| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Dave Hansen:
"The most important fix here adds a missing CPU model to the recent
Gather Data Sampling (GDS) mitigation list to ensure that mitigations
are available on that CPU.
There are also a pair of warning fixes, and closure of a covert
channel that pops up when protection keys are disabled.
Summary:
- Mark all Skylake CPUs as vulnerable to GDS
- Fix PKRU covert channel
- Fix -Wmissing-variable-declarations warning for ia32_xyz_class
- Fix kernel-doc annotation warning"
* tag 'x86-urgent-2023-09-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/fpu/xstate: Fix PKRU covert channel
x86/irq/i8259: Fix kernel-doc annotation warning
x86/speculation: Mark all Skylake CPUs as vulnerable to GDS
x86/audit: Fix -Wmissing-variable-declarations warning for ia32_xyz_class
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When XCR0[9] is set, PKRU can be read and written from userspace with
XSAVE and XRSTOR, even when CR4.PKE is clear.
Clear XCR0[9] when protection keys are disabled.
Reported-by: Tavis Ormandy <taviso@google.com>
Signed-off-by: Jim Mattson <jmattson@google.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lore.kernel.org/r/20230831043228.1194256-1-jmattson@google.com
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Fix this warning:
arch/x86/kernel/i8259.c:235: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
* ELCR registers (0x4d0, 0x4d1) control edge/level of IRQ
CC arch/x86/kernel/irqinit.o
Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230830131211.88226-1-vincenzopalazzodev@gmail.com
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The Gather Data Sampling (GDS) vulnerability is common to all Skylake
processors. However, the "client" Skylakes* are now in this list:
https://www.intel.com/content/www/us/en/support/articles/000022396/processors.html
which means they are no longer included for new vulnerabilities here:
https://www.intel.com/content/www/us/en/developer/topic-technology/software-security-guidance/processors-affected-consolidated-product-cpu-model.html
or in other GDS documentation. Thus, they were not included in the
original GDS mitigation patches.
Mark SKYLAKE and SKYLAKE_L as vulnerable to GDS to match all the
other Skylake CPUs (which include Kaby Lake). Also group the CPUs
so that the ones that share the exact same vulnerabilities are next
to each other.
Last, move SRBDS to the end of each line. This makes it clear at a
glance that SKYLAKE_X is unique. Of the five Skylakes, it is the
only "server" CPU and has a different implementation from the
clients of the "special register" hardware, making it immune to SRBDS.
This makes the diff much harder to read, but the resulting table is
worth it.
I very much appreciate the report from Michael Zhivich about this
issue. Despite what level of support a hardware vendor is providing,
the kernel very much needs an accurate and up-to-date list of
vulnerable CPUs. More reports like this are very welcome.
* Client Skylakes are CPUID 406E3/506E3 which is family 6, models
0x4E and 0x5E, aka INTEL_FAM6_SKYLAKE and INTEL_FAM6_SKYLAKE_L.
Reported-by: Michael Zhivich <mzhivich@akamai.com>
Fixes: 8974eb588283 ("x86/speculation: Add Gather Data Sampling mitigation")
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Daniel Sneddon <daniel.sneddon@linux.intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When building x86 defconfig with Clang-18 I get the following warnings:
| arch/x86/ia32/audit.c:6:10: warning: no previous extern declaration for non-static variable 'ia32_dir_class' [-Wmissing-variable-declarations]
| 6 | unsigned ia32_dir_class[] = {
| arch/x86/ia32/audit.c:11:10: warning: no previous extern declaration for non-static variable 'ia32_chattr_class' [-Wmissing-variable-declarations]
| 11 | unsigned ia32_chattr_class[] = {
| arch/x86/ia32/audit.c:16:10: warning: no previous extern declaration for non-static variable 'ia32_write_class' [-Wmissing-variable-declarations]
| 16 | unsigned ia32_write_class[] = {
| arch/x86/ia32/audit.c:21:10: warning: no previous extern declaration for non-static variable 'ia32_read_class' [-Wmissing-variable-declarations]
| 21 | unsigned ia32_read_class[] = {
| arch/x86/ia32/audit.c:26:10: warning: no previous extern declaration for non-static variable 'ia32_signal_class' [-Wmissing-variable-declarations]
| 26 | unsigned ia32_signal_class[] = {
These warnings occur due to their respective extern declarations being
scoped inside of audit_classes_init as well as only being enabled with
`CONFIG_IA32_EMULATION=y`:
| static int __init audit_classes_init(void)
| {
| #ifdef CONFIG_IA32_EMULATION
| extern __u32 ia32_dir_class[];
| extern __u32 ia32_write_class[];
| extern __u32 ia32_read_class[];
| extern __u32 ia32_chattr_class[];
| audit_register_class(AUDIT_CLASS_WRITE_32, ia32_write_class);
| audit_register_class(AUDIT_CLASS_READ_32, ia32_read_class);
| audit_register_class(AUDIT_CLASS_DIR_WRITE_32, ia32_dir_class);
| audit_register_class(AUDIT_CLASS_CHATTR_32, ia32_chattr_class);
| #endif
| audit_register_class(AUDIT_CLASS_WRITE, write_class);
| audit_register_class(AUDIT_CLASS_READ, read_class);
| audit_register_class(AUDIT_CLASS_DIR_WRITE, dir_class);
| audit_register_class(AUDIT_CLASS_CHATTR, chattr_class);
| return 0;
| }
Lift the extern declarations to their own header and resolve scoping
issues (and thus fix the warnings).
