Suppose:
- the source extent is: [0, 100]
- the src offset is 10
- the clone length is 90
- the dest offset is 0
This statement:
new_key.offset = key.offset + destoff - off
will produce such an extent for the dest file:
[ino, BTRFS_EXTENT_DATA_KEY, -10]
, which is obviously wrong.
Signed-off-by: Li Zefan <[email protected]>
After returing extents from a cluster to the block group, some
extents in the block group may be mergeable.
Reviewed-by: Josef Bacik <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
When adding a new extent, we'll firstly see if we can merge
this extent to the left or/and right extent. Extract this as
a helper try_merge_free_space().
As a side effect, we fix a small bug that if the new extent
has non-bitmap left entry but is unmergeble, we'll directly
link the extent without trying to drop it into bitmap.
This also prepares for the next patch.
Reviewed-by: Josef Bacik <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
When allocating extent entry from a cluster, we should update
the free_space and free_extents fields of the block group.
Reviewed-by: Josef Bacik <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
Remove some duplicated code.
This prepares for the next patch.
Reviewed-by: Josef Bacik <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
If a block group is smaller than 1GB, the extent entry threadhold
calculation will always set the threshold to 0.
So as free space gets fragmented, btrfs will switch to use bitmap
to manage free space, but then will never switch back to extents
due to this bug.
Reviewed-by: Josef Bacik <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
While running my ftrace stress test, this showed up:
BUG: sleeping function called from invalid context at mm/mmap.c:233
...
note: cat[3293] exited with preempt_count 1
The bug was introduced by commit 91e86e560d
("tracing: Fix recursive user stack trace")
Cc: <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
This allows us to set a snapshot or a subvolume readonly or writable
on the fly.
Usage:
Set BTRFS_SUBVOL_RDONLY of btrfs_ioctl_vol_arg_v2->flags, and then
call ioctl(BTRFS_IOCTL_SUBVOL_SETFLAGS);
Changelog for v3:
- Change to pass __u64 as ioctl parameter.
Changelog for v2:
- Add _GETFLAGS ioctl.
- Check if the passed fd is the root of a subvolume.
- Change the name from _SNAP_SETFLAGS to _SUBVOL_SETFLAGS.
Signed-off-by: Li Zefan <[email protected]>
Usage:
Set BTRFS_SUBVOL_RDONLY of btrfs_ioctl_vol_arg_v2->flags, and call
ioctl(BTRFS_I0CTL_SNAP_CREATE_V2).
Implementation:
- Set readonly bit of btrfs_root_item->flags.
- Add readonly checks in btrfs_permission (inode_permission),
btrfs_setattr, btrfs_set/remove_xattr and some ioctls.
Changelog for v3:
- Eliminate btrfs_root->readonly, but check btrfs_root->root_item.flags.
- Rename BTRFS_ROOT_SNAP_RDONLY to BTRFS_ROOT_SUBVOL_RDONLY.
Signed-off-by: Li Zefan <[email protected]>
Update defrag ioctl, so one can choose lzo or zlib when turning
on compression in defrag operation.
Changelog:
v1 -> v2
- Add incompability flag.
- Fix to check invalid compress type.
Signed-off-by: Li Zefan <[email protected]>
Lzo is a much faster compression algorithm than gzib, so would allow
more users to enable transparent compression, and some users can
choose from compression ratio and speed for different applications
Usage:
# mount -t btrfs -o compress[=<zlib,lzo>] dev /mnt
or
# mount -t btrfs -o compress-force[=<zlib,lzo>] dev /mnt
"-o compress" without argument is still allowed for compatability.
Compatibility:
If we mount a filesystem with lzo compression, it will not be able be
mounted in old kernels. One reason is, otherwise btrfs will directly
dump compressed data, which sits in inline extent, to user.
Performance:
The test copied a linux source tarball (~400M) from an ext4 partition
to the btrfs partition, and then extracted it.
(time in second)
lzo zlib nocompress
copy: 10.6 21.7 14.9
extract: 70.1 94.4 66.6
(data size in MB)
lzo zlib nocompress
copy: 185.87 108.69 394.49
extract: 193.80 132.36 381.21
Changelog:
v1 -> v2:
- Select LZO_COMPRESS and LZO_DECOMPRESS in btrfs Kconfig.
- Add incompability flag.
- Fix error handling in compress code.
Signed-off-by: Li Zefan <[email protected]>
Make the code aware of compression type, instead of always assuming
zlib compression.
Also make the zlib workspace function as common code for all
compression types.
