svn commit: r366746 - stable/12/sys/arm/allwinner
Author: avg Date: Fri Oct 16 06:42:25 2020 New Revision: 366746 URL: https://svnweb.freebsd.org/changeset/base/366746 Log: MFC r366140: aw_pwm: fix programming of the period The programmed value is biased by one: 0 means 1 cycle, 1 means 2 cycles, etc. Modified: stable/12/sys/arm/allwinner/aw_pwm.c Directory Properties: stable/12/ (props changed) Modified: stable/12/sys/arm/allwinner/aw_pwm.c == --- stable/12/sys/arm/allwinner/aw_pwm.cFri Oct 16 06:40:59 2020 (r366745) +++ stable/12/sys/arm/allwinner/aw_pwm.cFri Oct 16 06:42:25 2020 (r366746) @@ -314,7 +314,7 @@ aw_pwm_channel_config(device_t dev, u_int channel, u_i AW_PWM_WRITE(sc, AW_PWM_CTRL, reg); /* Write the total/active cycles */ - reg = ((clk_rate / period_freq) << AW_PWM_PERIOD_TOTAL_SHIFT) | + reg = ((clk_rate / period_freq - 1) << AW_PWM_PERIOD_TOTAL_SHIFT) | ((clk_rate / duty_freq) << AW_PWM_PERIOD_ACTIVE_SHIFT); AW_PWM_WRITE(sc, AW_PWM_PERIOD, reg); ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366745 - stable/12/sys/arm/allwinner
Author: avg
Date: Fri Oct 16 06:40:59 2020
New Revision: 366745
URL: https://svnweb.freebsd.org/changeset/base/366745
Log:
MFC r366138: aw_pwm: remove the busy bit check
The bit seems to always be set on my hardware, H3.
However, programming the hardware seems to work just fine.
Modified:
stable/12/sys/arm/allwinner/aw_pwm.c
Directory Properties:
stable/12/ (props changed)
Modified: stable/12/sys/arm/allwinner/aw_pwm.c
==
--- stable/12/sys/arm/allwinner/aw_pwm.cFri Oct 16 02:19:33 2020
(r366744)
+++ stable/12/sys/arm/allwinner/aw_pwm.cFri Oct 16 06:40:59 2020
(r366745)
@@ -300,10 +300,6 @@ aw_pwm_channel_config(device_t dev, u_int channel, u_i
}
reg = AW_PWM_READ(sc, AW_PWM_CTRL);
- if (reg & AW_PWM_CTRL_PERIOD_BUSY) {
- device_printf(sc->dev, "pwm busy\n");
- return (EBUSY);
- }
/* Write the prescalar */
reg &= ~AW_PWM_CTRL_PRESCALE_MASK;
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366744 - head/sys/kern
Author: mjg
Date: Fri Oct 16 02:19:33 2020
New Revision: 366744
URL: https://svnweb.freebsd.org/changeset/base/366744
Log:
cache: flip inverted condition in previous
It happened to not affect correctness in that the fallback code would
simply neglect to promote the entry.
Modified:
head/sys/kern/vfs_cache.c
Modified: head/sys/kern/vfs_cache.c
==
--- head/sys/kern/vfs_cache.c Fri Oct 16 00:56:13 2020(r366743)
+++ head/sys/kern/vfs_cache.c Fri Oct 16 02:19:33 2020(r366744)
@@ -1624,7 +1624,7 @@ negative_success:
vfs_smr_exit();
goto out_fallback;
}
- if (neg_hot) {
+ if (!neg_hot) {
vfs_smr_exit();
if (!cache_negative_promote_cond(dvp, cnp, ncp, hash))
goto out_fallback;
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366743 - head/sys/kern
Author: mjg
Date: Fri Oct 16 00:56:13 2020
New Revision: 366743
URL: https://svnweb.freebsd.org/changeset/base/366743
Log:
cache: support negative entry promotion in slowpath smr
Modified:
head/sys/kern/vfs_cache.c
Modified: head/sys/kern/vfs_cache.c
==
--- head/sys/kern/vfs_cache.c Fri Oct 16 00:55:57 2020(r366742)
+++ head/sys/kern/vfs_cache.c Fri Oct 16 00:56:13 2020(r366743)
@@ -806,6 +806,9 @@ cache_negative_init(struct namecache *ncp)
ns->neg_flag = 0;
}
+/*
+ * Move a negative entry to the hot list.
+ */
static void
cache_negative_promote(struct namecache *ncp)
{
@@ -823,6 +826,87 @@ cache_negative_promote(struct namecache *ncp)
}
}
+/*
+ * Move a negative entry to the hot list if it matches the lookup.
+ *
+ * We have to take locks, but they may be contended and in the worst
+ * case we may need to go off CPU. We don't want to spin within the
+ * smr section and we can't block with it. Exiting the section means
+ * the found entry could have been evicted. We are going to look it
+ * up again.
+ */
+static bool
+cache_negative_promote_cond(struct vnode *dvp, struct componentname *cnp,
+struct namecache *oncp, uint32_t hash)
+{
+ struct namecache *ncp;
+ struct neglist *nl;
+ u_char nc_flag;
+
+ nl = NCP2NEGLIST(oncp);
+
+ mtx_lock(&nl->nl_lock);
+ /*
+* For hash iteration.
+*/
+ vfs_smr_enter();
+
+ /*
+* Avoid all surprises by only succeeding if we got the same entry and
+* bailing completely otherwise.
+* XXX There are no provisions to keep the vnode around, meaning we may
+* end up promoting a negative entry for a *new* vnode and returning
+* ENOENT on its account. This is the error we want to return anyway
+* and promotion is harmless.
+*
+* In particular at this point there can be a new ncp which matches the
+* search but hashes to a different neglist.
+*/
+ CK_SLIST_FOREACH(ncp, (NCHHASH(hash)), nc_hash) {
+ if (ncp == oncp)
+ break;
+ }
+
+ /*
+* No match to begin with.
+*/
+ if (__predict_false(ncp == NULL)) {
+ goto out_abort;
+ }
+
+ /*
+* The newly found entry may be something different...
+*/
+ if (!(ncp->nc_dvp == dvp && ncp->nc_nlen == cnp->cn_namelen &&
+ !bcmp(ncp->nc_name, cnp->cn_nameptr, ncp->nc_nlen))) {
+ goto out_abort;
+ }
+
+ /*
+* ... and not even negative.
+*/
+ nc_flag = atomic_load_char(&ncp->nc_flag);
+ if ((nc_flag & NCF_NEGATIVE) == 0) {
+ goto out_abort;
+ }
+
+ if (__predict_false(!cache_ncp_canuse(ncp))) {
+ goto out_abort;
+ }
+
+ cache_negative_promote(ncp);
+
+ SDT_PROBE2(vfs, namecache, lookup, hit__negative, dvp, ncp->nc_name);
+ counter_u64_add(numneghits, 1);
+ vfs_smr_exit();
+ mtx_unlock(&nl->nl_lock);
+ return (true);
+out_abort:
+ vfs_smr_exit();
+ mtx_unlock(&nl->nl_lock);
+ return (false);
+}
+
static void
cache_negative_hit(struct namecache *ncp)
{
@@ -1455,7 +1539,7 @@ cache_lookup(struct vnode *dvp, struct vnode **vpp, st
uint32_t hash;
enum vgetstate vs;
int error;
- bool whiteout;
+ bool whiteout, neg_hot;
u_short nc_flag;
MPASS((tsp == NULL && ticksp == NULL) || (tsp != NULL && ticksp !=
NULL));
@@ -1532,21 +1616,23 @@ negative_success:
}
}
- SDT_PROBE2(vfs, namecache, lookup, hit__negative, dvp, ncp->nc_name);
cache_out_ts(ncp, tsp, ticksp);
- counter_u64_add(numneghits, 1);
whiteout = (ncp->nc_flag & NCF_WHITE);
- /*
-* TODO: We need to take locks to promote an entry. Code doing it
-* in SMR lookup can be modified to be shared.
-*/
ns = NCP2NEGSTATE(ncp);
- if ((ns->neg_flag & NEG_HOT) == 0 ||
- !cache_ncp_canuse(ncp)) {
+ neg_hot = ((ns->neg_flag & NEG_HOT) != 0);
+ if (__predict_false(!cache_ncp_canuse(ncp))) {
vfs_smr_exit();
goto out_fallback;
}
- vfs_smr_exit();
+ if (neg_hot) {
+ vfs_smr_exit();
+ if (!cache_negative_promote_cond(dvp, cnp, ncp, hash))
+ goto out_fallback;
+ } else {
+ SDT_PROBE2(vfs, namecache, lookup, hit__negative, dvp,
ncp->nc_name);
+ counter_u64_add(numneghits, 1);
+ vfs_smr_exit();
+ }
if (whiteout)
cnp->cn_flags |= ISWHITEOUT;
return (ENOENT);
@@ -3373,90 +3459,21 @@ cache_fplookup_vnode_supported(struct vnode *vp)
return (vp->v_type != VLNK);
}
-/*
- * Move a negative entry to the hot list.