Moreover, change __u32 to unsigned so that we match the definitions:
| unsigned ia32_dir_class[] = {
| #include <asm-generic/audit_dir_write.h>
| ~0U
| };
|
| unsigned ia32_chattr_class[] = {
| #include <asm-generic/audit_change_attr.h>
| ~0U
| };
| ...
This patch is similar to commit:
0e5e3d4461a22d73 ("x86/audit: Fix a -Wmissing-prototypes warning for ia32_classify_syscall()") [1]
Signed-off-by: Justin Stitt <justinstitt@google.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/all/20200516123816.2680-1-b.thiel@posteo.de/ [1]
Link: https://github.com/ClangBuiltLinux/linux/issues/1920
Link: https://lore.kernel.org/r/20230829-missingvardecl-audit-v1-1-34efeb7f3539@google.com
|
| |\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull tracing updates from Steven Rostedt:
"User visible changes:
- Added a way to easier filter with cpumasks:
# echo 'cpumask & CPUS{17-42}' > /sys/kernel/tracing/events/ipi_send_cpumask/filter
- Show actual size of ring buffer after modifying the ring buffer
size via buffer_size_kb.
Currently it just returns what was written, but the actual size
rounds up to the sub buffer size. Show that real size instead.
Major changes:
- Added "eventfs". This is the code that handles the inodes and
dentries of tracefs/events directory. As there are thousands of
events, and each event has several inodes and dentries that
currently exist even when tracing is never used, they take up
precious memory. Instead, eventfs will allocate the inodes and
dentries in a JIT way (similar to what procfs does). There is now
metadata that handles the events and subdirectories, and will
create the inodes and dentries when they are used.
Note, I also have patches that remove the subdirectory meta data,
but will wait till the next merge window before applying them. It's
a little more complex, and I want to make sure the dynamic code
works properly before adding more complexity, making it easier to
revert if need be.
Minor changes:
- Optimization to user event list traversal
- Remove intermediate permission of tracefs files (note the
intermediate permission removes all access to the files so it is
not a security concern, but just a clean up)
- Add the complex fix to FORTIFY_SOURCE to the kernel stack event
logic
- Other minor cleanups"
* tag 'trace-v6.6' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (29 commits)
tracefs: Remove kerneldoc from struct eventfs_file
tracefs: Avoid changing i_mode to a temp value
tracing/user_events: Optimize safe list traversals
ftrace: Remove empty declaration ftrace_enable_daemon() and ftrace_disable_daemon()
tracing: Remove unused function declarations
tracing/filters: Document cpumask filtering
tracing/filters: Further optimise scalar vs cpumask comparison
tracing/filters: Optimise CPU vs cpumask filtering when the user mask is a single CPU
tracing/filters: Optimise scalar vs cpumask filtering when the user mask is a single CPU
tracing/filters: Optimise cpumask vs cpumask filtering when user mask is a single CPU
tracing/filters: Enable filtering the CPU common field by a cpumask
tracing/filters: Enable filtering a scalar field by a cpumask
tracing/filters: Enable filtering a cpumask field by another cpumask
tracing/filters: Dynamically allocate filter_pred.regex
test: ftrace: Fix kprobe test for eventfs
eventfs: Move tracing/events to eventfs
eventfs: Implement removal of meta data from eventfs
eventfs: Implement functions to create files and dirs when accessed
eventfs: Implement eventfs lookup, read, open functions
eventfs: Implement eventfs file add functions
...
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The struct eventfs_file is a local structure and should not be parsed by
kernel doc. It also does not fully follow the kerneldoc format and is
causing kerneldoc to spit out errors. Replace the /** to /* so that
kerneldoc no longer processes this structure.
Also format the comments of the delete union of the structure to be a bit
better.
Link: https://lore.kernel.org/linux-trace-kernel/20230818201414.2729745-1-willy@infradead.org/
Link: https://lore.kernel.org/linux-trace-kernel/20230822053313.77aa3397@rorschach.local.home
Cc: Mark Rutland <mark.rutland@arm.com>
Reported-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Right now inode->i_mode is updated twice to reach the desired value
in tracefs_apply_options(). Because there is no lock protecting the two
writes, other threads might read the intermediate value of inode->i_mode.
Thread-1 Thread-2
// tracefs_apply_options() //e.g., acl_permission_check
inode->i_mode &= ~S_IALLUGO;
unsigned int mode = inode->i_mode;
inode->i_mode |= opts->mode;
I think there is no need to introduce a lock but it is better to
only update inode->i_mode ONCE, so the readers will either see the old
or latest value, rather than an intermediate/temporary value.