Signed-off-by: Li Zefan <[email protected]>
If we had reserved some bytes in struct btrfs_ioctl_vol_args, we
wouldn't have to create a new structure for async snapshot creation.
Here we convert async snapshot ioctl to use a more generic ABI, as
we'll add more ioctls for snapshots/subvolumes in the future, readonly
snapshots for example.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Chris Mason <[email protected]>
We should drop dentry before deactivating the superblock, otherwise
we can hit this bug:
BUG: Dentry f349a690{i=100,n=/} still in use (1) [unmount of btrfs loop1]
...
Steps to reproduce the bug:
# mount /dev/loop1 /mnt
# mkdir save
# btrfs subvolume snapshot /mnt save/snap1
# umount /mnt
# mount -o subvol=save/snap1 /dev/loop1 /mnt
(crash)
Reported-by: Michael Niederle <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Chris Mason <[email protected]>
Set src_offset = 0, src_length = 20K, dest_offset = 20K. And the
original filesize of the dest file 'file2' is 30K:
# ls -l /mnt/file2
-rw-r--r-- 1 root root 30720 Nov 18 16:42 /mnt/file2
Now clone file1 to file2, the dest file should be 40K, but it
still shows 30K:
# ls -l /mnt/file2
-rw-r--r-- 1 root root 30720 Nov 18 16:42 /mnt/file2
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Chris Mason <[email protected]>
We've done the check for src_offset and src_length, and We should
also check dest_offset, otherwise we'll corrupt the destination
file:
(After cloning file1 to file2 with unaligned dest_offset)
# cat /mnt/file2
cat: /mnt/file2: Input/output error
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Chris Mason <[email protected]>
Dave reported an rcu lockdep warning on 2.6.35.4 kernel
task->cgroups and task->cgroups->subsys[i] are protected by RCU.
So we avoid accessing invalid pointers here. This might happen,
for example, when you are deref-ing those pointers while someone
move @task from one cgroup to another.
Reported-by: Dave Jones <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
While we are reading trace_stat/functionX and someone just
disabled function_profile at that time, we can trigger this:
divide error: 0000 [#1] PREEMPT SMP
...
EIP is at function_stat_show+0x90/0x230
...
This fix just takes the ftrace_profile_lock and checks if
rec->counter is 0. If it's 0, we know the profile buffer
has been reset.
Signed-off-by: Li Zefan <[email protected]>
Cc: [email protected]
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
Remove the nasty hack that marks a pointer's LSB to distinguish common
fields from event fields. Replace it with a more sane approach.
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
We found that even enabling a single trace event that will rarely be
triggered can add big overhead to context switch.
(lmbench context switch test)
-------------------------------------------------
2p/0K 2p/16K 2p/64K 8p/16K 8p/64K 16p/16K 16p/64K
ctxsw ctxsw ctxsw ctxsw ctxsw ctxsw ctxsw
------ ------ ------ ------ ------ ------- -------
2.19 2.3 2.21 2.56 2.13 2.54 2.07
2.39 2.51 2.35 2.75 2.27 2.81 2.24
The overhead is 6% ~ 11%.
It's because when a trace event is enabled 3 tracepoints (sched_switch,
sched_wakeup, sched_wakeup_new) will be activated to map pid to cmdname.
We'd like to avoid this overhead, so add a trace option '(no)record-cmd'
to allow to disable cmdline recording.
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
Let trace_module_add_events() and event_trace_init() call
__trace_add_event_call().
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
Every event (or event class) has it's define_fields callback,
so the test is redundant.
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
Convert sched_wait_task to DEFINE_EVENT, and save ~1K:
text data bss dec hex filename
104595 9424 4992 119011 1d0e3 kernel/sched.o.orig
103619 9344 4992 117955 1ccc3 kernel/sched.o
No change in functionality.
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
Use DECLARE_EVENT_CLASS, and save ~2.3K:
text data bss dec hex filename
7018823 2031888 7251132 16301843 f8bf13 vmlinux.o.orig
7016727 2031696 7251132 16299555 f8b623 vmlinux.o
5 events are converted:
timer_class: timer_init, timer_expire_exit, timer_cancel
hrtimer_class: hrtimer_init, hrtimer_cancel
No change in functionality.
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
Every event has the same common fields, so it's a big waste of
memory to have a copy of those fields for every event.
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
All syscall exit events have the same fields.
The kernel size drops 2.5K:
text data bss dec hex filename
7018612 2034376 7251132 16304120 f8c7f8 vmlinux.o.orig
7018612 2031888 7251132 16301632 f8be40 vmlinux.o
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
We have been resisting new ftrace plugins and removing existing
ones, and kmemtrace has been superseded by kmem trace events
and perf-kmem, so we remove it.