svn commit: r366741 - head/sys/kern
Author: mjg
Date: Fri Oct 16 00:55:31 2020
New Revision: 366741
URL: https://svnweb.freebsd.org/changeset/base/366741
Log:
cache: dedup code for negative promotion
Modified:
head/sys/kern/vfs_cache.c
Modified: head/sys/kern/vfs_cache.c
==
--- head/sys/kern/vfs_cache.c Fri Oct 16 00:55:09 2020(r366740)
+++ head/sys/kern/vfs_cache.c Fri Oct 16 00:55:31 2020(r366741)
@@ -807,22 +807,34 @@ cache_negative_init(struct namecache *ncp)
}
static void
-cache_negative_hit(struct namecache *ncp)
+cache_negative_promote(struct namecache *ncp)
{
struct neglist *nl;
struct negstate *ns;
ns = NCP2NEGSTATE(ncp);
- if ((ns->neg_flag & NEG_HOT) != 0)
- return;
nl = NCP2NEGLIST(ncp);
- mtx_lock(&nl->nl_lock);
+ mtx_assert(&nl->nl_lock, MA_OWNED);
if ((ns->neg_flag & NEG_HOT) == 0) {
TAILQ_REMOVE(&nl->nl_list, ncp, nc_dst);
TAILQ_INSERT_TAIL(&nl->nl_hotlist, ncp, nc_dst);
nl->nl_hotnum++;
ns->neg_flag |= NEG_HOT;
}
+}
+
+static void
+cache_negative_hit(struct namecache *ncp)
+{
+ struct neglist *nl;
+ struct negstate *ns;
+
+ ns = NCP2NEGSTATE(ncp);
+ if ((ns->neg_flag & NEG_HOT) != 0)
+ return;
+ nl = NCP2NEGLIST(ncp);
+ mtx_lock(&nl->nl_lock);
+ cache_negative_promote(ncp);
mtx_unlock(&nl->nl_lock);
}
@@ -3376,7 +3388,6 @@ cache_fplookup_negative_promote(struct cache_fpl *fpl,
struct componentname *cnp;
struct namecache *ncp;
struct neglist *nl;
- struct negstate *ns;
struct vnode *dvp;
u_char nc_flag;
@@ -3434,13 +3445,7 @@ cache_fplookup_negative_promote(struct cache_fpl *fpl,
goto out_abort;
}
- ns = NCP2NEGSTATE(ncp);
- if ((ns->neg_flag & NEG_HOT) == 0) {
- TAILQ_REMOVE(&nl->nl_list, ncp, nc_dst);
- TAILQ_INSERT_TAIL(&nl->nl_hotlist, ncp, nc_dst);
- nl->nl_hotnum++;
- ns->neg_flag |= NEG_HOT;
- }
+ cache_negative_promote(ncp);
SDT_PROBE2(vfs, namecache, lookup, hit__negative, dvp, ncp->nc_name);
counter_u64_add(numneghits, 1);
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366742 - head/sys/kern
Author: mjg Date: Fri Oct 16 00:55:57 2020 New Revision: 366742 URL: https://svnweb.freebsd.org/changeset/base/366742 Log: cache: elide vhold/vdrop around promoting negative entry Modified: head/sys/kern/vfs_cache.c Modified: head/sys/kern/vfs_cache.c == --- head/sys/kern/vfs_cache.c Fri Oct 16 00:55:31 2020(r366741) +++ head/sys/kern/vfs_cache.c Fri Oct 16 00:55:57 2020(r366742) @@ -3394,9 +3394,6 @@ cache_fplookup_negative_promote(struct cache_fpl *fpl, cnp = fpl->cnp; dvp = fpl->dvp; - if (!vhold_smr(dvp)) - return (cache_fpl_aborted(fpl)); - nl = NCP2NEGLIST(oncp); cache_fpl_smr_exit(fpl); @@ -3409,6 +3406,10 @@ cache_fplookup_negative_promote(struct cache_fpl *fpl, /* * Avoid all surprises by only succeeding if we got the same entry and * bailing completely otherwise. +* XXX There are no provisions to keep the vnode around, meaning we may +* end up promoting a negative entry for a *new* vnode and returning +* ENOENT on its account. This is the error we want to return anyway +* and promotion is harmless. * * In particular at this point there can be a new ncp which matches the * search but hashes to a different neglist. @@ -3451,12 +3452,10 @@ cache_fplookup_negative_promote(struct cache_fpl *fpl, counter_u64_add(numneghits, 1); cache_fpl_smr_exit(fpl); mtx_unlock(&nl->nl_lock); - vdrop(dvp); return (cache_fpl_handled(fpl, ENOENT)); out_abort: cache_fpl_smr_exit(fpl); mtx_unlock(&nl->nl_lock); - vdrop(dvp); return (cache_fpl_aborted(fpl)); } ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366740 - head/sys/kern
Author: mjg
Date: Fri Oct 16 00:55:09 2020
New Revision: 366740
URL: https://svnweb.freebsd.org/changeset/base/366740
Log:
cache: neglist -> nl; negstate -> ns
No functional changes.
Modified:
head/sys/kern/vfs_cache.c
Modified: head/sys/kern/vfs_cache.c
==
--- head/sys/kern/vfs_cache.c Fri Oct 16 00:01:01 2020(r366739)
+++ head/sys/kern/vfs_cache.c Fri Oct 16 00:55:09 2020(r366740)
@@ -799,85 +799,85 @@ SYSCTL_PROC(_debug_hashstat, OID_AUTO, nchash, CTLTYPE
static void
cache_negative_init(struct namecache *ncp)
{
- struct negstate *negstate;
+ struct negstate *ns;
ncp->nc_flag |= NCF_NEGATIVE;
- negstate = NCP2NEGSTATE(ncp);
- negstate->neg_flag = 0;
+ ns = NCP2NEGSTATE(ncp);
+ ns->neg_flag = 0;
}
static void
cache_negative_hit(struct namecache *ncp)
{
- struct neglist *neglist;
- struct negstate *negstate;
+ struct neglist *nl;
+ struct negstate *ns;
- negstate = NCP2NEGSTATE(ncp);
- if ((negstate->neg_flag & NEG_HOT) != 0)
+ ns = NCP2NEGSTATE(ncp);
+ if ((ns->neg_flag & NEG_HOT) != 0)
return;
- neglist = NCP2NEGLIST(ncp);
- mtx_lock(&neglist->nl_lock);
- if ((negstate->neg_flag & NEG_HOT) == 0) {
- TAILQ_REMOVE(&neglist->nl_list, ncp, nc_dst);
- TAILQ_INSERT_TAIL(&neglist->nl_hotlist, ncp, nc_dst);
- neglist->nl_hotnum++;
- negstate->neg_flag |= NEG_HOT;
+ nl = NCP2NEGLIST(ncp);
+ mtx_lock(&nl->nl_lock);
+ if ((ns->neg_flag & NEG_HOT) == 0) {
+ TAILQ_REMOVE(&nl->nl_list, ncp, nc_dst);
+ TAILQ_INSERT_TAIL(&nl->nl_hotlist, ncp, nc_dst);
+ nl->nl_hotnum++;
+ ns->neg_flag |= NEG_HOT;
}
- mtx_unlock(&neglist->nl_lock);
+ mtx_unlock(&nl->nl_lock);
}
static void
cache_negative_insert(struct namecache *ncp)
{
- struct neglist *neglist;
+ struct neglist *nl;
MPASS(ncp->nc_flag & NCF_NEGATIVE);
cache_assert_bucket_locked(ncp);
- neglist = NCP2NEGLIST(ncp);
- mtx_lock(&neglist->nl_lock);
- TAILQ_INSERT_TAIL(&neglist->nl_list, ncp, nc_dst);
- mtx_unlock(&neglist->nl_lock);
+ nl = NCP2NEGLIST(ncp);
+ mtx_lock(&nl->nl_lock);
+ TAILQ_INSERT_TAIL(&nl->nl_list, ncp, nc_dst);
+ mtx_unlock(&nl->nl_lock);
atomic_add_long(&numneg, 1);
}
static void
cache_negative_remove(struct namecache *ncp)
{
- struct neglist *neglist;
- struct negstate *negstate;
+ struct neglist *nl;
+ struct negstate *ns;
cache_assert_bucket_locked(ncp);
- neglist = NCP2NEGLIST(ncp);
- negstate = NCP2NEGSTATE(ncp);
- mtx_lock(&neglist->nl_lock);
- if ((negstate->neg_flag & NEG_HOT) != 0) {
- TAILQ_REMOVE(&neglist->nl_hotlist, ncp, nc_dst);
- neglist->nl_hotnum--;
+ nl = NCP2NEGLIST(ncp);
+ ns = NCP2NEGSTATE(ncp);
+ mtx_lock(&nl->nl_lock);
+ if ((ns->neg_flag & NEG_HOT) != 0) {
+ TAILQ_REMOVE(&nl->nl_hotlist, ncp, nc_dst);
+ nl->nl_hotnum--;
} else {
- TAILQ_REMOVE(&neglist->nl_list, ncp, nc_dst);
+ TAILQ_REMOVE(&nl->nl_list, ncp, nc_dst);
}
- mtx_unlock(&neglist->nl_lock);
+ mtx_unlock(&nl->nl_lock);
atomic_subtract_long(&numneg, 1);
}
static struct neglist *
cache_negative_shrink_select(void)
{
- struct neglist *neglist;
+ struct neglist *nl;
static u_int cycle;
u_int i;
cycle++;
for (i = 0; i < numneglists; i++) {
- neglist = &neglists[(cycle + i) % numneglists];
- if (TAILQ_FIRST(&neglist->nl_list) == NULL &&
- TAILQ_FIRST(&neglist->nl_hotlist) == NULL)
+ nl = &neglists[(cycle + i) % numneglists];
+ if (TAILQ_FIRST(&nl->nl_list) == NULL &&
+ TAILQ_FIRST(&nl->nl_hotlist) == NULL)
continue;
- mtx_lock(&neglist->nl_lock);
- if (TAILQ_FIRST(&neglist->nl_list) != NULL ||
- TAILQ_FIRST(&neglist->nl_hotlist) != NULL)
- return (neglist);
- mtx_unlock(&neglist->nl_lock);
+ mtx_lock(&nl->nl_lock);
+ if (TAILQ_FIRST(&nl->nl_list) != NULL ||
+ TAILQ_FIRST(&nl->nl_hotlist) != NULL)
+ return (nl);
+ mtx_unlock(&nl->nl_lock);
}
return (NULL);
@@ -887,8 +887,8 @@ static void
cache_negative_zap_one(void)
{
struct namecache *ncp, *ncp2;
- struct neglist *neglist;
- struct negstate *negstate;
+ struct neglist *nl;
+ struct negstate *ns;
struct mtx *dvlp;
struct mtx *blp;
@@ -898,26 +898,26
svn commit: r366739 - releng/12.2/sys/conf
Author: gjb
Date: Fri Oct 16 00:01:01 2020
New Revision: 366739
URL: https://svnweb.freebsd.org/changeset/base/366739
Log:
Update releng/12.2 to RC3 as part of the 12.2-RELEASE cycle.