Note, the race is not a security concern as the intermediate value is more
locked down than either the start or end version. This is more just to do
the conversion cleanly.
Link: https://lore.kernel.org/linux-trace-kernel/AB5B0A1C-75D9-4E82-A7F0-CF7D0715587B@gmail.com
Signed-off-by: Sishuai Gong <sishuai.system@gmail.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Several of the list traversals in the user_events facility use safe list
traversals where they could be using the unsafe versions instead.
Replace these safe traversals with their unsafe counterparts in the
interest of optimization.
Link: https://lore.kernel.org/linux-trace-kernel/20230810194337.695983-1-ervaughn@linux.microsoft.com
Suggested-by: Beau Belgrave <beaub@linux.microsoft.com>
Signed-off-by: Eric Vaughn <ervaughn@linux.microsoft.com>
Acked-by: Beau Belgrave <beaub@linux.microsoft.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
ftrace_disable_daemon()
The definition of ftrace_enable_daemon() and ftrace_disable_daemon() has
been removed since commit cb7be3b2fc2c ("ftrace: remove daemon"), remain
the declarations in the header files, so remove it.
Link: https://lore.kernel.org/linux-trace-kernel/20230804013636.115940-1-zhangzekun11@huawei.com
Cc: <mhiramat@kernel.org>
Cc: <mark.rutland@arm.com>
Signed-off-by: Zhang Zekun <zhangzekun11@huawei.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Commit 9457158bbc0e ("tracing: Fix reset of time stamps during trace_clock changes")
left behind tracing_reset_current() declaration.
Also commit 6954e415264e ("tracing: Place trace_pid_list logic into abstract functions")
removed trace_free_pid_list() implementation but leave declaration.
Link: https://lore.kernel.org/linux-trace-kernel/20230803144028.25492-1-yuehaibing@huawei.com
Cc: <mhiramat@kernel.org>
Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Cpumask, scalar and CPU fields can now be filtered by a user-provided
cpumask, document the syntax.
Link: https://lkml.kernel.org/r/20230707172155.70873-10-vschneid@redhat.com
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Leonardo Bras <leobras@redhat.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Per the previous commits, we now only enter do_filter_scalar_cpumask() with
a mask of weight greater than one. Optimise the equality checks.
Link: https://lkml.kernel.org/r/20230707172155.70873-9-vschneid@redhat.com
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Leonardo Bras <leobras@redhat.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Suggested-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
single CPU
Steven noted that when the user-provided cpumask contains a single CPU,
then the filtering function can use a scalar as input instead of a
full-fledged cpumask.
In this case we can directly re-use filter_pred_cpu(), we just need to
transform '&' into '==' before executing it.
Link: https://lkml.kernel.org/r/20230707172155.70873-8-vschneid@redhat.com
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Leonardo Bras <leobras@redhat.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Suggested-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
a single CPU
Steven noted that when the user-provided cpumask contains a single CPU,
then the filtering function can use a scalar as input instead of a
full-fledged cpumask.
When the mask contains a single CPU, directly re-use the unsigned field
predicate functions. Transform '&' into '==' beforehand.
Link: https://lkml.kernel.org/r/20230707172155.70873-7-vschneid@redhat.com
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Leonardo Bras <leobras@redhat.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Suggested-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
single CPU
Steven noted that when the user-provided cpumask contains a single CPU,
then the filtering function can use a scalar as input instead of a
full-fledged cpumask.
Reuse do_filter_scalar_cpumask() when the input mask has a weight of one.
Link: https://lkml.kernel.org/r/20230707172155.70873-6-vschneid@redhat.com
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Leonardo Bras <leobras@redhat.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Suggested-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The tracing_cpumask lets us specify which CPUs are traced in a buffer
instance, but doesn't let us do this on a per-event basis (unless one
creates an instance per event).
A previous commit added filtering scalar fields by a user-given cpumask,
make this work with the CPU common field as well.
This enables doing things like
$ trace-cmd record -e 'sched_switch' -f 'CPU & CPUS{12-52}' \
-e 'sched_wakeup' -f 'target_cpu & CPUS{12-52}'
Link: https://lkml.kernel.org/r/20230707172155.70873-5-vschneid@redhat.com
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Leonardo Bras <leobras@redhat.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Several events use a scalar field to denote a CPU:
o sched_wakeup.target_cpu
o sched_migrate_task.orig_cpu,dest_cpu
o sched_move_numa.src_cpu,dst_cpu
o ipi_send_cpu.cpu
o ...
Filtering these currently requires using arithmetic comparison functions,
which can be tedious when dealing with interleaved SMT or NUMA CPU ids.
