x86 CPUs can suffer severe performance drops if a tight loop, such as
the ones in __clear_user(), straddles a 16-byte instruction fetch
window, or worse, a 64-byte cacheline. This issues was discovered in the
SUSE kernel with the following commit,
1153933703d9 ("x86/asm/64: Micro-optimize __clear_user() - Use immediate constants")
which increased the code object size from 10 bytes to 15 bytes and
caused the 8-byte copy loop in __clear_user() to be split across a
64-byte cacheline.
Aligning the start of the loop to 16-bytes makes this fit neatly inside
a single instruction fetch window again and restores the performance of
__clear_user() which is used heavily when reading from /dev/zero.
Here are some numbers from running libmicro's read_z* and pread_z*
microbenchmarks which read from /dev/zero:
Zen 1 (Naples)
libmicro-file
5.7.0-rc6 5.7.0-rc6 5.7.0-rc6
revert-1153933703d9+ align16+
Time mean95-pread_z100k 9.9195 ( 0.00%) 5.9856 ( 39.66%) 5.9938 ( 39.58%)
Time mean95-pread_z10k 1.1378 ( 0.00%) 0.7450 ( 34.52%) 0.7467 ( 34.38%)
Time mean95-pread_z1k 0.2623 ( 0.00%) 0.2251 ( 14.18%) 0.2252 ( 14.15%)
Time mean95-pread_zw100k 9.9974 ( 0.00%) 6.0648 ( 39.34%) 6.0756 ( 39.23%)
Time mean95-read_z100k 9.8940 ( 0.00%) 5.9885 ( 39.47%) 5.9994 ( 39.36%)
Time mean95-read_z10k 1.1394 ( 0.00%) 0.7483 ( 34.33%) 0.7482 ( 34.33%)
Note that this doesn't affect Haswell or Broadwell microarchitectures
which seem to avoid the alignment issue by executing the loop straight
out of the Loop Stream Detector (verified using perf events).
Fixes: 1153933703d9 ("x86/asm/64: Micro-optimize __clear_user() - Use immediate constants")
Cc: "Grimm, Jon" <Jon.Grimm(a)amd.com>
Cc: "Kumar, Venkataramanan" <Venkataramanan.Kumar(a)amd.com>
CC: Jan Kara <jack(a)suse.cz>
Cc: <stable(a)vger.kernel.org> # v4.19+
Signed-off-by: Matt Fleming <matt(a)codeblueprint.co.uk>
---
arch/x86/lib/usercopy_64.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/arch/x86/lib/usercopy_64.c b/arch/x86/lib/usercopy_64.c
index fff28c6f73a2..b0dfac3d3df7 100644
--- a/arch/x86/lib/usercopy_64.c
+++ b/arch/x86/lib/usercopy_64.c
@@ -24,6 +24,7 @@ unsigned long __clear_user(void __user *addr, unsigned long size)
asm volatile(
" testq %[size8],%[size8]\n"
" jz 4f\n"
+ " .align 16\n"
"0: movq $0,(%[dst])\n"
" addq $8,%[dst]\n"
" decl %%ecx ; jnz 0b\n"
--
2.17.1
The patch below does not apply to the 4.9-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable(a)vger.kernel.org>.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
>From 80526491c2ca6abc028c0f0dbb0707a1f35fb18a Mon Sep 17 00:00:00 2001
From: Masami Hiramatsu <mhiramat(a)kernel.org>
Date: Thu, 23 Apr 2020 20:01:04 +0900
Subject: [PATCH] perf probe: Fix to check blacklist address correctly
Fix to check kprobe blacklist address correctly with relocated address
by adjusting debuginfo address.
Since the address in the debuginfo is same as objdump, it is different
from relocated kernel address with KASLR. Thus, 'perf probe' always
misses to catch the blacklisted addresses.
Without this patch, 'perf probe' can not detect the blacklist addresses
on a KASLR enabled kernel.
# perf probe kprobe_dispatcher
Failed to write event: Invalid argument
Error: Failed to add events.
#
With this patch, it correctly shows the error message.