Signed-off-by: Li Zefan <[email protected]>
Acked-by: Pekka Enberg <[email protected]>
Acked-by: Eduard - Gabriel Munteanu <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Steven Rostedt <[email protected]>
[ remove kmemtrace from the makefile, handle slob too ]
Signed-off-by: Frederic Weisbecker <[email protected]>
Truncate list items to fit in a single line, otherwise those items
which have long prompts will cover some other items.
This follows the behavior of menubox.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
Run:
make ARCH=arm menuconfig
And then select "System Type" -> "ARM system type". The kconfig
"choice" menu at this point looks empty.
It's because config ARCH_S3C2410 has a long prompt:
config ARCH_S3C2410
bool "Samsung S3C2410, S3C2412, S3C2413, S3C2416, S3C2440, S3C2442, S3C2443, S3C2450"
...
menuconfig centers the checklist according to this prompt without
considering the width of the list, and then things get wrong.
Reported-by: Nobin Mathew <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
This feature has been supported in menuconfig and gconfig, so
here add it to xconfig.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
@ok is a pointer to a bool var, so we should check the value of
*ok. But actually we don't need to check it, so just remove the
if statement.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
In gconfig if you enable "Show all options", you'll see some "(null)"
config options, and clicking those options triggers a warning:
(gconf:9368): Gtk-CRITICAL **: gtk_text_buffer_insert_with_tags: assertion `text != NULL' failed
Signed-off-by: Li Zefan <[email protected]>
Acked-by: Randy Dunlap <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
Those configs are not new:
$ cat .config
...
CONFIG_NAMESPACES=y
...
CONFIG_BLOCK=y
...
But are tagged as NEW:
$ yes "" | make config > myconf
$ cat myconf | grep '(NEW)'
Namespaces support (NAMESPACES) [Y/?] (NEW) y
...
Enable the block layer (BLOCK) [Y/?] (NEW) y
...
You can also notice this bug when using gconfig/xconfig.
It's because the SYMBOL_DEF_USER bit of an invisible symbol is cleared
when the config file is read:
int conf_read(const char *name)
{
...
for_all_symbols(i, sym) {
if (sym_has_value(sym) && !sym_is_choice_value(sym)) {
/* Reset values of generates values, so they'll appear
* as new, if they should become visible, but that
* doesn't quite work if the Kconfig and the saved
* configuration disagree.
*/
if (sym->visible == no && !conf_unsaved)
sym->flags &= ~SYMBOL_DEF_USER;
...
}
But a menu item which represents an invisible symbol is still
visible, if it's sub-menu is visible, so its SYMBOL_DEF_USER
bit should be set to indicate it's not NEW.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
Without this patch, one has to refer to the Kconfig file to find
out the range of an integer/hex symbol.
│ Symbol: NR_CPUS [=4]
│ Type : integer
│ Range : [2 8]
│ Prompt: Maximum number of CPUs
│ Defined at arch/x86/Kconfig:761
│ Depends on: SMP [=y] && !MAXSMP [=n]
│ Location:
│ -> Processor type and features
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
Randy suggested to print out the symbol type in gconfig.
Note this change does more than Randy's suggestion, that it also
affects menuconfig and "make config".
│ Symbol: BLOCK [=y]
│ Type : boolean
│ Prompt: Enable the block layer
│ Defined at block/Kconfig:4
│ Depends on: EMBEDDED [=n]
Signed-off-by: Li Zefan <[email protected]>
Acked-by: Randy Dunlap <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
Use DECLARE_EVENT_CLASS, and save ~15K:
text data bss dec hex filename
171949 43028 48 215025 347f1 fs/xfs/linux-2.6/xfs_trace.o.orig
156521 43028 36 199585 30ba1 fs/xfs/linux-2.6/xfs_trace.o
No change in functionality.
Signed-off-by: Li Zefan <[email protected]>
Acked-by: Steven Rostedt <[email protected]>
Signed-off-by: Alex Elder <[email protected]>
Convert i915_gem_object_clflush to DEFINE_EVENT, and save ~0.5K:
text data bss dec hex filename
13204 2732 12 15948 3e4c i915_trace_points.o.orig
12668 2732 12 15412 3c34 i915_trace_points.o
No change in functionality.