Allow these to be filtered by a user-provided cpumask, which enables e.g.:
$ trace-cmd record -e 'sched_wakeup' -f 'target_cpu & CPUS{2,4,6,8-32}'
Link: https://lkml.kernel.org/r/20230707172155.70873-4-vschneid@redhat.com
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Leonardo Bras <leobras@redhat.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The recently introduced ipi_send_cpumask trace event contains a cpumask
field, but it currently cannot be used in filter expressions.
Make event filtering aware of cpumask fields, and allow these to be
filtered by a user-provided cpumask.
The user-provided cpumask is to be given in cpulist format and wrapped as:
"CPUS{$cpulist}". The use of curly braces instead of parentheses is to
prevent predicate_parse() from parsing the contents of CPUS{...} as a
full-fledged predicate subexpression.
This enables e.g.:
$ trace-cmd record -e 'ipi_send_cpumask' -f 'cpumask & CPUS{2,4,6,8-32}'
Link: https://lkml.kernel.org/r/20230707172155.70873-3-vschneid@redhat.com
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Leonardo Bras <leobras@redhat.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Every predicate allocation includes a MAX_FILTER_STR_VAL (256) char array
in the regex field, even if the predicate function does not use the field.
A later commit will introduce a dynamically allocated cpumask to struct
filter_pred, which will require a dedicated freeing function. Bite the
bullet and make filter_pred.regex dynamically allocated.
While at it, reorder the fields of filter_pred to fill in the byte
holes. The struct now fits on a single cacheline.
No change in behaviour intended.
The kfree()'s were patched via Coccinelle:
@@
struct filter_pred *pred;
@@
-kfree(pred);
+free_predicate(pred);
Link: https://lkml.kernel.org/r/20230707172155.70873-2-vschneid@redhat.com
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Leonardo Bras <leobras@redhat.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
kprobe_args_char.tc, kprobe_args_string.tc has validation check
for tracefs_create_dir, for eventfs it should be eventfs_create_dir.
Link: https://lkml.kernel.org/r/1690568452-46553-11-git-send-email-akaher@vmware.com
Signed-off-by: Ajay Kaher <akaher@vmware.com>
Co-developed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Tested-by: Ching-lin Yu <chinglinyu@google.com>
Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Up until now, /sys/kernel/tracing/events was no different than any other
part of tracefs. The files and directories within the events directory was
created when the tracefs was mounted, and also created for the instances in
/sys/kernel/tracing/instances/<instance>/events. Most of these files and
directories will never be referenced. Since there are thousands of these
files and directories they spend their time wasting precious memory
resources.
Move the "events" directory to the new eventfs. The eventfs will take the
meta data of the events that they represent and store that. When the files
in the events directory are referenced, the dentry and inodes to represent
them are then created. When the files are no longer referenced, they are
freed. This saves the precious memory resources that were wasted on these
seldom referenced dentries and inodes.
Running the following:
~# cat /proc/meminfo /proc/slabinfo > before.out
~# mkdir /sys/kernel/tracing/instances/foo
~# cat /proc/meminfo /proc/slabinfo > after.out
to test the changes produces the following deltas:
Before this change:
Before after deltas for meminfo:
MemFree: -32260
MemAvailable: -21496
KReclaimable: 21528
Slab: 22440
SReclaimable: 21528
SUnreclaim: 912
VmallocUsed: 16
Before after deltas for slabinfo:
<slab>: <objects> [ * <size> = <total>]
tracefs_inode_cache: 14472 [* 1184 = 17134848]
buffer_head: 24 [* 168 = 4032]
hmem_inode_cache: 28 [* 1480 = 41440]
dentry: 14450 [* 312 = 4508400]
lsm_inode_cache: 14453 [* 32 = 462496]
vma_lock: 11 [* 152 = 1672]
vm_area_struct: 2 [* 184 = 368]
trace_event_file: 1748 [* 88 = 153824]
kmalloc-256: 1072 [* 256 = 274432]
kmalloc-64: 2842 [* 64 = 181888]
Total slab additions in size: 22,763,400 bytes
With this change:
Before after deltas for meminfo:
MemFree: -12600
MemAvailable: -12580
Cached: 24
Active: 12
Inactive: 68
Inactive(anon): 48
Active(file): 12
Inactive(file): 20
Dirty: -4
AnonPages: 68
KReclaimable: 12
Slab: 1856
SReclaimable: 12
SUnreclaim: 1844
KernelStack: 16
PageTables: 36
VmallocUsed: 16
Before after deltas for slabinfo:
<slab>: <objects> [ * <size> = <total>]
tracefs_inode_cache: 108 [* 1184 = 127872]
buffer_head: 24 [* 168 = 4032]
hmem_inode_cache: 18 [* 1480 = 26640]
dentry: 127 [* 312 = 39624]
lsm_inode_cache: 152 [* 32 = 4864]
vma_lock: 67 [* 152 = 10184]
vm_area_struct: -12 [* 184 = -2208]
trace_event_file: 1764 [* 96 = 169344]
kmalloc-96: 14322 [* 96 = 1374912]
kmalloc-64: 2814 [* 64 = 180096]
kmalloc-32: 1103 [* 32 = 35296]
kmalloc-16: 2308 [* 16 = 36928]
kmalloc-8: 12800 [* 8 = 102400]
Total slab additions in size: 2,109,984 bytes
Which is a savings of 20,653,416 bytes (20 MB) per tracing instance.