Approved by: re (implicit)
Sponsored by: Rubicon Communications, LLC (netgate.com)
Modified:
releng/12.2/sys/conf/newvers.sh
Modified: releng/12.2/sys/conf/newvers.sh
==
--- releng/12.2/sys/conf/newvers.sh Thu Oct 15 23:05:13 2020
(r366738)
+++ releng/12.2/sys/conf/newvers.sh Fri Oct 16 00:01:01 2020
(r366739)
@@ -49,7 +49,7 @@
TYPE="FreeBSD"
REVISION="12.2"
-BRANCH="RC2"
+BRANCH="RC3"
if [ -n "${BRANCH_OVERRIDE}" ]; then
BRANCH=${BRANCH_OVERRIDE}
fi
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366738 - in head/release: amd64 arm64
Author: gjb
Date: Thu Oct 15 23:05:13 2020
New Revision: 366738
URL: https://svnweb.freebsd.org/changeset/base/366738
Log:
Bump the ISO EFI partition size from 1024 to 2048, following r366732.
Suggested by: imp
Sponsored by: Rubicon Communications, LLC (netgate.com)
Modified:
head/release/amd64/mkisoimages.sh
head/release/arm64/mkisoimages.sh
Modified: head/release/amd64/mkisoimages.sh
==
--- head/release/amd64/mkisoimages.sh Thu Oct 15 20:21:15 2020
(r366737)
+++ head/release/amd64/mkisoimages.sh Thu Oct 15 23:05:13 2020
(r366738)
@@ -48,7 +48,7 @@ if [ "$1" = "-b" ]; then
# Make EFI system partition.
espfilename=$(mktemp /tmp/efiboot.XX)
# ESP file size in KB.
- espsize="1024"
+ espsize="2048"
make_esp_file ${espfilename} ${espsize} ${BASEBITSDIR}/boot/loader.efi
bootable="$bootable -o bootimage=i386;${espfilename} -o no-emul-boot -o
platformid=efi"
Modified: head/release/arm64/mkisoimages.sh
==
--- head/release/arm64/mkisoimages.sh Thu Oct 15 20:21:15 2020
(r366737)
+++ head/release/arm64/mkisoimages.sh Thu Oct 15 23:05:13 2020
(r366738)
@@ -42,7 +42,7 @@ if [ "$1" = "-b" ]; then
# Make an EFI system partition.
espfilename=$(mktemp /tmp/efiboot.XX)
# ESP file size in KB.
- espsize="1024"
+ espsize="2048"
make_esp_file ${espfilename} ${espsize} ${BASEBITSDIR}/boot/loader.efi
bootable="-o bootimage=efi;${espfilename} -o no-emul-boot -o
platformid=efi"
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366737 - in head/sys: amd64/amd64 arm/arm arm64/arm64 riscv/riscv
Author: mhorne
Date: Thu Oct 15 20:21:15 2020
New Revision: 366737
URL: https://svnweb.freebsd.org/changeset/base/366737
Log:
Simplify preload_dump() condition
Hiding this feature behind RB_VERBOSE is gratuitous. The tunable is enough
to limit its use to only those who explicitly request it.
Suggested by: kevans
Modified:
head/sys/amd64/amd64/machdep.c
head/sys/arm/arm/machdep.c
head/sys/arm64/arm64/machdep.c
head/sys/riscv/riscv/machdep.c
Modified: head/sys/amd64/amd64/machdep.c
==
--- head/sys/amd64/amd64/machdep.c Thu Oct 15 18:03:14 2020
(r366736)
+++ head/sys/amd64/amd64/machdep.c Thu Oct 15 20:21:15 2020
(r366737)
@@ -1859,8 +1859,7 @@ hammer_time(u_int64_t modulep, u_int64_t physfree)
* output is required. If it's grossly incorrect the kernel will never
* make it this far.
*/
- if ((boothowto & RB_VERBOSE) &&
- getenv_is_true("debug.dump_modinfo_at_boot"))
+ if (getenv_is_true("debug.dump_modinfo_at_boot"))
preload_dump();
#ifdef DEV_ISA
Modified: head/sys/arm/arm/machdep.c
==
--- head/sys/arm/arm/machdep.c Thu Oct 15 18:03:14 2020(r366736)
+++ head/sys/arm/arm/machdep.c Thu Oct 15 20:21:15 2020(r366737)
@@ -1032,8 +1032,7 @@ initarm(struct arm_boot_params *abp)
* output is required. If it's grossly incorrect the kernel will never
* make it this far.
*/
- if ((boothowto & RB_VERBOSE) &&
- getenv_is_true("debug.dump_modinfo_at_boot"))
+ if (getenv_is_true("debug.dump_modinfo_at_boot"))
preload_dump();
env = kern_getenv("kernelname");
Modified: head/sys/arm64/arm64/machdep.c
==
--- head/sys/arm64/arm64/machdep.c Thu Oct 15 18:03:14 2020
(r366736)
+++ head/sys/arm64/arm64/machdep.c Thu Oct 15 20:21:15 2020
(r366737)
@@ -1247,8 +1247,7 @@ initarm(struct arm64_bootparams *abp)
* output is required. If it's grossly incorrect the kernel will never
* make it this far.
*/
- if ((boothowto & RB_VERBOSE) &&
- getenv_is_true("debug.dump_modinfo_at_boot"))
+ if (getenv_is_true("debug.dump_modinfo_at_boot"))
preload_dump();
init_proc0(abp->kern_stack);
Modified: head/sys/riscv/riscv/machdep.c
==
--- head/sys/riscv/riscv/machdep.c Thu Oct 15 18:03:14 2020
(r366736)
+++ head/sys/riscv/riscv/machdep.c Thu Oct 15 20:21:15 2020
(r366737)
@@ -954,8 +954,7 @@ initriscv(struct riscv_bootparams *rvbp)
* output is required. If it's grossly incorrect the kernel will never
* make it this far.
*/
- if ((boothowto & RB_VERBOSE) &&
- getenv_is_true("debug.dump_modinfo_at_boot"))
+ if (getenv_is_true("debug.dump_modinfo_at_boot"))
preload_dump();
init_proc0(rvbp->kern_stack);
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
Re: svn commit: r366732 - in head/release: amd64 arm64
And on a related note, I can make it so we can build w/o ZSTD in the CD
boot loader because it's not needed there. I'd rather not, but if the 800k
limit is a real thing in relevant places, we have that as an option.
Warner
On Thu, Oct 15, 2020 at 1:34 PM Warner Losh wrote:
> I'd recommend bumping this to 2MB or larger.
>
> This most likely is from the zstd landing which pushed arm64's size over
> the 1MB limit in memory, so it's only a matter of time before the file
> grows past 1MB in size.
>
> Warner
>
> On Thu, Oct 15, 2020 at 11:13 AM Glen Barber wrote:
>
>> Author: gjb
>> Date: Thu Oct 15 17:12:58 2020
>> New Revision: 366732
>> URL: https://svnweb.freebsd.org/changeset/base/366732
>>
>> Log:
>> Increase the amd64 ISO ESP file size from 800KB to 1024KB.
>>
>> At some poing over the last week, the bootx64.efi file has grown
>> past the 800KB threshold, resulting in being unable to copy it to
>> the EFI/BOOT directory.
>>
>># stat -f %z efiboot.znWo7m
>>819200
>># stat -f %z stand-test.PIEugN/EFI/BOOT/bootx64.efi
>>842752
>>
>> The comment in the script that creates the ISOs suggests that 800KB
>> is the maximum allowed for the boot code, however I was able to
>> boot an ISO with a 1024KB boot partition. Additionally, I verified
>> against an ISO from OtherOS, where the boot EFI partition is 2.4MB.
>>
>> Sponsored by: Rubicon Communications, LLC (netgate.com)
>>
>> Modified:
>> head/release/amd64/mkisoimages.sh
>> head/release/arm64/mkisoimages.sh
>>
>> Modified: head/release/amd64/mkisoimages.sh
>>
>> ==
>> --- head/release/amd64/mkisoimages.sh Thu Oct 15 17:05:21 2020
>> (r366731)
>> +++ head/release/amd64/mkisoimages.sh Thu Oct 15 17:12:58 2020
>> (r366732)
>> @@ -46,10 +46,10 @@ if [ "$1" = "-b" ]; then
>> bootable="-o bootimage=i386;$BASEBITSDIR/boot/cdboot -o
>> no-emul-boot"
>>
>> # Make EFI system partition.
>> - # The ISO file is a special case, in that it only has a maximum of
>> - # 800 KB available for the boot code. So make an 800 KB ESP
>> espfilename=$(mktemp /tmp/efiboot.XX)
>> - make_esp_file ${espfilename} 800 ${BASEBITSDIR}/boot/loader.efi
>> + # ESP file size in KB.
