])
+{
+ ng_hci_le_clear_white_list_rp rp;
+ int n;
+
+ n = sizeof(rp);
+
+ if (hci_simple_request(s, NG_HCI_OPCODE(NG_HCI_OGF_LE,
+ NG_HCI_OCF_LE_CLEAR_WHITE_LIST),
+ (void *), ) == ERROR)
+ return (ERROR);
+
+ if (rp.status != 0x00) {
+ fprintf(stdout, "Status: %s [%#02x]\n",
+ hci_status2str(rp.status), rp.status);
+ return (FAILED);
+ }
+
+fprintf(stdout, "White list cleared\n");
+
+ return (OK);
+}
+
+static int
+le_add_device_to_white_list(int s, int argc, char *argv[])
+{
+ ng_hci_le_add_device_to_white_list_cp cp;
+ ng_hci_le_add_device_to_white_list_rp rp;
+ int n;
+ char ch;
+ optreset = 1;
+ optind = 0;
+ bool addr_set = false;
+
+ n = sizeof(rp);
+
+ cp.address_type = 0x00;
+
+ while ((ch = getopt(argc, argv , "t:a:")) != -1) {
+ switch(ch) {
+ case 't':
+ if (strcmp(optarg, "public") == 0)
+ cp.address_type = 0x00;
+ else if (strcmp(optarg, "random") == 0)
+ cp.address_type = 0x01;
+ else
+ return (USAGE);
+ break;
+ case 'a':
+ addr_set = true;
+ if (!bt_aton(optarg, )) {
+ struct hostent *he = NULL;
+
+ if ((he = bt_gethostbyname(optarg)) == NULL)
+ return (USAGE);
+
+ memcpy(, he->h_addr,
+ sizeof(cp.address));
+ }
+ break;
+ }
+ }
+
+ if (addr_set == false)
+ return (USAGE);
+
+ if (hci_request(s, NG_HCI_OPCODE(NG_HCI_OGF_LE,
+ NG_HCI_OCF_LE_ADD_DEVICE_TO_WHITE_LIST),
+ (void *), sizeof(cp), (void *), ) == ERROR)
+ return (ERROR);
+
+ if (rp.status != 0x00) {
+ fprintf(stdout, "Status: %s [%#02x]\n",
+
svn commit: r361250 - head/sbin/swapon
Author: bcr (doc committer) Date: Tue May 19 12:16:44 2020 New Revision: 361250 URL: https://svnweb.freebsd.org/changeset/base/361250 Log: The -F flag of swapon(8) requires -a to work. Before this change, swapon(8) implied that -F works as a standalone option, which is not the case and would produce a usage message. This change extends the description of the -F option to mention that -a is required with it. PR: 238551 Submitted by: Christian Baltini MFC after:5 days Modified: head/sbin/swapon/swapon.8 Modified: head/sbin/swapon/swapon.8 == --- head/sbin/swapon/swapon.8 Tue May 19 11:05:27 2020(r361249) +++ head/sbin/swapon/swapon.8 Tue May 19 12:16:44 2020(r361250) @@ -28,7 +28,7 @@ .\" @(#)swapon.8 8.1 (Berkeley) 6/5/93 .\" $FreeBSD$ .\" -.Dd June 21, 2019 +.Dd May 19, 2020 .Dt SWAPON 8 .Os .Sh NAME @@ -98,7 +98,7 @@ To delay .Nm swapon for a device until after .Nm savecore -has copied the crash dump to another location, use the +has copied the crash dump to another location, use the .Dq late option. .Pp @@ -143,7 +143,8 @@ are as follows: .It Fl F Ar fstab Specify the .Pa fstab -file to use. +file to use (in conjunction with +.Fl a Ns ). .El .Pp The ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361249 - head/lib/libc/stdlib
Author: bcr (doc committer) Date: Tue May 19 11:05:27 2020 New Revision: 361249 URL: https://svnweb.freebsd.org/changeset/base/361249 Log: Fix a typo: argments -> arguments PR: 243294 Submitted by: Igor Ostapenko MFC after:5 days Modified: head/lib/libc/stdlib/getopt_long.3 Modified: head/lib/libc/stdlib/getopt_long.3 == --- head/lib/libc/stdlib/getopt_long.3 Tue May 19 09:17:20 2020 (r361248) +++ head/lib/libc/stdlib/getopt_long.3 Tue May 19 11:05:27 2020 (r361249) @@ -244,7 +244,7 @@ matching the default behavior of .Xr getopt 3 . The default behavior without .Ql \&+ -is to permute non-option argments to the end of +is to permute non-option arguments to the end of .Fa argv . .Pp A leading ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361248 - head/usr.bin/sed