Link: https://lkml.kernel.org/r/1690568452-46553-10-git-send-email-akaher@vmware.com
Signed-off-by: Ajay Kaher <akaher@vmware.com>
Co-developed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Tested-by: Ching-lin Yu <chinglinyu@google.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When events are removed from tracefs, the eventfs must be aware of this.
The eventfs_remove() removes the meta data from eventfs so that it will no
longer create the files associated with that event.
When an instance is removed from tracefs, eventfs_remove_events_dir() will
remove and clean up the entire "events" directory.
The helper function eventfs_remove_rec() is used to clean up and free the
associated data from eventfs for both of the added functions. SRCU is used
to protect the lists of meta data stored in the eventfs. The eventfs_mutex
is used to protect the content of the items in the list.
As lookups may be happening as deletions of events are made, the freeing
of dentry/inodes and relative information is done after the SRCU grace
period has passed.
Link: https://lkml.kernel.org/r/1690568452-46553-9-git-send-email-akaher@vmware.com
Signed-off-by: Ajay Kaher <akaher@vmware.com>
Co-developed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Tested-by: Ching-lin Yu <chinglinyu@google.com>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202305030611.Kas747Ev-lkp@intel.com/
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Add create_file() and create_dir() functions to create the files and
directories respectively when they are accessed. The functions will be
called from the lookup operation of the inode_operations or from the open
function of file_operations.
Link: https://lkml.kernel.org/r/1690568452-46553-8-git-send-email-akaher@vmware.com
Signed-off-by: Ajay Kaher <akaher@vmware.com>
Co-developed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Tested-by: Ching-lin Yu <chinglinyu@google.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Add the inode_operations, file_operations, and helper functions to eventfs:
dcache_dir_open_wrapper()
eventfs_root_lookup()
eventfs_release()
eventfs_set_ef_status_free()
eventfs_post_create_dir()
The inode_operations and file_operations functions will be called from the
VFS layer.
create_file() and create_dir() are added as stub functions and will be
filled in later.
Link: https://lkml.kernel.org/r/1690568452-46553-7-git-send-email-akaher@vmware.com
Signed-off-by: Ajay Kaher <akaher@vmware.com>
Co-developed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Tested-by: Ching-lin Yu <chinglinyu@google.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Add the following functions to add files to evenfs:
eventfs_add_events_file() to add the data needed to create a specific file
located at the top level events directory. The dentry/inode will be
created when the events directory is scanned.
eventfs_add_file() to add the data needed for files within the directories
below the top level events directory. The dentry/inode of the file will be
created when the directory that the file is in is scanned.
Link: https://lkml.kernel.org/r/1690568452-46553-6-git-send-email-akaher@vmware.com
Signed-off-by: Ajay Kaher <akaher@vmware.com>
Co-developed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Tested-by: Ching-lin Yu <chinglinyu@google.com>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202305051619.9a469a9a-yujie.liu@intel.com
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Add eventfs_file structure which will hold the properties of the eventfs
files and directories.
Add following functions to create the directories in eventfs:
eventfs_create_events_dir() will create the top level "events" directory
within the tracefs file system.
eventfs_add_subsystem_dir() creates an eventfs_file descriptor with the
given name of the subsystem.
eventfs_add_dir() creates an eventfs_file descriptor with the given name of
the directory and attached to a eventfs_file of a subsystem.
Add tracefs_inode structure to hold the inodes, flags and pointers to
private data used by eventfs.
Link: https://lkml.kernel.org/r/1690568452-46553-5-git-send-email-akaher@vmware.com
Signed-off-by: Ajay Kaher <akaher@vmware.com>
Co-developed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Tested-by: Ching-lin Yu <chinglinyu@google.com>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202305051619.9a469a9a-yujie.liu@intel.com
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Export a few tracefs functions that will be needed by the eventfs dynamic
file system. Rename them to start with "tracefs_" to keep with the name
space.
start_creating -> tracefs_start_creating
failed_creating -> tracefs_failed_creating
end_creating -> tracefs_end_creating
Link: https://lkml.kernel.org/r/1690568452-46553-4-git-send-email-akaher@vmware.com
Signed-off-by: Ajay Kaher <akaher@vmware.com>
Co-developed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Tested-by: Ching-lin Yu <chinglinyu@google.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Create a kmem cache of tracefs_inodes. To be more efficient, as there are
lots of tracefs inodes, create its own cache. This also allows to see how
many tracefs inodes have been created.
Add helper functions:
tracefs_alloc_inode()
tracefs_free_inode()
get_tracefs()
Link: https://lkml.kernel.org/r/1690568452-46553-3-git-send-email-akaher@vmware.com
Signed-off-by: Ajay Kaher <akaher@vmware.com>
Co-developed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Tested-by: Ching-lin Yu <chinglinyu@google.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The creation of the trace event directory requires that a TRACE_SYSTEM is
defined that the trace event directory is added within the system it was
defined in.