>> + espsize="1024"
>> + make_esp_file ${espfilename} ${espsize}
>> ${BASEBITSDIR}/boot/loader.efi
>> bootable="$bootable -o bootimage=i386;${espfilename} -o
>> no-emul-boot -o platformid=efi"
>>
>> shift
>>
>> Modified: head/release/arm64/mkisoimages.sh
>>
>> ==
>> --- head/release/arm64/mkisoimages.sh Thu Oct 15 17:05:21 2020
>> (r366731)
>> +++ head/release/arm64/mkisoimages.sh Thu Oct 15 17:12:58 2020
>> (r366732)
>> @@ -40,10 +40,10 @@ if [ "$1" = "-b" ]; then
>> BASEBITSDIR="$4"
>>
>> # Make an EFI system partition.
>> - # The ISO file is a special case, in that it only has a maximum of
>> - # 800 KB available for the boot code. So make an 800 KB ESP
>> espfilename=$(mktemp /tmp/efiboot.XX)
>> - make_esp_file ${espfilename} 800 ${BASEBITSDIR}/boot/loader.efi
>> + # ESP file size in KB.
>> + espsize="1024"
>> + make_esp_file ${espfilename} ${espsize}
>> ${BASEBITSDIR}/boot/loader.efi
>>
>> bootable="-o bootimage=efi;${espfilename} -o no-emul-boot -o
>> platformid=efi"
>>
>>
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
Re: svn commit: r366732 - in head/release: amd64 arm64
I'd recommend bumping this to 2MB or larger.
This most likely is from the zstd landing which pushed arm64's size over
the 1MB limit in memory, so it's only a matter of time before the file
grows past 1MB in size.
Warner
On Thu, Oct 15, 2020 at 11:13 AM Glen Barber wrote:
> Author: gjb
> Date: Thu Oct 15 17:12:58 2020
> New Revision: 366732
> URL: https://svnweb.freebsd.org/changeset/base/366732
>
> Log:
> Increase the amd64 ISO ESP file size from 800KB to 1024KB.
>
> At some poing over the last week, the bootx64.efi file has grown
> past the 800KB threshold, resulting in being unable to copy it to
> the EFI/BOOT directory.
>
># stat -f %z efiboot.znWo7m
>819200
># stat -f %z stand-test.PIEugN/EFI/BOOT/bootx64.efi
>842752
>
> The comment in the script that creates the ISOs suggests that 800KB
> is the maximum allowed for the boot code, however I was able to
> boot an ISO with a 1024KB boot partition. Additionally, I verified
> against an ISO from OtherOS, where the boot EFI partition is 2.4MB.
>
> Sponsored by: Rubicon Communications, LLC (netgate.com)
>
> Modified:
> head/release/amd64/mkisoimages.sh
> head/release/arm64/mkisoimages.sh
>
> Modified: head/release/amd64/mkisoimages.sh
>
> ==
> --- head/release/amd64/mkisoimages.sh Thu Oct 15 17:05:21 2020
> (r366731)
> +++ head/release/amd64/mkisoimages.sh Thu Oct 15 17:12:58 2020
> (r366732)
> @@ -46,10 +46,10 @@ if [ "$1" = "-b" ]; then
> bootable="-o bootimage=i386;$BASEBITSDIR/boot/cdboot -o
> no-emul-boot"
>
> # Make EFI system partition.
> - # The ISO file is a special case, in that it only has a maximum of
> - # 800 KB available for the boot code. So make an 800 KB ESP
> espfilename=$(mktemp /tmp/efiboot.XX)
> - make_esp_file ${espfilename} 800 ${BASEBITSDIR}/boot/loader.efi
> + # ESP file size in KB.
> + espsize="1024"
> + make_esp_file ${espfilename} ${espsize}
> ${BASEBITSDIR}/boot/loader.efi
> bootable="$bootable -o bootimage=i386;${espfilename} -o
> no-emul-boot -o platformid=efi"
>
> shift
>
> Modified: head/release/arm64/mkisoimages.sh
>
> ==
> --- head/release/arm64/mkisoimages.sh Thu Oct 15 17:05:21 2020
> (r366731)
> +++ head/release/arm64/mkisoimages.sh Thu Oct 15 17:12:58 2020
> (r366732)
> @@ -40,10 +40,10 @@ if [ "$1" = "-b" ]; then
> BASEBITSDIR="$4"
>
> # Make an EFI system partition.
> - # The ISO file is a special case, in that it only has a maximum of
> - # 800 KB available for the boot code. So make an 800 KB ESP
> espfilename=$(mktemp /tmp/efiboot.XX)
> - make_esp_file ${espfilename} 800 ${BASEBITSDIR}/boot/loader.efi
> + # ESP file size in KB.
> + espsize="1024"
> + make_esp_file ${espfilename} ${espsize}
> ${BASEBITSDIR}/boot/loader.efi
>
> bootable="-o bootimage=efi;${espfilename} -o no-emul-boot -o
> platformid=efi"
>
>
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366736 - head/usr.sbin/kldxref
Author: jrtc27
Date: Thu Oct 15 18:03:14 2020
New Revision: 366736
URL: https://svnweb.freebsd.org/changeset/base/366736
Log:
kldxref: Avoid buffer overflows in parse_pnp_list
We convert a string like "W32:vendor/device" into "I:vendor;I:device",
where the output is longer than the input, but only allocate space equal
to the length of the input, leading to a buffer overflow.
Instead use open_memstream so we get a safe dynamically-grown buffer.
Found by: CHERI
Reviewed by: imp, jhb (mentor)
Approved by: imp, jhb (mentor)
Obtained from:CheriBSD
Differential Revision:https://reviews.freebsd.org/D26637
Modified:
head/usr.sbin/kldxref/kldxref.c
Modified: head/usr.sbin/kldxref/kldxref.c
==
--- head/usr.sbin/kldxref/kldxref.c Thu Oct 15 17:44:17 2020
(r366735)
+++ head/usr.sbin/kldxref/kldxref.c Thu Oct 15 18:03:14 2020
(r366736)
@@ -235,14 +235,17 @@ parse_pnp_list(const char *desc, char **new_desc, pnp_
const char *walker, *ep;
const char *colon, *semi;
struct pnp_elt *elt;
- char *nd;
char type[8], key[32];
int off;
+ size_t new_desc_size;
+ FILE *fp;
walker = desc;
ep = desc + strlen(desc);
off = 0;
- nd = *new_desc = malloc(strlen(desc) + 1);
+ fp = open_memstream(new_desc, &new_desc_size);
+ if (fp == NULL)
+ err(1, "Could not open new memory stream");
if (verbose > 1)
printf("Converting %s into a list\n", desc);
while (walker < ep) {
@@ -336,42 +339,44 @@ parse_pnp_list(const char *desc, char **new_desc, pnp_
off = elt->pe_offset + sizeof(void *);
}
if (elt->pe_kind & TYPE_PAIRED) {
- char *word, *ctx;
+ char *word, *ctx, newtype;
for (word = strtok_r(key, "/", &ctx);
word; word = strtok_r(NULL, "/", &ctx)) {
- sprintf(nd, "%c:%s;", elt->pe_kind &
TYPE_FLAGGED ? 'J' : 'I',
- word);
- nd += strlen(nd);
+ newtype = elt->pe_kind & TYPE_FLAGGED ? 'J' :
'I';
+ fprintf(fp, "%c:%s;", newtype, word);
}
-
}
else {
+ char newtype;
+
if (elt->pe_kind & TYPE_FLAGGED)
- *nd++ = 'J';
+ newtype = 'J';
else if (elt->pe_kind & TYPE_GE)
- *nd++ = 'G';
+ newtype = 'G';
else if (elt->pe_kind & TYPE_LE)
- *nd++ = 'L';
+ newtype = 'L';
else if (elt->pe_kind & TYPE_MASK)
- *nd++ = 'M';
+ newtype = 'M';
else if (elt->pe_kind & TYPE_INT)
- *nd++ = 'I';
+ newtype = 'I';
else if (elt->pe_kind == TYPE_D)
- *nd++ = 'D';
+ newtype = 'D';
else if (elt->pe_kind == TYPE_Z || elt->pe_kind ==
TYPE_E)
- *nd++ = 'Z';
+ newtype = 'Z';
else if (elt->pe_kind == TYPE_T)
- *nd++ = 'T';
+ newtype = 'T';
else
errx(1, "Impossible type %x\n", elt->pe_kind);
- *nd++ = ':';
- strcpy(nd, key);
- nd += strlen(nd);
- *nd++ = ';';
+ fprintf(fp, "%c:%s;", newtype, key);
}
}
- *nd++ = '\0';
+ if (ferror(fp) != 0) {
+ fclose(fp);
+ errx(1, "Exhausted space converting description %s", desc);
+ }
+ if (fclose(fp) != 0)
+ errx(1, "Failed to close memory stream");
return (0);
err:
errx(1, "Parse error of description string %s", desc);
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366735 - head/sys/kern
Author: mjg
Date: Thu Oct 15 17:44:17 2020
New Revision: 366735
URL: https://svnweb.freebsd.org/changeset/base/366735
Log:
cache: split hotlist between existing negative lists
This simplifies the code while allowing for concurrent negative eviction
down the road.
Cache misses increased slightly due to higher rate of evictions allowed by
the change.
The current algorithm remains too aggressive.