Author: bcr (doc committer) Date: Tue May 19 09:17:20 2020 New Revision: 361248 URL: https://svnweb.freebsd.org/changeset/base/361248 Log: Update SYNOPSIS section to be consistent regarding -u, -i, and -I. Apparently, when the -u, -i and -I options where added to sed(1), it was forgotten to add them to both lines in the SYNOPSIS section. They were only added to the second line, although they apply to both. With the updated SYNOPSIS, it is now allowed (and consistent) to run: sed -i BAK s/foo/bar/g some_file PR: 240556 Submitted by: Oliver Fromme MFC after:5 days Modified: head/usr.bin/sed/sed.1 Modified: head/usr.bin/sed/sed.1 == --- head/usr.bin/sed/sed.1 Tue May 19 09:04:35 2020(r361247) +++ head/usr.bin/sed/sed.1 Tue May 19 09:17:20 2020(r361248) @@ -31,7 +31,7 @@ .\"@(#)sed.1 8.2 (Berkeley) 12/30/93 .\" $FreeBSD$ .\" -.Dd March 27, 2017 +.Dd May 19, 2020 .Dt SED 1 .Os .Sh NAME @@ -41,9 +41,11 @@ .Nm .Op Fl Ealnru .Ar command +.Op Fl I Ar extension +.Op Fl i Ar extension .Op Ar .Nm -.Op Fl Ealnr +.Op Fl Ealnru .Op Fl e Ar command .Op Fl f Ar command_file .Op Fl I Ar extension @@ -197,7 +199,8 @@ option was specified); .It a context address that consists of a regular expression preceded and followed by a -delimiter. The closing delimiter can also optionally be followed by the +delimiter. +The closing delimiter can also optionally be followed by the .Dq I character, to indicate that the regular expression is to be matched in a case-insensitive way. ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361247 - head/sys/compat/linuxkpi/common/include/linux
Author: manu
Date: Tue May 19 09:04:35 2020
New Revision: 361247
URL: https://svnweb.freebsd.org/changeset/base/361247
Log:
linuxkpi: Add irq_work.h
Since handlers are call in a thread context we can simply use a workqueue
to emulate those functions.
The DRM code was patched to do that already, having it in linuxkpi allows us
to not patch the upstream code.
Sponsored-by: The FreeBSD Foundation
Reviewed by: hselasky
Differential Revision:https://reviews.freebsd.org/D24859
Added:
head/sys/compat/linuxkpi/common/include/linux/irq_work.h (contents, props
changed)
Added: head/sys/compat/linuxkpi/common/include/linux/irq_work.h
==
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ head/sys/compat/linuxkpi/common/include/linux/irq_work.hTue May 19
09:04:35 2020(r361247)
@@ -0,0 +1,52 @@
+/*-
+ * Copyright (c) 2020 The FreeBSD Foundation
+ *
+ * This software was developed by Emmanuel Vadot under sponsorship
+ * from the FreeBSD Foundation.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *notice, this list of conditions and the following disclaimer in the
+ *documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef __LINUX_IRQ_WORK_H__
+#define__LINUX_IRQ_WORK_H__
+
+#include
+
+struct irq_work {
+ struct work_struct work;
+};
+
+static inline void
+init_irq_work(struct irq_work *irqw, void (*func)(struct irq_work *))
+{
+ INIT_WORK(>work, (work_func_t)func);
+}
+
+static inline void
+irq_work_queue(struct irq_work *irqw)
+{
+ schedule_work(>work);
+}
+
+#endif /* __LINUX_IRQ_WORK_H__ */
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361246 - head/sys/compat/linuxkpi/common/include/linux
Author: manu
Date: Tue May 19 08:44:33 2020
New Revision: 361246
URL: https://svnweb.freebsd.org/changeset/base/361246
Log:
linuxkpi: add pci_dev_present
pci_dev_present shows if a set of pci ids are present in the system.