The code handled the case where a TRACE_SYSTEM was not added, and would
then add the event at the events directory. But nothing should be doing
this. This code also prevents the implementation of creating dynamic
dentrys for the eventfs system.
As this path has never been hit on correct code, remove it. If it does get
hit, issues a WARN_ON_ONCE() and return ENODEV.
Link: https://lkml.kernel.org/r/1690568452-46553-2-git-send-email-akaher@vmware.com
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Ajay Kaher <akaher@vmware.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Currently we can resize trace ringbuffer by writing a value into file
'buffer_size_kb', then by reading the file, we get the value that is
usually what we wrote. However, this value may be not actual size of
trace ring buffer because of the round up when doing resize in kernel,
and the actual size would be more useful.
Link: https://lore.kernel.org/linux-trace-kernel/20230705002705.576633-1-zhengyejian1@huawei.com
Cc: <mhiramat@kernel.org>
Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
As the trace iterator is created and used by various interfaces, the clean
up of it needs to be consistent. Create a free_trace_iter_content() helper
function that frees the content of the iterator and use that to clean it
up in all places that it is used.
Link: https://lkml.kernel.org/r/20230715141348.341887497@goodmis.org
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The iterator allocated a descriptor to copy the current_trace. This was done
with the assumption that the function pointers might change. But this was a
false assuption, as it does not change. There's no reason to make a copy of the
current_trace and just use the pointer it points to. This removes needing to
manage freeing the descriptor. Worse yet, there's locations that the iterator
is used but does make a copy and just uses the pointer. This could cause the
actual pointer to the trace descriptor to be freed and not the allocated copy.
This is more of a clean up than a fix.
Link: https://lkml.kernel.org/r/20230715141348.135792275@goodmis.org
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Fixes: d7350c3f45694 ("tracing/core: make the read callbacks reentrants")
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Use try_cmpxchg instead of cmpxchg (*ptr, old, new) == old in
ring_buffer.c. x86 CMPXCHG instruction returns success in ZF flag,
so this change saves a compare after cmpxchg (and related move
instruction in front of cmpxchg).
No functional change intended.
Link: https://lore.kernel.org/linux-trace-kernel/20230714154418.8884-1-ubizjak@gmail.com
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
For backward compatibility, older tooling expects to see the kernel_stack
event with a "caller" field that is a fixed size array of 8 addresses. The
code now supports more than 8 with an added "size" field that states the
real number of entries. But the "caller" field still just looks like a
fixed size to user space.
Since the tracing macros that create the user space format files also
creates the structures that those files represent, the kernel_stack event
structure had its "caller" field a fixed size of 8, but in reality, when
it is allocated on the ring buffer, it can hold more if the stack trace is
bigger that 8 functions. The copying of these entries was simply done with
a memcpy():
size = nr_entries * sizeof(unsigned long);
memcpy(entry->caller, fstack->calls, size);
The FORTIFY_SOURCE logic noticed at runtime that when the nr_entries was
larger than 8, that the memcpy() was writing more than what the structure
stated it can hold and it complained about it. This is because the
FORTIFY_SOURCE code is unaware that the amount allocated is actually
enough to hold the size. It does not expect that a fixed size field will
hold more than the fixed size.
This was originally solved by hiding the caller assignment with some
pointer arithmetic.
ptr = ring_buffer_data();
entry = ptr;
ptr += offsetof(typeof(*entry), caller);
memcpy(ptr, fstack->calls, size);
But it is considered bad form to hide from kernel hardening. Instead, make
it work nicely with FORTIFY_SOURCE by adding a new __stack_array() macro
that is specific for this one special use case. The macro will take 4
arguments: type, item, len, field (whereas the __array() macro takes just
the first three). This macro will act just like the __array() macro when
creating the code to deal with the format file that is exposed to user
space. But for the kernel, it will turn the caller field into:
type item[] __counted_by(field);
or for this instance:
unsigned long caller[] __counted_by(size);
Now the kernel code can expose the assignment of the caller to the
FORTIFY_SOURCE and everyone is happy!
Link: https://lore.kernel.org/linux-trace-kernel/20230712105235.5fc441aa@gandalf.local.home/
Link: https://lore.kernel.org/linux-trace-kernel/20230713092605.2ddb9788@rorschach.local.home
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Sven Schnelle <svens@linux.ibm.com>
Suggested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
|
| |\ \ \
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Pull workqueue updates from Tejun Heo:
- Unbound workqueues now support more flexible affinity scopes.
The default behavior is to soft-affine according to last level cache
boundaries. A work item queued from a given LLC is executed by a
worker running on the same LLC but the worker may be moved across
cache boundaries as the scheduler sees fit. On machines which
multiple L3 caches, which are becoming more popular along with
chiplet designs, this improves cache locality while not harming work
conservation too much.