Modified:
head/sys/kern/vfs_cache.c
Modified: head/sys/kern/vfs_cache.c
==
--- head/sys/kern/vfs_cache.c Thu Oct 15 17:42:22 2020(r366734)
+++ head/sys/kern/vfs_cache.c Thu Oct 15 17:44:17 2020(r366735)
@@ -311,11 +311,11 @@ static struct mtx __exclusive_cache_line ncneg_shrink_
struct neglist {
struct mtx nl_lock;
TAILQ_HEAD(, namecache) nl_list;
+ TAILQ_HEAD(, namecache) nl_hotlist;
+ u_long nl_hotnum;
} __aligned(CACHE_LINE_SIZE);
static struct neglist neglists[numneglists];
-static struct neglist ncneg_hot;
-static u_long numhotneg;
static inline struct neglist *
NCP2NEGLIST(struct namecache *ncp)
@@ -471,7 +471,6 @@ static long zap_and_exit_bucket_fail2; STATNODE_ULONG(
static long cache_lock_vnodes_cel_3_failures;
STATNODE_ULONG(cache_lock_vnodes_cel_3_failures,
"Number of times 3-way vnode locking failed");
-STATNODE_ULONG(numhotneg, "Number of hot negative entries");
STATNODE_COUNTER(numneg_evicted,
"Number of negative entries evicted when adding a new entry");
STATNODE_COUNTER(shrinking_skipped,
@@ -682,6 +681,21 @@ SYSCTL_PROC(_vfs_cache, OID_AUTO, nchstats, CTLTYPE_OP
CTLFLAG_MPSAFE, 0, 0, sysctl_nchstats, "LU",
"VFS cache effectiveness statistics");
+static int
+sysctl_hotnum(SYSCTL_HANDLER_ARGS)
+{
+ int i, out;
+
+ out = 0;
+ for (i = 0; i < numneglists; i++)
+ out += neglists[i].nl_hotnum;
+
+ return (SYSCTL_OUT(req, &out, sizeof(out)));
+}
+SYSCTL_PROC(_vfs_cache, OID_AUTO, hotnum, CTLTYPE_INT | CTLFLAG_RD |
+CTLFLAG_MPSAFE, 0, 0, sysctl_hotnum, "I",
+"Number of hot negative entries");
+
#ifdef DIAGNOSTIC
/*
* Grab an atomic snapshot of the name cache hash chain lengths
@@ -802,16 +816,14 @@ cache_negative_hit(struct namecache *ncp)
if ((negstate->neg_flag & NEG_HOT) != 0)
return;
neglist = NCP2NEGLIST(ncp);
- mtx_lock(&ncneg_hot.nl_lock);
mtx_lock(&neglist->nl_lock);
if ((negstate->neg_flag & NEG_HOT) == 0) {
- numhotneg++;
TAILQ_REMOVE(&neglist->nl_list, ncp, nc_dst);
- TAILQ_INSERT_TAIL(&ncneg_hot.nl_list, ncp, nc_dst);
+ TAILQ_INSERT_TAIL(&neglist->nl_hotlist, ncp, nc_dst);
+ neglist->nl_hotnum++;
negstate->neg_flag |= NEG_HOT;
}
mtx_unlock(&neglist->nl_lock);
- mtx_unlock(&ncneg_hot.nl_lock);
}
static void
@@ -833,72 +845,42 @@ cache_negative_remove(struct namecache *ncp)
{
struct neglist *neglist;
struct negstate *negstate;
- bool hot_locked = false;
- bool list_locked = false;
cache_assert_bucket_locked(ncp);
neglist = NCP2NEGLIST(ncp);
negstate = NCP2NEGSTATE(ncp);
+ mtx_lock(&neglist->nl_lock);
if ((negstate->neg_flag & NEG_HOT) != 0) {
- hot_locked = true;
- mtx_lock(&ncneg_hot.nl_lock);
- if ((negstate->neg_flag & NEG_HOT) == 0) {
- list_locked = true;
- mtx_lock(&neglist->nl_lock);
- }
+ TAILQ_REMOVE(&neglist->nl_hotlist, ncp, nc_dst);
+ neglist->nl_hotnum--;
} else {
- list_locked = true;
- mtx_lock(&neglist->nl_lock);
- /*
-* We may be racing against promotion in lockless lookup.
-*/
- if ((negstate->neg_flag & NEG_HOT) != 0) {
- mtx_unlock(&neglist->nl_lock);
- hot_locked = true;
- mtx_lock(&ncneg_hot.nl_lock);
- mtx_lock(&neglist->nl_lock);
- }
- }
- if ((negstate->neg_flag & NEG_HOT) != 0) {
- mtx_assert(&ncneg_hot.nl_lock, MA_OWNED);
- TAILQ_REMOVE(&ncneg_hot.nl_list, ncp, nc_dst);
- numhotneg--;
- } else {
- mtx_assert(&neglist->nl_lock, MA_OWNED);
TAILQ_REMOVE(&neglist->nl_list, ncp, nc_dst);
}
- if (list_locked)
- mtx_unlock(&neglist->nl_lock);
- if (hot_locked)
- mtx_unlock(&ncneg_hot.nl_lock);
+ mtx_unlock(&neglist->nl_lock);
atomic_subtract_long(&numneg, 1);
}
-static void
-cache_negative_shrink_select(struct namecache **ncpp,
-struct neglist **neglistpp)
+static struct neglist *
+cache_negative_shrink_select
svn commit: r366734 - head/sys/kern
Author: mjg
Date: Thu Oct 15 17:42:22 2020
New Revision: 366734
URL: https://svnweb.freebsd.org/changeset/base/366734
Log:
cache: make neglist an array given the static size
Modified:
head/sys/kern/vfs_cache.c
Modified: head/sys/kern/vfs_cache.c
==
--- head/sys/kern/vfs_cache.c Thu Oct 15 17:40:02 2020(r366733)
+++ head/sys/kern/vfs_cache.c Thu Oct 15 17:42:22 2020(r366734)
@@ -305,17 +305,18 @@ SYSCTL_BOOL(_vfs, OID_AUTO, cache_fast_revlookup, CTLF
static struct mtx __exclusive_cache_line ncneg_shrink_lock;
+#define ncneghash 3
+#definenumneglists (ncneghash + 1)
+
struct neglist {
struct mtx nl_lock;
TAILQ_HEAD(, namecache) nl_list;
} __aligned(CACHE_LINE_SIZE);
-static struct neglist __read_mostly*neglists;
+static struct neglist neglists[numneglists];
static struct neglist ncneg_hot;
static u_long numhotneg;
-#define ncneghash 3
-#definenumneglists (ncneghash + 1)
static inline struct neglist *
NCP2NEGLIST(struct namecache *ncp)
{
@@ -2091,8 +2092,6 @@ nchinit(void *dummy __unused)
for (i = 0; i < numvnodelocks; i++)
mtx_init(&vnodelocks[i], "ncvn", NULL, MTX_DUPOK | MTX_RECURSE);
- neglists = malloc(sizeof(*neglists) * numneglists, M_VFSCACHE,
- M_WAITOK | M_ZERO);
for (i = 0; i < numneglists; i++) {
mtx_init(&neglists[i].nl_lock, "ncnegl", NULL, MTX_DEF);
TAILQ_INIT(&neglists[i].nl_list);
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366733 - head/sys/dev/sound/pci/hda
Author: mav Date: Thu Oct 15 17:40:02 2020 New Revision: 366733 URL: https://svnweb.freebsd.org/changeset/base/366733 Log: Drop unsolicited responses to the still attaching CODECs. It is reported to fix kernel panics when early unsolicited responses delivered to the CODEC device not having driver attached yet. PR: 250248 Reported by: Rajeev Pillai Reviewed by: avg MFC after:2 weeks Modified: head/sys/dev/sound/pci/hda/hdac.c Modified: head/sys/dev/sound/pci/hda/hdac.c == --- head/sys/dev/sound/pci/hda/hdac.c Thu Oct 15 17:12:58 2020 (r366732) +++ head/sys/dev/sound/pci/hda/hdac.c Thu Oct 15 17:40:02 2020 (r366733) @@ -990,7 +990,8 @@ hdac_unsolq_flush(struct hdac_softc *sc) sc->unsolq_rp %= HDAC_UNSOLQ_MAX; cad = sc->unsolq[sc->unsolq_rp++]; sc->unsolq_rp %= HDAC_UNSOLQ_MAX; - if ((child = sc->codecs[cad].dev) != NULL) + if ((child = sc->codecs[cad].dev) != NULL && + device_is_attached(child)) HDAC_UNSOL_INTR(child, resp); ret++; } ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366732 - in head/release: amd64 arm64
Author: gjb
Date: Thu Oct 15 17:12:58 2020
New Revision: 366732
URL: https://svnweb.freebsd.org/changeset/base/366732
Log:
Increase the amd64 ISO ESP file size from 800KB to 1024KB.
At some poing over the last week, the bootx64.efi file has grown
past the 800KB threshold, resulting in being unable to copy it to
the EFI/BOOT directory.
# stat -f %z efiboot.znWo7m
819200
# stat -f %z stand-test.PIEugN/EFI/BOOT/bootx64.efi
842752
The comment in the script that creates the ISOs suggests that 800KB
is the maximum allowed for the boot code, however I was able to
boot an ISO with a 1024KB boot partition. Additionally, I verified
against an ISO from OtherOS, where the boot EFI partition is 2.4MB.