It just wraps pci_find_device.
Needed by DRMv5.2
Submitted by: Austing Shafer (asha...@badland.io)
Differential Revision:https://reviews.freebsd.org/D24796
Modified:
head/sys/compat/linuxkpi/common/include/linux/pci.h
Modified: head/sys/compat/linuxkpi/common/include/linux/pci.h
==
--- head/sys/compat/linuxkpi/common/include/linux/pci.h Tue May 19 08:43:17
2020(r361245)
+++ head/sys/compat/linuxkpi/common/include/linux/pci.h Tue May 19 08:44:33
2020(r361246)
@@ -1050,4 +1050,16 @@ extern int linux_pci_attach_device(device_t, struct pc
const struct pci_device_id *, struct pci_dev *);
extern int linux_pci_detach_device(struct pci_dev *);
+static inline int
+pci_dev_present(const struct pci_device_id *cur)
+{
+ while (cur != NULL && (cur->vendor || cur->device)) {
+ if (pci_find_device(cur->vendor, cur->device) != NULL) {
+ return (1);
+ }
+ cur++;
+ }
+ return (0);
+}
+
#endif /* _LINUX_PCI_H_ */
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361245 - head/sys/compat/linuxkpi/common/include/linux
Author: manu Date: Tue May 19 08:43:17 2020 New Revision: 361245 URL: https://svnweb.freebsd.org/changeset/base/361245 Log: linuxkpi: Add __init_waitqueue_head The only difference with init_waitqueue_head is that the name and the lock class key are provided but we don't use those so use init_waitqueue_head directly. Sponsored-by: The FreeBSD Foundation Differential Revision:https://reviews.freebsd.org/D24861 Modified: head/sys/compat/linuxkpi/common/include/linux/wait.h Modified: head/sys/compat/linuxkpi/common/include/linux/wait.h == --- head/sys/compat/linuxkpi/common/include/linux/wait.hTue May 19 07:47:59 2020(r361244) +++ head/sys/compat/linuxkpi/common/include/linux/wait.hTue May 19 08:43:17 2020(r361245) @@ -119,6 +119,8 @@ extern wait_queue_func_t default_wake_function; INIT_LIST_HEAD(&(wqh)->task_list); \ } while (0) +#define__init_waitqueue_head(wqh, name, lk) init_waitqueue_head(wqh) + void linux_init_wait_entry(wait_queue_t *, int); void linux_wake_up(wait_queue_head_t *, unsigned int, int, bool); ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r347418 - head/sys/net
On 10 May 2019, at 2:41, Eric Joyner wrote:
Author: erj
Date: Fri May 10 00:41:42 2019
New Revision: 347418
URL: https://svnweb.freebsd.org/changeset/base/347418
Log:
iflib: use default ntxd and nrxd when user value is not power of 2
From Jake:
A user may set a sysctl to override the default number of Tx or Rx
descriptors. However, certain calculations in the iflib core expect
the
number of descriptors to be a power of 2.
Update _iflib_assert to verify that all of the shared context
parameters
for the number of descriptors are powers of 2.
Modify iflib_reset_qvalues to check that the provided isc_nrxd value
is
a power of 2. If it's not, print a warning message and then use the
default value.
An alternative might be to try rounding the number down instead.
However, this creates problems in case the rounded down value is
below
the minimum value that the driver would support.
This commit appears to trigger a panic I see on a system with a Broadcom
BCM57416 (if_bnxt) nic.