# perf probe kprobe_dispatcher
kprobe_dispatcher is blacklisted function, skip it.
Probe point 'kprobe_dispatcher' not found.
Error: Failed to add events.
#
Fixes: 9aaf5a5f479b ("perf probe: Check kprobes blacklist when adding new events")
Signed-off-by: Masami Hiramatsu <mhiramat(a)kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme(a)redhat.com>
Cc: Jiri Olsa <jolsa(a)kernel.org>
Cc: Namhyung Kim <namhyung(a)kernel.org>
Cc: stable(a)vger.kernel.org
Link: http://lore.kernel.org/lkml/158763966411.30755.5882376357738273695.stgit@de…
Signed-off-by: Arnaldo Carvalho de Melo <acme(a)redhat.com>
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index c6bcf5709564..63d936f6e993 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -102,7 +102,7 @@ void exit_probe_symbol_maps(void)
symbol__exit();
}
-static struct ref_reloc_sym *kernel_get_ref_reloc_sym(void)
+static struct ref_reloc_sym *kernel_get_ref_reloc_sym(struct map **pmap)
{
/* kmap->ref_reloc_sym should be set if host_machine is initialized */
struct kmap *kmap;
@@ -114,6 +114,10 @@ static struct ref_reloc_sym *kernel_get_ref_reloc_sym(void)
kmap = map__kmap(map);
if (!kmap)
return NULL;
+
+ if (pmap)
+ *pmap = map;
+
return kmap->ref_reloc_sym;
}
@@ -125,7 +129,7 @@ static int kernel_get_symbol_address_by_name(const char *name, u64 *addr,
struct map *map;
/* ref_reloc_sym is just a label. Need a special fix*/
- reloc_sym = kernel_get_ref_reloc_sym();
+ reloc_sym = kernel_get_ref_reloc_sym(NULL);
if (reloc_sym && strcmp(name, reloc_sym->name) == 0)
*addr = (reloc) ? reloc_sym->addr : reloc_sym->unrelocated_addr;
else {
@@ -745,6 +749,7 @@ post_process_kernel_probe_trace_events(struct probe_trace_event *tevs,
int ntevs)
{
struct ref_reloc_sym *reloc_sym;
+ struct map *map;
char *tmp;
int i, skipped = 0;
@@ -753,7 +758,7 @@ post_process_kernel_probe_trace_events(struct probe_trace_event *tevs,
return post_process_offline_probe_trace_events(tevs, ntevs,
symbol_conf.vmlinux_name);
- reloc_sym = kernel_get_ref_reloc_sym();
+ reloc_sym = kernel_get_ref_reloc_sym(&map);
if (!reloc_sym) {
pr_warning("Relocated base symbol is not found!\n");
return -EINVAL;
@@ -764,9 +769,13 @@ post_process_kernel_probe_trace_events(struct probe_trace_event *tevs,
continue;
if (tevs[i].point.retprobe && !kretprobe_offset_is_supported())
continue;
- /* If we found a wrong one, mark it by NULL symbol */
+ /*
+ * If we found a wrong one, mark it by NULL symbol.
+ * Since addresses in debuginfo is same as objdump, we need
+ * to convert it to addresses on memory.
+ */
if (kprobe_warn_out_range(tevs[i].point.symbol,
- tevs[i].point.address)) {
+ map__objdump_2mem(map, tevs[i].point.address))) {
tmp = NULL;
skipped++;
} else {
@@ -2935,7 +2944,7 @@ static int find_probe_trace_events_from_map(struct perf_probe_event *pev,
/* Note that the symbols in the kmodule are not relocated */
if (!pev->uprobes && !pev->target &&
(!pp->retprobe || kretprobe_offset_is_supported())) {
- reloc_sym = kernel_get_ref_reloc_sym();
+ reloc_sym = kernel_get_ref_reloc_sym(NULL);
if (!reloc_sym) {
pr_warning("Relocated base symbol is not found!\n");
ret = -EINVAL;
The patch below does not apply to the 4.4-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable(a)vger.kernel.org>.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
>From 80526491c2ca6abc028c0f0dbb0707a1f35fb18a Mon Sep 17 00:00:00 2001
From: Masami Hiramatsu <mhiramat(a)kernel.org>
Date: Thu, 23 Apr 2020 20:01:04 +0900
Subject: [PATCH] perf probe: Fix to check blacklist address correctly
Fix to check kprobe blacklist address correctly with relocated address
by adjusting debuginfo address.