Sponsored by: Rubicon Communications, LLC (netgate.com)
Modified:
head/release/amd64/mkisoimages.sh
head/release/arm64/mkisoimages.sh
Modified: head/release/amd64/mkisoimages.sh
==
--- head/release/amd64/mkisoimages.sh Thu Oct 15 17:05:21 2020
(r366731)
+++ head/release/amd64/mkisoimages.sh Thu Oct 15 17:12:58 2020
(r366732)
@@ -46,10 +46,10 @@ if [ "$1" = "-b" ]; then
bootable="-o bootimage=i386;$BASEBITSDIR/boot/cdboot -o no-emul-boot"
# Make EFI system partition.
- # The ISO file is a special case, in that it only has a maximum of
- # 800 KB available for the boot code. So make an 800 KB ESP
espfilename=$(mktemp /tmp/efiboot.XX)
- make_esp_file ${espfilename} 800 ${BASEBITSDIR}/boot/loader.efi
+ # ESP file size in KB.
+ espsize="1024"
+ make_esp_file ${espfilename} ${espsize} ${BASEBITSDIR}/boot/loader.efi
bootable="$bootable -o bootimage=i386;${espfilename} -o no-emul-boot -o
platformid=efi"
shift
Modified: head/release/arm64/mkisoimages.sh
==
--- head/release/arm64/mkisoimages.sh Thu Oct 15 17:05:21 2020
(r366731)
+++ head/release/arm64/mkisoimages.sh Thu Oct 15 17:12:58 2020
(r366732)
@@ -40,10 +40,10 @@ if [ "$1" = "-b" ]; then
BASEBITSDIR="$4"
# Make an EFI system partition.
- # The ISO file is a special case, in that it only has a maximum of
- # 800 KB available for the boot code. So make an 800 KB ESP
espfilename=$(mktemp /tmp/efiboot.XX)
- make_esp_file ${espfilename} 800 ${BASEBITSDIR}/boot/loader.efi
+ # ESP file size in KB.
+ espsize="1024"
+ make_esp_file ${espfilename} ${espsize} ${BASEBITSDIR}/boot/loader.efi
bootable="-o bootimage=efi;${espfilename} -o no-emul-boot -o
platformid=efi"
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366731 - head/sys/kern
Author: brooks
Date: Thu Oct 15 17:05:21 2020
New Revision: 366731
URL: https://svnweb.freebsd.org/changeset/base/366731
Log:
physio: Don't store user addresses in bio_data
Only assign the address from the iovec to bio_data if it is a kernel
address. This was the single place where bio_data stored (however
briefly) a userspace pointer.
Reviewed by: imp, markj
Obtained from:CheriBSD
MFC after:1 week
Sponsored by: DARPA
Differential Revision:https://reviews.freebsd.org/D26783
Modified:
head/sys/kern/kern_physio.c
Modified: head/sys/kern/kern_physio.c
==
--- head/sys/kern/kern_physio.c Thu Oct 15 15:36:08 2020(r366730)
+++ head/sys/kern/kern_physio.c Thu Oct 15 17:05:21 2020(r366731)
@@ -45,7 +45,7 @@ physio(struct cdev *dev, struct uio *uio, int ioflag)
struct buf *pbuf;
struct bio *bp;
struct vm_page **pages;
- caddr_t sa;
+ char *base, *sa;
u_int iolen, poff;
int error, i, npages, maxpages;
vm_prot_t prot;
@@ -140,7 +140,7 @@ physio(struct cdev *dev, struct uio *uio, int ioflag)
curthread->td_ru.ru_oublock++;
}
bp->bio_offset = uio->uio_offset;
- bp->bio_data = uio->uio_iov[i].iov_base;
+ base = uio->uio_iov[i].iov_base;
bp->bio_length = uio->uio_iov[i].iov_len;
if (bp->bio_length > dev->si_iosize_max)
bp->bio_length = dev->si_iosize_max;
@@ -153,13 +153,13 @@ physio(struct cdev *dev, struct uio *uio, int ioflag)
* larger than MAXPHYS - PAGE_SIZE must be
* page aligned or it will be fragmented.
*/
- poff = (vm_offset_t)bp->bio_data & PAGE_MASK;
+ poff = (vm_offset_t)base & PAGE_MASK;
if (pbuf && bp->bio_length + poff > pbuf->b_kvasize) {
if (dev->si_flags & SI_NOSPLIT) {
uprintf("%s: request ptr %p is not "
"on a page boundary; cannot split "
"request\n", devtoname(dev),
- bp->bio_data);
+ base);
error = EFBIG;
goto doerror;
}
@@ -174,7 +174,7 @@ physio(struct cdev *dev, struct uio *uio, int ioflag)
if (pages) {
if ((npages = vm_fault_quick_hold_pages(
&curproc->p_vmspace->vm_map,
- (vm_offset_t)bp->bio_data, bp->bio_length,
+ (vm_offset_t)base, bp->bio_length,
prot, pages, maxpages)) < 0) {
error = EFAULT;
goto doerror;
@@ -190,7 +190,8 @@ physio(struct cdev *dev, struct uio *uio, int ioflag)
bp->bio_data = unmapped_buf;
bp->bio_flags |= BIO_UNMAPPED;
}
- }
+ } else
+ bp->bio_data = base;
csw->d_strategy(bp);
if (uio->uio_rw == UIO_READ)
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366730 - head/sbin/nvmecontrol
Author: mav
Date: Thu Oct 15 15:36:08 2020
New Revision: 366730
URL: https://svnweb.freebsd.org/changeset/base/366730
Log:
Fix nvmecontrol logpage -i parameter.
MFC after:3 days
Modified:
head/sbin/nvmecontrol/logpage.c
Modified: head/sbin/nvmecontrol/logpage.c
==
--- head/sbin/nvmecontrol/logpage.c Thu Oct 15 15:07:25 2020
(r366729)
+++ head/sbin/nvmecontrol/logpage.c Thu Oct 15 15:36:08 2020
(r366730)
@@ -84,7 +84,7 @@ static const struct opts logpage_opts[] = {
"Page to dump"),
OPT("lsp", 'f', arg_uint8, opt, lsp,
"Log Specific Field"),
- OPT("lsi", 'i', arg_uint16, opt, lsp,
+ OPT("lsi", 'i', arg_uint16, opt, lsi,
"Log Specific Identifier"),
OPT("rae", 'r', arg_none, opt, rae,
"Retain Asynchronous Event"),
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366729 - releng/12.2/stand/libsa/zfs
Author: allanjude
Date: Thu Oct 15 15:07:25 2020
New Revision: 366729
URL: https://svnweb.freebsd.org/changeset/base/366729
Log:
ZFS: whitelist zstd and encryption in the loader
MFC r364787:
MFS r366593:
Please note that neither zstd nor encryption is
supported by the loader at this instant. This
change makes it safe to use those features in
one's root pool, but not in one's root dataset.
Approved by: re (gjb)
Modified:
releng/12.2/stand/libsa/zfs/zfsimpl.c
Directory Properties:
releng/12.2/ (props changed)
Modified: releng/12.2/stand/libsa/zfs/zfsimpl.c
==
--- releng/12.2/stand/libsa/zfs/zfsimpl.c Thu Oct 15 14:56:51 2020
(r366728)
+++ releng/12.2/stand/libsa/zfs/zfsimpl.c Thu Oct 15 15:07:25 2020
(r366729)
@@ -127,6 +127,8 @@ static const char *features_for_read[] = {
"com.delphix:device_removal",
"com.delphix:obsolete_counts",
"com.intel:allocation_classes",
+ "org.freebsd:zstd_compress",
+ "com.datto:encryption",
NULL
};
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366728 - head/sbin/pfctl/tests
Author: adrian
Date: Thu Oct 15 14:56:51 2020
New Revision: 366728
URL: https://svnweb.freebsd.org/changeset/base/366728
Log:
[pfctl_tests] Add missing void to empty function declaration
Our gcc-6.4 flags require non-empty function declarations.
Fix this to match the rest of the codebase.