It trips over the power of two assertion:
panic: Assertion powerof2(sctx->isc_nrxd_max[i]) failed at
/usr/src/sys/net/iflib.c:5320
Tracing pid 0 tid 10 td 0x81c8c640
kdb_enter() at kdb_enter+0x37/frame 0x825be990
vpanic() at vpanic+0x19e/frame 0x825be9e0
panic() at panic+0x43/frame 0x825bea40
iflib_register() at iflib_register+0x340/frame 0x825bea80
iflib_device_register() at iflib_device_register+0x9f/frame
0x825bee10
iflib_device_attach() at iflib_device_attach+0xb5/frame
0x825bee40
device_attach() at device_attach+0x3ca/frame 0x825bee80
device_probe_and_attach() at device_probe_and_attach+0x70/frame
0x825beeb0
bus_generic_attach() at bus_generic_attach+0x18/frame
0x825beed0
pci_attach() at pci_attach+0xe0/frame 0x825bef10
acpi_pci_attach() at acpi_pci_attach+0x19/frame 0x825bf150
device_attach() at device_attach+0x3ca/frame 0x825bf190
device_probe_and_attach() at device_probe_and_attach+0x70/frame
0x825bf1c0
bus_generic_attach() at bus_generic_attach+0x18/frame
0x825bf1e0
acpi_pcib_acpi_attach() at acpi_pcib_acpi_attach+0x431/frame
0x825bf250
device_attach() at device_attach+0x3ca/frame 0x825bf290
device_probe_and_attach() at device_probe_and_attach+0x70/frame
0x825bf2c0
bus_generic_attach() at bus_generic_attach+0x18/frame
0x825bf2e0
acpi_attach() at acpi_attach+0xbb7/frame 0x825bf370
device_attach() at device_attach+0x3ca/frame 0x825bf3b0
device_probe_and_attach() at device_probe_and_attach+0x70/frame
0x825bf3e0
bus_generic_attach() at bus_generic_attach+0x18/frame
0x825bf400
device_attach() at device_attach+0x3ca/frame 0x825bf440
device_probe_and_attach() at device_probe_and_attach+0x70/frame
0x825bf470
bus_generic_new_pass() at bus_generic_new_pass+0xed/frame
0x825bf4a0
bus_set_pass() at bus_set_pass+0x46/frame 0x825bf4d0
configure() at configure+0x9/frame 0x825bf4e0
mi_startup() at mi_startup+0xec/frame 0x825bf530
btext() at btext+0x2c
The if_bnxt driver initialises `.isc_nrxd_max = {INT32_MAX, INT32_MAX,
INT32_MAX},`, so presumably that’s the cause.
I don’t know what a sane value would be though. I’ve defaulted to
4096 (because that’s what some other iflib users seems to do) for now,
and that seems to work. It doesn’t panic and I can get traffic through
it at least:
diff --git a/sys/dev/bnxt/if_bnxt.c b/sys/dev/bnxt/if_bnxt.c
index 50827106024..3958d95cab9 100644
--- a/sys/dev/bnxt/if_bnxt.c
+++ b/sys/dev/bnxt/if_bnxt.c
@@ -316,11 +316,11 @@ static struct if_shared_ctx bnxt_sctx_init = {
.isc_nrxd_default = {PAGE_SIZE / sizeof(struct cmpl_base) * 8,
PAGE_SIZE / sizeof(struct rx_prod_pkt_bd),
PAGE_SIZE / sizeof(struct rx_prod_pkt_bd)},
- .isc_nrxd_max = {INT32_MAX, INT32_MAX, INT32_MAX},
+ .isc_nrxd_max = {4096, 4096, 4096},
.isc_ntxd_min = {16, 16, 16},
.isc_ntxd_default = {PAGE_SIZE / sizeof(struct cmpl_base) * 2,
PAGE_SIZE / sizeof(struct tx_bd_short)},
- .isc_ntxd_max = {INT32_MAX, INT32_MAX, INT32_MAX},
+ .isc_ntxd_max = {4096, 4096, 4096},
.isc_admin_intrcnt = 1,
.isc_vendor_info = bnxt_vendor_info_array,
Best regards,
Kristof
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r361243 - head/sys/netinet
Author: tuexen
Date: Tue May 19 07:23:35 2020
New Revision: 361243
URL: https://svnweb.freebsd.org/changeset/base/361243
Log:
Replace snprintf() by SCTP_SNPRINTF() and let SCTP_SNPRINTF() map
to snprintf() on FreeBSD. This allows to check for failures of snprintf()
on platforms other than FreeBSD kernel.