Since the address in the debuginfo is same as objdump, it is different
from relocated kernel address with KASLR. Thus, 'perf probe' always
misses to catch the blacklisted addresses.
Without this patch, 'perf probe' can not detect the blacklist addresses
on a KASLR enabled kernel.
# perf probe kprobe_dispatcher
Failed to write event: Invalid argument
Error: Failed to add events.
#
With this patch, it correctly shows the error message.
# perf probe kprobe_dispatcher
kprobe_dispatcher is blacklisted function, skip it.
Probe point 'kprobe_dispatcher' not found.
Error: Failed to add events.
#
Fixes: 9aaf5a5f479b ("perf probe: Check kprobes blacklist when adding new events")
Signed-off-by: Masami Hiramatsu <mhiramat(a)kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme(a)redhat.com>
Cc: Jiri Olsa <jolsa(a)kernel.org>
Cc: Namhyung Kim <namhyung(a)kernel.org>
Cc: stable(a)vger.kernel.org
Link: http://lore.kernel.org/lkml/158763966411.30755.5882376357738273695.stgit@de…
Signed-off-by: Arnaldo Carvalho de Melo <acme(a)redhat.com>
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index c6bcf5709564..63d936f6e993 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -102,7 +102,7 @@ void exit_probe_symbol_maps(void)
symbol__exit();
}
-static struct ref_reloc_sym *kernel_get_ref_reloc_sym(void)
+static struct ref_reloc_sym *kernel_get_ref_reloc_sym(struct map **pmap)
{
/* kmap->ref_reloc_sym should be set if host_machine is initialized */
struct kmap *kmap;
@@ -114,6 +114,10 @@ static struct ref_reloc_sym *kernel_get_ref_reloc_sym(void)
kmap = map__kmap(map);
if (!kmap)
return NULL;
+
+ if (pmap)
+ *pmap = map;
+
return kmap->ref_reloc_sym;
}
@@ -125,7 +129,7 @@ static int kernel_get_symbol_address_by_name(const char *name, u64 *addr,
struct map *map;
/* ref_reloc_sym is just a label. Need a special fix*/
- reloc_sym = kernel_get_ref_reloc_sym();
+ reloc_sym = kernel_get_ref_reloc_sym(NULL);
if (reloc_sym && strcmp(name, reloc_sym->name) == 0)
*addr = (reloc) ? reloc_sym->addr : reloc_sym->unrelocated_addr;
else {
@@ -745,6 +749,7 @@ post_process_kernel_probe_trace_events(struct probe_trace_event *tevs,
int ntevs)
{
struct ref_reloc_sym *reloc_sym;
+ struct map *map;
char *tmp;
int i, skipped = 0;
@@ -753,7 +758,7 @@ post_process_kernel_probe_trace_events(struct probe_trace_event *tevs,
return post_process_offline_probe_trace_events(tevs, ntevs,
symbol_conf.vmlinux_name);
- reloc_sym = kernel_get_ref_reloc_sym();
+ reloc_sym = kernel_get_ref_reloc_sym(&map);
if (!reloc_sym) {
pr_warning("Relocated base symbol is not found!\n");
return -EINVAL;
@@ -764,9 +769,13 @@ post_process_kernel_probe_trace_events(struct probe_trace_event *tevs,
continue;
if (tevs[i].point.retprobe && !kretprobe_offset_is_supported())
continue;
- /* If we found a wrong one, mark it by NULL symbol */
+ /*
+ * If we found a wrong one, mark it by NULL symbol.
+ * Since addresses in debuginfo is same as objdump, we need
+ * to convert it to addresses on memory.