Signed-off-by: Li Zefan <[email protected]>
Acked-by: Steven Rostedt <[email protected]>
Signed-off-by: Eric Anholt <[email protected]>
Use DECLARE_EVENT_CLASS, and save ~2.7K:
text data bss dec hex filename
274441 7200 260 281901 44d2d fs/ext4/ext4.o.orig
271881 7040 256 279177 44289 fs/ext4/ext4.o
4 events are converted:
ext4__mb_new_pa: ext4_mb_new_inode_pa, ext4_mb_new_group_pa
ext4__mballoc: ext4_mballoc_discard, ext4_mballoc_free
No change in functionality.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>
ftrace_trace_stack() and frace_trace_userstacke() take a
struct ring_buffer argument, not struct trace_array. Commit
e77405ad("tracing: pass around ring buffer instead of tracer")
made this change.
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
with CONFIG_PROVE_RCU=y, a warning can be triggered:
# mount -t cgroup -o blkio xxx /mnt
# mkdir /mnt/subgroup
...
kernel/cgroup.c:4442 invoked rcu_dereference_check() without protection!
...
To fix this, we avoid caling css_depth() here, which is a bit simpler
than the original code.
Signed-off-by: Li Zefan <[email protected]>
Acked-by: Vivek Goyal <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
Expand task_subsys_state()'s rcu_dereference_check() to include the full
locking rule as documented in Documentation/cgroups/cgroups.txt by adding
a check for task->alloc_lock being held.
This fixes an RCU false positive when resuming from suspend. The warning
comes from freezer cgroup in cgroup_freezing_or_frozen().
Signed-off-by: Li Zefan <[email protected]>
Acked-by: Matt Helsley <[email protected]>
Signed-off-by: Paul E. McKenney <[email protected]>
With CONFIG_PROVE_RCU=y, a warning can be triggered:
$ cat /proc/sched_debug
...
kernel/cgroup.c:1649 invoked rcu_dereference_check() without protection!
...
Both cgroup_path() and task_group() should be called with either
rcu_read_lock or cgroup_mutex held.
The rcu_dereference_check() does include cgroup_lock_is_held(), so we
know that this lock is not held. Therefore, in a CONFIG_PREEMPT kernel,
to say nothing of a CONFIG_PREEMPT_RT kernel, the original code could
have ended up copying a string out of the freelist.
This patch inserts RCU read-side primitives needed to prevent this
scenario.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Paul E. McKenney <[email protected]>
With CONFIG_PROVE_RCU=y, a warning can be triggered:
# mount -t cgroup -o memory xxx /mnt
# mkdir /mnt/0
...
kernel/cgroup.c:4442 invoked rcu_dereference_check() without protection!
...
This is a false-positive. It's safe to directly access parent_css->id.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Paul E. McKenney <[email protected]>
with CONFIG_PROVE_RCU=y, a warning can be triggered:
# mount -t cgroup -o debug xxx /mnt
# cat /proc/$$/cgroup
...
kernel/cgroup.c:1649 invoked rcu_dereference_check() without protection!
...
This is a false-positive, because cgroup_path() can be called
with either rcu_read_lock() held or cgroup_mutex held.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Paul E. McKenney <[email protected]>
There's a button in gconfig to "Show all options", but I think
normally we are not interested in those configs which have no
prompt and thus can't be changed, so here I add a new button to
show hidden options which have prompts.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
Usage:
Press <Z> to show all config symbols which have prompts.
Quote Tim Bird:
| I've been bitten by this numerous times. I most often
| use ftrace on ARM, but when I go back to x86, I almost
| always go through a sequence of searching for the
| function graph tracer in the menus, then realizing it's
| completely missing until I disable CC_OPTIMIZE_FOR_SIZE.
|
| Is there any way to have the menu item appear, but be
| unsettable unless the SIZE option is disabled? I'm
| not a Kconfig guru...
I myself found this useful too. For example, I need to test
ftrace/tracing and want to be sure all the tracing features are
enabled, so I enter the "Tracers" menu, and press <Z> to
see if there is any config hidden.
I also noticed gconfig and xconfig have a button "Show all options",
but that's a bit too much, and I think normally what we are not
interested in those configs which have no prompt thus can't be
changed by users.
Exmaple:
--- Tracers
-*- Kernel Function Tracer
- - Kernel Function Graph Tracer
[*] Interrupts-off Latency Tracer
- - Preemption-off Latency Tracer
[*] Sysprof Tracer
Here you can see 2 tracers are not selectable, and then can find
out how to make them selectable.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
zconfdump(), which is used for debugging, can't recognize P_SELECT,
P_RANGE and P_MENU (if associated with a symbol, aka "menuconfig"),
and output something like this:
config X86
boolean
default y
unknown prop 6!
unknown prop 6!
unknown prop 6!