Modified:
head/sys/netinet/sctp_asconf.c
head/sys/netinet/sctp_indata.c
head/sys/netinet/sctp_input.c
head/sys/netinet/sctp_os_bsd.h
head/sys/netinet/sctp_output.c
head/sys/netinet/sctp_pcb.c
Modified: head/sys/netinet/sctp_asconf.c
==
--- head/sys/netinet/sctp_asconf.c Tue May 19 07:21:11 2020
(r361242)
+++ head/sys/netinet/sctp_asconf.c Tue May 19 07:23:35 2020
(r361243)
@@ -1706,8 +1706,7 @@ sctp_handle_asconf_ack(struct mbuf *m, int offset,
char msg[SCTP_DIAG_INFO_LEN];
SCTPDBG(SCTP_DEBUG_ASCONF1, "handle_asconf_ack: got unexpected
next serial number! Aborting asoc!\n");
- snprintf(msg, sizeof(msg), "Never sent serial number %8.8x",
- serial_num);
+ SCTP_SNPRINTF(msg, sizeof(msg), "Never sent serial number
%8.8x", serial_num);
op_err = sctp_generate_cause(SCTP_CAUSE_PROTOCOL_VIOLATION,
msg);
sctp_abort_an_association(stcb->sctp_ep, stcb, op_err,
SCTP_SO_NOT_LOCKED);
*abort_no_unlock = 1;
Modified: head/sys/netinet/sctp_indata.c
==
--- head/sys/netinet/sctp_indata.c Tue May 19 07:21:11 2020
(r361242)
+++ head/sys/netinet/sctp_indata.c Tue May 19 07:23:35 2020
(r361243)
@@ -434,7 +434,7 @@ sctp_abort_in_reasm(struct sctp_tcb *stcb,
struct mbuf *oper;
if (stcb->asoc.idata_supported) {
- snprintf(msg, sizeof(msg),
+ SCTP_SNPRINTF(msg, sizeof(msg),
"Reass %x,CF:%x,TSN=%8.8x,SID=%4.4x,FSN=%8.8x,MID:%8.8x",
opspot,
control->fsn_included,
@@ -442,7 +442,7 @@ sctp_abort_in_reasm(struct sctp_tcb *stcb,
chk->rec.data.sid,
chk->rec.data.fsn, chk->rec.data.mid);
} else {
- snprintf(msg, sizeof(msg),
+ SCTP_SNPRINTF(msg, sizeof(msg),
"Reass %x,CI:%x,TSN=%8.8x,SID=%4.4x,FSN=%4.4x,SSN:%4.4x",
opspot,
control->fsn_included,
@@ -533,11 +533,11 @@ sctp_queue_data_to_stream(struct sctp_tcb *stcb,
*/
TAILQ_INSERT_HEAD(>inqueue, control, next_instrm);
if (asoc->idata_supported) {
- snprintf(msg, sizeof(msg), "Delivered MID=%8.8x, got
TSN=%8.8x, SID=%4.4x, MID=%8.8x",
+ SCTP_SNPRINTF(msg, sizeof(msg), "Delivered MID=%8.8x,
got TSN=%8.8x, SID=%4.4x, MID=%8.8x",
strm->last_mid_delivered, control->sinfo_tsn,
control->sinfo_stream, control->mid);
} else {
- snprintf(msg, sizeof(msg), "Delivered SSN=%4.4x, got
TSN=%8.8x, SID=%4.4x, SSN=%4.4x",
+ SCTP_SNPRINTF(msg, sizeof(msg), "Delivered SSN=%4.4x,
got TSN=%8.8x, SID=%4.4x, SSN=%4.4x",
(uint16_t)strm->last_mid_delivered,
control->sinfo_tsn,
control->sinfo_stream,
@@ -648,9 +648,8 @@ sctp_queue_data_to_stream(struct sctp_tcb *stcb,
* to put it on the queue.