Unbound workqueues are now also a lot more flexible in terms of
execution affinity. Differeing levels of affinity scopes are
supported and both the default and per-workqueue affinity settings
can be modified dynamically. This should help working around amny of
sub-optimal behaviors observed recently with asymmetric ARM CPUs.
This involved signficant restructuring of workqueue code. Nothing was
reported yet but there's some risk of subtle regressions. Should keep
an eye out.
- Rescuer workers now has more identifiable comms.
- workqueue.unbound_cpus added so that CPUs which can be used by
workqueue can be constrained early during boot.
- Now that all the in-tree users have been flushed out, trigger warning
if system-wide workqueues are flushed.
* tag 'wq-for-6.6' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq: (31 commits)
workqueue: fix data race with the pwq->stats[] increment
workqueue: Rename rescuer kworker
workqueue: Make default affinity_scope dynamically updatable
workqueue: Add "Affinity Scopes and Performance" section to documentation
workqueue: Implement non-strict affinity scope for unbound workqueues
workqueue: Add workqueue_attrs->__pod_cpumask
workqueue: Factor out need_more_worker() check and worker wake-up
workqueue: Factor out work to worker assignment and collision handling
workqueue: Add multiple affinity scopes and interface to select them
workqueue: Modularize wq_pod_type initialization
workqueue: Add tools/workqueue/wq_dump.py which prints out workqueue configuration
workqueue: Generalize unbound CPU pods
workqueue: Factor out clearing of workqueue-only attrs fields
workqueue: Factor out actual cpumask calculation to reduce subtlety in wq_update_pod()
workqueue: Initialize unbound CPU pods later in the boot
workqueue: Move wq_pod_init() below workqueue_init()
workqueue: Rename NUMA related names to use pod instead
workqueue: Rename workqueue_attrs->no_numa to ->ordered
workqueue: Make unbound workqueues to use per-cpu pool_workqueues
workqueue: Call wq_update_unbound_numa() on all CPUs in NUMA node on CPU hotplug
...
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
KCSAN has discovered a data race in kernel/workqueue.c:2598:
[ 1863.554079] ==================================================================
[ 1863.554118] BUG: KCSAN: data-race in process_one_work / process_one_work
[ 1863.554142] write to 0xffff963d99d79998 of 8 bytes by task 5394 on cpu 27:
[ 1863.554154] process_one_work (kernel/workqueue.c:2598)
[ 1863.554166] worker_thread (./include/linux/list.h:292 kernel/workqueue.c:2752)
[ 1863.554177] kthread (kernel/kthread.c:389)
[ 1863.554186] ret_from_fork (arch/x86/kernel/process.c:145)
[ 1863.554197] ret_from_fork_asm (arch/x86/entry/entry_64.S:312)
[ 1863.554213] read to 0xffff963d99d79998 of 8 bytes by task 5450 on cpu 12:
[ 1863.554224] process_one_work (kernel/workqueue.c:2598)
[ 1863.554235] worker_thread (./include/linux/list.h:292 kernel/workqueue.c:2752)
[ 1863.554247] kthread (kernel/kthread.c:389)
[ 1863.554255] ret_from_fork (arch/x86/kernel/process.c:145)
[ 1863.554266] ret_from_fork_asm (arch/x86/entry/entry_64.S:312)
[ 1863.554280] value changed: 0x0000000000001766 -> 0x000000000000176a
[ 1863.554295] Reported by Kernel Concurrency Sanitizer on:
[ 1863.554303] CPU: 12 PID: 5450 Comm: kworker/u64:1 Tainted: G L 6.5.0-rc6+ #44
[ 1863.554314] Hardware name: ASRock X670E PG Lightning/X670E PG Lightning, BIOS 1.21 04/26/2023
[ 1863.554322] Workqueue: btrfs-endio btrfs_end_bio_work [btrfs]
[ 1863.554941] ==================================================================
lockdep_invariant_state(true);
→ pwq->stats[PWQ_STAT_STARTED]++;
trace_workqueue_execute_start(work);
worker->current_func(work);
Moving pwq->stats[PWQ_STAT_STARTED]++; before the line
raw_spin_unlock_irq(&pool->lock);
resolves the data race without performance penalty.