Tested:
* compiled on gcc-6.4 for amd64
Reviewed by: imp
Differential Revision:https://reviews.freebsd.org/D26795
Modified:
head/sbin/pfctl/tests/pfctl_test.c
Modified: head/sbin/pfctl/tests/pfctl_test.c
==
--- head/sbin/pfctl/tests/pfctl_test.c Thu Oct 15 14:55:07 2020
(r366727)
+++ head/sbin/pfctl/tests/pfctl_test.c Thu Oct 15 14:56:51 2020
(r366728)
@@ -68,7 +68,7 @@ __FBSDID("$FreeBSD$");
*/
static bool
-check_pf_module_available()
+check_pf_module_available(void)
{
int modid;
struct module_stat stat;
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366727 - head/contrib/netbsd-tests/lib/libc/sys
Author: adrian Date: Thu Oct 15 14:55:07 2020 New Revision: 366727 URL: https://svnweb.freebsd.org/changeset/base/366727 Log: [tests] Fix itimer test warning-errors on gcc-6.4 This fixes a "suggested parens" compile warning-into-error that shows up on gcc-6.4. Reviewed by: ngie Differential Revision:https://reviews.freebsd.org/D26789 Modified: head/contrib/netbsd-tests/lib/libc/sys/t_getitimer.c Modified: head/contrib/netbsd-tests/lib/libc/sys/t_getitimer.c == --- head/contrib/netbsd-tests/lib/libc/sys/t_getitimer.cThu Oct 15 14:37:51 2020(r366726) +++ head/contrib/netbsd-tests/lib/libc/sys/t_getitimer.cThu Oct 15 14:55:07 2020(r366727) @@ -195,8 +195,8 @@ ATF_TC_BODY(setitimer_old, tc) ATF_REQUIRE(setitimer(ITIMER_REAL, &it, &ot) == 0); #ifdef __FreeBSD__ - ATF_REQUIRE_MSG(ot.it_value.tv_sec < 4 || - ot.it_value.tv_sec == 4 && ot.it_value.tv_usec <= 3, + ATF_REQUIRE_MSG((ot.it_value.tv_sec < 4) || + (ot.it_value.tv_sec == 4 && ot.it_value.tv_usec <= 3), "setitimer(2) returned invalid it_value: %jd %jd", (intmax_t)ot.it_value.tv_sec, (intmax_t)ot.it_value.tv_usec); #else ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
Re: svn commit: r366725 - head/lib/geom/eli
Hi Mateusz, On Thu, Oct 15, 2020 at 02:17:45PM +, Mateusz Piotrowski wrote: > Author: 0mp (doc,ports committer) > Date: Thu Oct 15 14:17:45 2020 > New Revision: 366725 > URL: https://svnweb.freebsd.org/changeset/base/366725 > > Log: > Fix formatting of SYNOPSIS > > There was an unnecessary newline being added before Nm. > > MFC after: 3 days > > Modified: > head/lib/geom/eli/geli.8 > > Modified: head/lib/geom/eli/geli.8 > == > --- head/lib/geom/eli/geli.8 Thu Oct 15 13:47:52 2020(r366724) > +++ head/lib/geom/eli/geli.8 Thu Oct 15 14:17:45 2020(r366725) > @@ -24,7 +24,7 @@ > .\" > .\" $FreeBSD$ > .\" > -.Dd July 22, 2020 > +.Dd October 15, 2020 > .Dt GELI 8 > .Os > .Sh NAME > @@ -45,8 +45,7 @@ to your > geom_eli_load="YES" > .Ed > .Pp > -Usage of the > -.Nm > +.No Usage of the Nm > utility: > .Pp > .Nm The formatting of the SYNOPSIS was correct before this change. It's a common practice in man pages to use The .Nm utility ... Your change is looking somewhat strange, since ".No Usage of the Nm" reads like the man page would renders "Nm", since it is not used as macro. A .Dd bump is also not necessary since no user visible changes were made to the man page. -- Gordon ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366726 - head/sys/conf
Author: emaste Date: Thu Oct 15 14:37:51 2020 New Revision: 366726 URL: https://svnweb.freebsd.org/changeset/base/366726 Log: move vmware pv drivers to sys/conf/files VMware now has arm64 support; move these to MI files in advance of building them on arm64. PR: 250308 Reported by: Vincent Milum Jr MFC after:1 week Sponsored by: The FreeBSD Foundation Modified: head/sys/conf/files head/sys/conf/files.amd64 head/sys/conf/files.i386 Modified: head/sys/conf/files == --- head/sys/conf/files Thu Oct 15 14:17:45 2020(r366725) +++ head/sys/conf/files Thu Oct 15 14:37:51 2020(r366726) @@ -3455,6 +3455,18 @@ dev/virtio/random/virtio_random.coptional virtio_rand dev/virtio/console/virtio_console.coptionalvirtio_console dev/vkbd/vkbd.coptional vkbd dev/vmgenc/vmgenc_acpi.c optional acpi +dev/vmware/vmxnet3/if_vmx.coptional vmx +dev/vmware/vmci/vmci.c optional vmci +dev/vmware/vmci/vmci_datagram.coptional vmci +dev/vmware/vmci/vmci_doorbell.coptional vmci +dev/vmware/vmci/vmci_driver.c optional vmci +dev/vmware/vmci/vmci_event.c optional vmci +dev/vmware/vmci/vmci_hashtable.c optional vmci +dev/vmware/vmci/vmci_kernel_if.c optional vmci +dev/vmware/vmci/vmci_qpair.c optional vmci +dev/vmware/vmci/vmci_queue_pair.c optional vmci +dev/vmware/vmci/vmci_resource.coptional vmci +dev/vmware/pvscsi/pvscsi.c optional pvscsi dev/vr/if_vr.c optional vr pci dev/vt/colors/vt_termcolors.c optional vt dev/vt/font/vt_font_default.c optional vt Modified: head/sys/conf/files.amd64 == --- head/sys/conf/files.amd64 Thu Oct 15 14:17:45 2020(r366725) +++ head/sys/conf/files.amd64 Thu Oct 15 14:37:51 2020(r366726) @@ -378,18 +378,6 @@ dev/tpm/tpm_acpi.c optionaltpm acpi dev/tpm/tpm_isa.c optionaltpm isa dev/uart/uart_cpu_x86.coptionaluart dev/viawd/viawd.c optionalviawd -dev/vmware/vmxnet3/if_vmx.coptionalvmx -dev/vmware/vmci/vmci.c optionalvmci -dev/vmware/vmci/vmci_datagram.coptionalvmci -dev/vmware/vmci/vmci_doorbell.coptionalvmci -dev/vmware/vmci/vmci_driver.c optionalvmci -dev/vmware/vmci/vmci_event.c optionalvmci -dev/vmware/vmci/vmci_hashtable.c optionalvmci -dev/vmware/vmci/vmci_kernel_if.c optionalvmci -dev/vmware/vmci/vmci_qpair.c optionalvmci -dev/vmware/vmci/vmci_queue_pair.c optionalvmci -dev/vmware/vmci/vmci_resource.coptionalvmci -dev/vmware/pvscsi/pvscsi.c optionalpvscsi dev/vmd/vmd.c optionalvmd dev/vmd/vmd_bus.c optionalvmd_bus dev/wbwd/wbwd.coptionalwbwd Modified: head/sys/conf/files.i386 == --- head/sys/conf/files.i386Thu Oct 15 14:17:45 2020(r366725) +++ head/sys/conf/files.i386Thu Oct 15 14:37:51 2020(r366726) @@ -126,18 +126,6 @@ dev/tpm/tpm_acpi.c optional tpm acpi dev/tpm/tpm_isa.c optional tpm isa dev/uart/uart_cpu_x86.coptional uart dev/viawd/viawd.c optional viawd -dev/vmware/vmxnet3/if_vmx.coptional vmx -dev/vmware/vmci/vmci.c optionalvmci -dev/vmware/vmci/vmci_datagram.coptionalvmci -dev/vmware/vmci/vmci_doorbell.coptionalvmci -dev/vmware/vmci/vmci_driver.c optionalvmci -dev/vmware/vmci/vmci_event.c optionalvmci -dev/vmware/vmci/vmci_hashtable.c optionalvmci -dev/vmware/vmci/vmci_kernel_if.c optionalvmci -dev/vmware/vmci/vmci_qpair.c optionalvmci -dev/vmware/vmci/vmci_queue_pair.c optionalvmci -dev/vmware/vmci/vmci_resource.coptionalvmci -dev/vmware/pvscsi/pvscsi.c optionalpvscsi dev/acpi_support/acpi_wmi_if.m standard dev/wbwd/wbwd.coptional wbwd i386/acpica/acpi_machdep.c optional acpi ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366725 - head/lib/geom/eli
Author: 0mp (doc,ports committer) Date: Thu Oct 15 14:17:45 2020 New Revision: 366725 URL: https://svnweb.freebsd.org/changeset/base/366725 Log: Fix formatting of SYNOPSIS There was an unnecessary newline being added before Nm. MFC after:3 days Modified: head/lib/geom/eli/geli.8 Modified: head/lib/geom/eli/geli.8 == --- head/lib/geom/eli/geli.8Thu Oct 15 13:47:52 2020(r366724) +++ head/lib/geom/eli/geli.8Thu Oct 15 14:17:45 2020(r366725) @@ -24,7 +24,7 @@ .\" .\" $FreeBSD$ .\" -.Dd July 22, 2020 +.Dd October 15, 2020 .Dt GELI 8 .Os .Sh NAME @@ -45,8 +45,7 @@ to your geom_eli_load="YES" .Ed .Pp -Usage of the -.Nm +.No Usage of the Nm utility: .Pp .Nm ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366724 - head/sys/dev/iommu
Author: br
Date: Thu Oct 15 13:47:52 2020
New Revision: 366724
URL: https://svnweb.freebsd.org/changeset/base/366724
Log:
Split-out Guest Address Space (GAS) macroses to a separate header.