*/
if (sctp_place_control_in_stream(strm, asoc, control)) {
- snprintf(msg, sizeof(msg),
- "Queue to str MID: %u duplicate",
- control->mid);
+ SCTP_SNPRINTF(msg, sizeof(msg),
+ "Queue to str MID: %u duplicate", control->mid);
sctp_clean_up_control(stcb, control);
op_err =
sctp_generate_cause(SCTP_CAUSE_PROTOCOL_VIOLATION, msg);
stcb->sctp_ep->last_abort_code = SCTP_FROM_SCTP_INDATA
+ SCTP_LOC_3;
@@ -1881,8 +1880,7 @@ sctp_process_a_data_chunk(struct sctp_tcb *stcb, struc
* can *not* be fsn 0. XXX: This can happen in case of a
* wrap around. Ignore is for now.
*/
- snprintf(msg, sizeof(msg), "FSN zero for MID=%8.8x, but
flags=%2.2x",
- mid, chk_flags);
+ SCTP_SNPRINTF(msg, sizeof(msg), "FSN zero for MID=%8.8x, but
flags=%2.2x", mid, chk_flags);
goto err_out;
}
control = sctp_find_reasm_entry(>strmin[sid], mid, ordered,
asoc->idata_supported);
@@ -1893,7 +1891,7 @@ sctp_process_a_data_chunk(struct sctp_tcb *stcb, struc
if (control != NULL) {
svn commit: r361242 - head/sys/netinet
Author: tuexen
Date: Tue May 19 07:21:11 2020
New Revision: 361242
URL: https://svnweb.freebsd.org/changeset/base/361242
Log:
Revert r361209:
cem noted that on FreeBSD snprintf() can not fail and code should not
check for that.
A followup commit will replace the usage of snprintf() in the SCTP
sources with a variadic macro SCTP_SNPRINTF, which will simply map to
snprintf() on FreeBSD and do a checking similar to r361209 on
other platforms.
Modified:
head/sys/netinet/sctp_asconf.c
head/sys/netinet/sctp_indata.c
head/sys/netinet/sctp_input.c
head/sys/netinet/sctp_output.c
head/sys/netinet/sctp_pcb.c
Modified: head/sys/netinet/sctp_asconf.c
==
--- head/sys/netinet/sctp_asconf.c Tue May 19 07:16:09 2020
(r361241)
+++ head/sys/netinet/sctp_asconf.c Tue May 19 07:21:11 2020
(r361242)
@@ -1706,9 +1706,8 @@ sctp_handle_asconf_ack(struct mbuf *m, int offset,
char msg[SCTP_DIAG_INFO_LEN];
SCTPDBG(SCTP_DEBUG_ASCONF1, "handle_asconf_ack: got unexpected
next serial number! Aborting asoc!\n");
- if (snprintf(msg, sizeof(msg), "Never sent serial number
%8.8x", serial_num) < 0) {
- msg[0] = '\0';
- }
+ snprintf(msg, sizeof(msg), "Never sent serial number %8.8x",
+ serial_num);
op_err = sctp_generate_cause(SCTP_CAUSE_PROTOCOL_VIOLATION,
msg);
sctp_abort_an_association(stcb->sctp_ep, stcb, op_err,
SCTP_SO_NOT_LOCKED);
*abort_no_unlock = 1;
Modified: head/sys/netinet/sctp_indata.c
==
--- head/sys/netinet/sctp_indata.c Tue May 19 07:16:09 2020
(r361241)
+++ head/sys/netinet/sctp_indata.c Tue May 19 07:21:11 2020
(r361242)
@@ -434,26 +434,22 @@ sctp_abort_in_reasm(struct sctp_tcb *stcb,
struct mbuf *oper;
if (stcb->asoc.idata_supported) {
- if (snprintf(msg, sizeof(msg),