+ */
if (kprobe_warn_out_range(tevs[i].point.symbol,
- tevs[i].point.address)) {
+ map__objdump_2mem(map, tevs[i].point.address))) {
tmp = NULL;
skipped++;
} else {
@@ -2935,7 +2944,7 @@ static int find_probe_trace_events_from_map(struct perf_probe_event *pev,
/* Note that the symbols in the kmodule are not relocated */
if (!pev->uprobes && !pev->target &&
(!pp->retprobe || kretprobe_offset_is_supported())) {
- reloc_sym = kernel_get_ref_reloc_sym();
+ reloc_sym = kernel_get_ref_reloc_sym(NULL);
if (!reloc_sym) {
pr_warning("Relocated base symbol is not found!\n");
ret = -EINVAL;
The patch below does not apply to the 5.4-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable(a)vger.kernel.org>.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
>From fc3bb095ab02b9e7d89a069ade2cead15c64c504 Mon Sep 17 00:00:00 2001
From: Eric Biggers <ebiggers(a)google.com>
Date: Mon, 1 Jun 2020 13:08:05 -0700
Subject: [PATCH] f2fs: avoid utf8_strncasecmp() with unstable name
If the dentry name passed to ->d_compare() fits in dentry::d_iname, then
it may be concurrently modified by a rename. This can cause undefined
behavior (possibly out-of-bounds memory accesses or crashes) in
utf8_strncasecmp(), since fs/unicode/ isn't written to handle strings
that may be concurrently modified.
Fix this by first copying the filename to a stack buffer if needed.
This way we get a stable snapshot of the filename.
Fixes: 2c2eb7a300cd ("f2fs: Support case-insensitive file name lookups")
Cc: <stable(a)vger.kernel.org> # v5.4+
Cc: Al Viro <viro(a)zeniv.linux.org.uk>
Cc: Daniel Rosenberg <drosen(a)google.com>
Cc: Gabriel Krisman Bertazi <krisman(a)collabora.co.uk>
Signed-off-by: Eric Biggers <ebiggers(a)google.com>
Reviewed-by: Chao Yu <yuchao0(a)huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk(a)kernel.org>
diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c
index 29f70f2295cc..d35976785e8c 100644
--- a/fs/f2fs/dir.c
+++ b/fs/f2fs/dir.c
@@ -1114,11 +1114,27 @@ static int f2fs_d_compare(const struct dentry *dentry, unsigned int len,
const struct inode *dir = READ_ONCE(parent->d_inode);
const struct f2fs_sb_info *sbi = F2FS_SB(dentry->d_sb);
struct qstr entry = QSTR_INIT(str, len);
+ char strbuf[DNAME_INLINE_LEN];
int res;
if (!dir || !IS_CASEFOLDED(dir))
goto fallback;
+ /*
+ * If the dentry name is stored in-line, then it may be concurrently
+ * modified by a rename. If this happens, the VFS will eventually retry
+ * the lookup, so it doesn't matter what ->d_compare() returns.
+ * However, it's unsafe to call utf8_strncasecmp() with an unstable
+ * string. Therefore, we have to copy the name into a temporary buffer.
+ */
+ if (len <= DNAME_INLINE_LEN - 1) {
+ memcpy(strbuf, str, len);
+ strbuf[len] = 0;
+ entry.name = strbuf;
+ /* prevent compiler from optimizing out the temporary buffer */
+ barrier();
+ }
+
res = utf8_strncasecmp(sbi->s_encoding, name, &entry);
if (res >= 0)
return res;
Hi CKI maintainers,
thanks for doing automated tests.
I am interested in a report of currently released Linux v5.7.5-rc1
before doing my testing with Clang's Integrated Assembly on
Debian/testing AMD64.
Is there a browsable URL you can give me where I can see if AMD64
(x86-64) tests have passed OK?
Or is it "Be patient and wait".
Thanks.
Regards,
- Sedat -
[1] https://git.kernel.org/pub/scm/public-inbox/vger.kernel.org/stable/0.git/co…