Sponsored by: Innovate DSbD
Added:
head/sys/dev/iommu/iommu_gas.h (contents, props changed)
Modified:
head/sys/dev/iommu/busdma_iommu.h
head/sys/dev/iommu/iommu.h
head/sys/dev/iommu/iommu_gas.c
Modified: head/sys/dev/iommu/busdma_iommu.h
==
--- head/sys/dev/iommu/busdma_iommu.h Thu Oct 15 13:43:43 2020
(r366723)
+++ head/sys/dev/iommu/busdma_iommu.h Thu Oct 15 13:47:52 2020
(r366724)
@@ -35,6 +35,7 @@
#define __X86_IOMMU_BUSDMA_DMAR_H
#include
+#include
struct bus_dma_tag_iommu {
struct bus_dma_tag_common common;
Modified: head/sys/dev/iommu/iommu.h
==
--- head/sys/dev/iommu/iommu.h Thu Oct 15 13:43:43 2020(r366723)
+++ head/sys/dev/iommu/iommu.h Thu Oct 15 13:47:52 2020(r366724)
@@ -65,18 +65,6 @@ struct iommu_map_entry {
struct iommu_qi_genseq gseq;
};
-#defineIOMMU_MAP_ENTRY_PLACE 0x0001 /* Fake entry */
-#defineIOMMU_MAP_ENTRY_RMRR0x0002 /* Permanent, not linked by
- dmamap_link */
-#defineIOMMU_MAP_ENTRY_MAP 0x0004 /* Busdma created, linked by
- dmamap_link */
-#defineIOMMU_MAP_ENTRY_UNMAPPED0x0010 /* No backing pages */
-#defineIOMMU_MAP_ENTRY_QI_NF 0x0020 /* qi task, do not free entry */
-#defineIOMMU_MAP_ENTRY_READ0x1000 /* Read permitted */
-#defineIOMMU_MAP_ENTRY_WRITE 0x2000 /* Write permitted */
-#defineIOMMU_MAP_ENTRY_SNOOP 0x4000 /* Snoop */
-#defineIOMMU_MAP_ENTRY_TM 0x8000 /* Transient */
-
struct iommu_unit {
struct mtx lock;
int unit;
@@ -148,17 +136,6 @@ struct iommu_ctx {
page table */
#defineIOMMU_DOMAIN_RMRR 0x0020 /* Domain contains RMRR
entry,
cannot be turned off */
-
-/* Map flags */
-#defineIOMMU_MF_CANWAIT0x0001
-#defineIOMMU_MF_CANSPLIT 0x0002
-#defineIOMMU_MF_RMRR 0x0004
-
-#defineIOMMU_PGF_WAITOK0x0001
-#defineIOMMU_PGF_ZERO 0x0002
-#defineIOMMU_PGF_ALLOC 0x0004
-#defineIOMMU_PGF_NOALLOC 0x0008
-#defineIOMMU_PGF_OBJL 0x0010
#defineIOMMU_LOCK(unit)mtx_lock(&(unit)->lock)
#defineIOMMU_UNLOCK(unit) mtx_unlock(&(unit)->lock)
Modified: head/sys/dev/iommu/iommu_gas.c
==
--- head/sys/dev/iommu/iommu_gas.c Thu Oct 15 13:43:43 2020
(r366723)
+++ head/sys/dev/iommu/iommu_gas.c Thu Oct 15 13:47:52 2020
(r366724)
@@ -62,6 +62,7 @@ __FBSDID("$FreeBSD$");
#include
#include
#include
+#include
#include
#include
#include
Added: head/sys/dev/iommu/iommu_gas.h
==
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ head/sys/dev/iommu/iommu_gas.h Thu Oct 15 13:47:52 2020
(r366724)
@@ -0,0 +1,60 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
+ *
+ * Copyright (c) 2013 The FreeBSD Foundation
+ * All rights reserved.
+ *
+ * This software was developed by Konstantin Belousov
+ * 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
svn commit: r366723 - head/sys/dev/iicbus
Author: nwhitehorn
Date: Thu Oct 15 13:43:43 2020
New Revision: 366723
URL: https://svnweb.freebsd.org/changeset/base/366723
Log:
Provide a slightly more-tolerant set of thermal parameters for PowerMac
motherboard temperatures. In particular, the U4 northbridge die is very
hard to cool or heat effectively with fans and is not responsive to load.
It generally sits around 64C, where it seems happy, so (like Linux) just
declare that to be its target temperature.
This makes the PowerMac G5 much less loud, with no change in the
temperatures of any system components.
MFC after:2 weeks
Modified:
head/sys/dev/iicbus/max6690.c
Modified: head/sys/dev/iicbus/max6690.c
==
--- head/sys/dev/iicbus/max6690.c Thu Oct 15 12:48:30 2020
(r366722)
+++ head/sys/dev/iicbus/max6690.c Thu Oct 15 13:43:43 2020
(r366723)
@@ -213,8 +213,17 @@ max6690_fill_sensor_prop(device_t dev)
for (j = 0; j < i; j++) {
sc->sc_sensors[j].dev = dev;
- sc->sc_sensors[j].therm.target_temp = 400 + ZERO_C_TO_K;
- sc->sc_sensors[j].therm.max_temp = 800 + ZERO_C_TO_K;
+ /*
+* Target value for "KODIAK DIODE" (= northbridge die) should
+* be 64C (value from Linux). It operates fine at that
+* temperature and has limited responsivity to the fan aimed at
+* it, so no point in trying to cool it to 40C.
+*/
+ if (strcmp(sc->sc_sensors[j].therm.name, "KODIAK DIODE") == 0)
+ sc->sc_sensors[j].therm.target_temp = 640 + ZERO_C_TO_K;
+ else
+ sc->sc_sensors[j].therm.target_temp = 400 + ZERO_C_TO_K;
+ sc->sc_sensors[j].therm.max_temp = 850 + ZERO_C_TO_K;
sc->sc_sensors[j].therm.read =
(int (*)(struct pmac_therm *))(max6690_sensor_read);
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366722 - head/sys/compat/linprocfs
Author: trasz Date: Thu Oct 15 12:48:30 2020 New Revision: 366722 URL: https://svnweb.freebsd.org/changeset/base/366722 Log: With some popular multiplayer games (such as Counter-Strike: Global Offensive) the Linux Steam client likes to occasionally scan the game process memory, presumably as part anti-cheat measures. Turns out the client also expects each inode entry to be followed by a space character, otherwise the parsing code crashes. PR: 248216 Submitted by: Alex S MFC after:2 weeks Modified: head/sys/compat/linprocfs/linprocfs.c Modified: head/sys/compat/linprocfs/linprocfs.c == --- head/sys/compat/linprocfs/linprocfs.c Thu Oct 15 11:44:28 2020 (r366721) +++ head/sys/compat/linprocfs/linprocfs.c Thu Oct 15 12:48:30 2020 (r366722) @@ -1249,7 +1249,7 @@ linprocfs_doprocmaps(PFS_FILL_ARGS) 0, 0, (u_long)ino, - *name ? " " : "", + *name ? " " : " ", name ); if (freename) ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r366721 - head/sys/dev/hyperv/netvsc
Author: whu
Date: Thu Oct 15 11:44:28 2020
New Revision: 366721
URL: https://svnweb.freebsd.org/changeset/base/366721
Log:
Hyper-V: hn: Relinquish cpu in HN_LOCK to avoid deadlock
The try lock loop in HN_LOCK put the thread spinning on cpu if the lock
is not available. It is possible to cause deadlock if the thread holding
the lock is sleeping. Relinquish the cpu to work around this problem even
it doesn't completely solve the issue. The priority inversion could cause
the livelock no matter how less likely it could happen. A more complete
solution may be needed in the future.
Reported by: Microsoft, Netapp
MFC after:2 weeks
Sponsored by: Microsoft
Modified:
head/sys/dev/hyperv/netvsc/if_hn.c
Modified: head/sys/dev/hyperv/netvsc/if_hn.c
==
--- head/sys/dev/hyperv/netvsc/if_hn.c Thu Oct 15 05:57:20 2020
(r366720)
+++ head/sys/dev/hyperv/netvsc/if_hn.c Thu Oct 15 11:44:28 2020
(r366721)
@@ -71,8 +71,10 @@ __FBSDID("$FreeBSD$");
#include
#include
#include
+#include
#include
#include
+#include
#include
#include
#include
@@ -165,8 +167,11 @@ __FBSDID("$FreeBSD$");
#define HN_LOCK_ASSERT(sc) sx_assert(&(sc)->hn_lock, SA_XLOCKED)
#define HN_LOCK(sc)\
do { \
- while (sx_try_xlock(&(sc)->hn_lock) == 0) \
+ while (sx_try_xlock(&(sc)->hn_lock) == 0) { \
+ /* Relinquish cpu to avoid deadlock */ \
+ sched_relinquish(curthread);\
DELAY(1000);\
+ } \
} while (0)
#define HN_UNLOCK(sc) sx_xunlock(&(sc)->hn_lock)
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
Re: svn commit: r366697 - head/usr.bin/xinstall
On Thu, 15 Oct 2020 at 06:59, Enji Cooper wrote: > > > > On Oct 14, 2020, at 5:28 AM, Alex Richardson > > wrote: > > > > Author: arichardson > > Date: Wed Oct 14 12:28:41 2020 > > New Revision: 366697 > > URL: https://svnweb.freebsd.org/changeset/base/366697 > > > > Log: > > install(1): Avoid unncessary fstatfs() calls and use mmap() based on size > > > > According to git blame the trymmap() function was added in 1996 to skip > > mmap() calls for NFS file systems. However, nowadays mmap() should be > > perfectly safe even on NFS. Importantly, onl ufs and cd9660 file systems > > were whitelisted so we don't use mmap() on ZFS. It also prevents the use > > of mmap() when bootstrapping from macOS/Linux since on those systems the > > trymmap() function was always returning zero due to the missing MFSNAMELEN > > define. > > > > This change keeps the trymmap() function but changes it to check whether > > using mmap() can reduce the number of system calls that are required. > > Using mmap() only reduces the number of system calls if we need multiple > > read() > > syscalls, i.e. if the file size is > MAXBSIZE. However, mmap() is more > > expensive > > than read() so this sets the threshold at 4 fewer syscalls. Additionally, > > for > > larger file size mmap() can significantly increase the number of page > > faults, > > so avoid it in that case. > > > > It's unclear whether using mmap() is ever faster than a read with an > > appropriate > > buffer size, but this change at least removes two unnecessary system calls > > for every file that is installed. > > > > Reviewed By: markj > > Differential Revision: https://reviews.freebsd.org/D26041 > > * Has this change been tested out with source filesystems other than > UFS/ZFS? Not all filesystems support mmap(2). I've used ext4 and afps, but it doesn't matter since there's a fallback. > * trymmap(..) seems to be less about computing whether or not the > filesystem should use mmap(2) after this change, but how it should use > mmap(2). Seems like a misnamed function call now. There was always fallback code in case mmap fails: https://github.com/freebsd/freebsd/blob/8349de39d23fc152c3782ee3b9eeab3df4610944/usr.bin/xinstall/xinstall.c#L1253 and https://github.com/freebsd/freebsd/blob/8349de39d23fc152c3782ee3b9eeab3df4610944/usr.bin/xinstall/xinstall.c#L1253 so it is really "should I try to use mmap()". Alex ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