+ snprintf(msg, sizeof(msg),
"Reass %x,CF:%x,TSN=%8.8x,SID=%4.4x,FSN=%8.8x,MID:%8.8x",
opspot,
control->fsn_included,
chk->rec.data.tsn,
chk->rec.data.sid,
- chk->rec.data.fsn, chk->rec.data.mid) < 0) {
- msg[0] = '\0';
- }
+ chk->rec.data.fsn, chk->rec.data.mid);
} else {
- if (snprintf(msg, sizeof(msg),
+ snprintf(msg, sizeof(msg),
"Reass %x,CI:%x,TSN=%8.8x,SID=%4.4x,FSN=%4.4x,SSN:%4.4x",
opspot,
control->fsn_included,
chk->rec.data.tsn,
chk->rec.data.sid,
chk->rec.data.fsn,
- (uint16_t)chk->rec.data.mid) < 0) {
- msg[0] = '\0';
- }
+ (uint16_t)chk->rec.data.mid);
}
oper = sctp_generate_cause(SCTP_CAUSE_PROTOCOL_VIOLATION, msg);
sctp_m_freem(chk->data);
@@ -537,19 +533,15 @@ sctp_queue_data_to_stream(struct sctp_tcb *stcb,
*/
TAILQ_INSERT_HEAD(>inqueue, control, next_instrm);
if (asoc->idata_supported) {
- if (snprintf(msg, sizeof(msg), "Delivered MID=%8.8x,
got TSN=%8.8x, SID=%4.4x, MID=%8.8x",
+ snprintf(msg, sizeof(msg), "Delivered MID=%8.8x, got
TSN=%8.8x, SID=%4.4x, MID=%8.8x",
strm->last_mid_delivered, control->sinfo_tsn,
- control->sinfo_stream, control->mid) < 0) {
- msg[0] = '\0';
- }
+ control->sinfo_stream, control->mid);
} else {
- if (snprintf(msg, sizeof(msg), "Delivered SSN=%4.4x,
got TSN=%8.8x, SID=%4.4x, SSN=%4.4x",
+ snprintf(msg, sizeof(msg), "Delivered SSN=%4.4x, got
TSN=%8.8x, SID=%4.4x, SSN=%4.4x",
(uint16_t)strm->last_mid_delivered,
control->sinfo_tsn,
control->sinfo_stream,
- (uint16_t)control->mid) < 0) {
- msg[0] = '\0';
- }
+ (uint16_t)control->mid);
}
op_err = sctp_generate_cause(SCTP_CAUSE_PROTOCOL_VIOLATION,
msg);
stcb->sctp_ep->last_abort_code = SCTP_FROM_SCTP_INDATA +
SCTP_LOC_2;
@@ -656,10 +648,9 @@ sctp_queue_data_to_stream(struct sctp_tcb *stcb,
* to put it on the queue.
*/
if (sctp_place_control_in_stream(strm, asoc, control)) {
- if (snprintf(msg, sizeof(msg),
-
svn commit: r361241 - head/tests/sys/netinet
Author: lwhsu
Date: Tue May 19 07:16:09 2020
New Revision: 361241
URL: https://svnweb.freebsd.org/changeset/base/361241
Log:
Revert r361237, it's fixed by r361231
PR: 246560
Sponsored by: The FreeBSD Foundation
Modified:
head/tests/sys/netinet/so_reuseport_lb_test.c
Modified: head/tests/sys/netinet/so_reuseport_lb_test.c
==
--- head/tests/sys/netinet/so_reuseport_lb_test.c Tue May 19 03:24:45
2020(r361240)
+++ head/tests/sys/netinet/so_reuseport_lb_test.c Tue May 19 07:16:09
2020(r361241)
@@ -188,9 +188,6 @@ ATF_TC_BODY(basic_ipv6, tc)
int error, sds[16];
uint16_t port;
- if (atf_tc_get_config_var_as_bool_wd(tc, "ci", false))
- atf_tc_skip("https://bugs.freebsd.org/246560;);
-
sds[0] = lb_listen_socket(PF_INET6, SOCK_NONBLOCK);
memset(, 0, sizeof(addr));
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