...
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
- fix a typo in documentation
- fix a typo in a printk on error
- fix comments in dialog_inputbox()
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
Remove the @refcnt argument, because it has side-effects, and arguments with
side-effects are not skipped by the jump over disabled instrumentation and are
executed even when the tracepoint is disabled.
This was also causing a GPF as found by Randy Dunlap:
Subject: 2.6.33 GP fault only when built with tracing
LKML-Reference: <[email protected]>
Note, the current 2.6.34-rc has a fix for the actual cause of the GPF,
but this fixes one of its triggers.
Tested-by: Randy Dunlap <[email protected]>
Acked-by: Mathieu Desnoyers <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
# echo 1 > events/enable
# echo global > trace_clock
------------[ cut here ]------------
WARNING: at kernel/lockdep.c:3162 check_flags+0xb2/0x190()
...
---[ end trace 3f86734a89416623 ]---
possible reason: unannotated irqs-on.
...
There's no reason to use the raw_local_irq_save() in trace_clock_global.
The local_irq_save() version is fine, and does not cause the bug in lockdep.
Acked-by: Peter Zijlstra <[email protected]>
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
A symbol's value won't be recalc-ed until we save config file or
enter the menu where the symbol sits.
So If I enable OPTIMIZE_FOR_SIZE, and search FUNCTION_GRAPH_TRACER:
Symbol: FUNCTION_GRAPH_TRACER [=y]
Prompt: Kernel Function Graph Tracer
Defined at kernel/trace/Kconfig:140
Depends on: ... [=y] && (!X86_32 [=y] || !CC_OPTIMIZE_FOR_SIZE [=y])
...
From the dependency it should result in FUNCTION_GRAPH_TRACER=n,
but it still shows FUNCTION_GRAPH_TRACER=y.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
Make the config visible, so we can choose from CONFIG_BLK_CGROUP=y
and CONFIG_BLK_CGROUP=m when CONFIG_IOSCHED_CFQ=m.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
I don't see why we can only clear all functions from the filter.
After patching:
# echo sys_open > set_graph_function
# echo sys_close >> set_graph_function
# cat set_graph_function
sys_open
sys_close
# echo '!sys_close' >> set_graph_function
# cat set_graph_function
sys_open
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
Some places in kernel need to iterate over a hlist in seq_file,
so provide some common helpers.
Signed-off-by: Li Zefan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
In cgroup_create(), if alloc_css_id() returns failure, the errno is not
propagated to userspace, so mkdir will fail silently.
To trigger this bug, we mount blkio (or memory subsystem), and create more
then 65534 cgroups. (The number of cgroups is limited to 65535 if a
subsystem has use_id == 1)
# mount -t cgroup -o blkio xxx /mnt
# for ((i = 0; i < 65534; i++)); do mkdir /mnt/$i; done
# mkdir /mnt/65534
(should return ENOSPC)
#
Signed-off-by: Li Zefan <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Acked-by: Paul Menage <[email protected]>
Acked-by: KAMEZAWA Hiroyuki <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
We should be clear on 2 things:
- the length parameter of a match callback includes
tailing '\0'.
- the string to be searched might not be NULL-terminated.
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
MATCH_FULL matching for PTR_STRING is not working correctly:
# echo 'func == vt' > events/bkl/lock_kernel/filter
# echo 1 > events/bkl/lock_kernel/enable
...
# cat trace
Xorg-1484 [000] 1973.392586: lock_kernel: ... func=vt_ioctl()
gpm-1402 [001] 1974.027740: lock_kernel: ... func=vt_ioctl()
We should pass to regex.match(..., len) the length (including '\0')
of the source string instead of the length of the pattern string.
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Acked-by: Frederic Weisbecker <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
The @str might not be NULL-terminated if it's of type
DYN_STRING or STATIC_STRING, so we should use strnstr()
instead of strstr().
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Acked-by: Frederic Weisbecker <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
For '*foo' pattern, we should allow any string ending with
'foo', but event filtering incorrectly disallows strings
like bar_foo_foo:
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Acked-by: Frederic Weisbecker <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
MATCH_FRONT_ONLY actually is a full matching:
# ./perf record -R -f -a -e lock:lock_acquire \
--filter 'name ~rcu_*' sleep 1
# ./perf trace
(no output)
We should pass the length of the pattern string to strncmp().
Signed-off-by: Li Zefan <[email protected]>
LKML-Reference: <[email protected]>
Acked-by: Frederic Weisbecker <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
Li Zefan