KCSAN detected at least one additional data race:
[ 157.834751] ==================================================================
[ 157.834770] BUG: KCSAN: data-race in process_one_work / process_one_work
[ 157.834793] write to 0xffff9934453f77a0 of 8 bytes by task 468 on cpu 29:
[ 157.834804] process_one_work (/home/marvin/linux/kernel/linux_torvalds/kernel/workqueue.c:2606)
[ 157.834815] worker_thread (/home/marvin/linux/kernel/linux_torvalds/./include/linux/list.h:292 /home/marvin/linux/kernel/linux_torvalds/kernel/workqueue.c:2752)
[ 157.834826] kthread (/home/marvin/linux/kernel/linux_torvalds/kernel/kthread.c:389)
[ 157.834834] ret_from_fork (/home/marvin/linux/kernel/linux_torvalds/arch/x86/kernel/process.c:145)
[ 157.834845] ret_from_fork_asm (/home/marvin/linux/kernel/linux_torvalds/arch/x86/entry/entry_64.S:312)
[ 157.834859] read to 0xffff9934453f77a0 of 8 bytes by task 214 on cpu 7:
[ 157.834868] process_one_work (/home/marvin/linux/kernel/linux_torvalds/kernel/workqueue.c:2606)
[ 157.834879] worker_thread (/home/marvin/linux/kernel/linux_torvalds/./include/linux/list.h:292 /home/marvin/linux/kernel/linux_torvalds/kernel/workqueue.c:2752)
[ 157.834890] kthread (/home/marvin/linux/kernel/linux_torvalds/kernel/kthread.c:389)
[ 157.834897] ret_from_fork (/home/marvin/linux/kernel/linux_torvalds/arch/x86/kernel/process.c:145)
[ 157.834907] ret_from_fork_asm (/home/marvin/linux/kernel/linux_torvalds/arch/x86/entry/entry_64.S:312)
[ 157.834920] value changed: 0x000000000000052a -> 0x0000000000000532
[ 157.834933] Reported by Kernel Concurrency Sanitizer on:
[ 157.834941] CPU: 7 PID: 214 Comm: kworker/u64:2 Tainted: G L 6.5.0-rc7-kcsan-00169-g81eaf55a60fc #4
[ 157.834951] Hardware name: ASRock X670E PG Lightning/X670E PG Lightning, BIOS 1.21 04/26/2023
[ 157.834958] Workqueue: btrfs-endio btrfs_end_bio_work [btrfs]
[ 157.835567] ==================================================================
in code:
trace_workqueue_execute_end(work, worker->current_func);
→ pwq->stats[PWQ_STAT_COMPLETED]++;
lock_map_release(&lockdep_map);
lock_map_release(&pwq->wq->lockdep_map);
which needs to be resolved separately.
Fixes: 725e8ec59c56c ("workqueue: Add pwq->stats[] and a monitoring script")
Cc: Tejun Heo <tj@kernel.org>
Suggested-by: Lai Jiangshan <jiangshanlai@gmail.com>
Link: https://lore.kernel.org/lkml/20230818194448.29672-1-mirsad.todorovac@alu.unizg.hr/
Signed-off-by: Mirsad Goran Todorovac <mirsad.todorovac@alu.unizg.hr>
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Each CPU-specific and unbound kworker kthread conforms to a particular
naming scheme. However, this does not extend to the rescuer kworker.
At present, a rescuer kworker is simply named according to its
workqueue's name. This can be cryptic.
This patch modifies a rescuer to follow the kworker naming scheme.
The "R" is indicative of a rescuer and after "-" is its workqueue's
name e.g. "kworker/R-ext4-rsv-conver".
tj: Use "R" instead of "r" as the prefix to make it more distinctive and
consistent with how highpri pools are marked.
Signed-off-by: Aaron Tomlin <atomlin@atomlin.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
While workqueue.default_affinity_scope is writable, it only affects
workqueues which are created afterwards and isn't very useful. Instead,
let's introduce explicit "default" scope and update the effective scope
dynamically when workqueue.default_affinity_scope is changed.
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
With affinity scopes and their strictness setting added, unbound workqueues
should now be able to cover wide variety of configurations and use cases.
Unfortunately, the performance picture is not entirely straight-forward due
to a trade-off between efficiency and work-conservation in some situations
necessitating manual configuration.
This patch adds "Affinity Scopes and Performance" section to
Documentation/core-api/workqueue.rst which illustrates the trade-off with a
set of experiments and provides some guidelines.
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
An unbound workqueue can be served by multiple worker_pools to improve
locality. The segmentation is achieved by grouping CPUs into pods. By
default, the cache boundaries according to cpus_share_cache() define the
CPUs are grouped. Let's a workqueue is allowed to run on all CPUs and the
system has two L3 caches. The workqueue would be mapped to two worker_pools
each serving one L3 cache domains.
While this improves locality, because the pod boundaries are strict, it
limits the total bandwidth a given issuer can consume. For example, let's
say there is a thread pinned to a CPU issuing enough work items to saturate
the whole machine. With the machine segmented into two pods, no matter how
many work items it issues, it can only use half of the CPUs on the system.
While this limitation has existed for a very long time, it wasn't very
pronounced because the affinity grouping used to be always by NUMA nodes.
With cache boundaries as the default and support for even finer grained
scopes (smt and cpu), it is now an a lot more pressing problem.
This patch implements non-strict affinity scope where the pod boundaries
aren't enforced strictly. Going back to the previous example, the workqueue
would still be mapped to two worker_pools; however, the affinity enforcement
would be soft. The workers in both pools would have their cpus_allowed set
to the whole machine thus allowing the scheduler to migrate them anywhere on
the machine. However, whenever an idle worker is woken up, the workqueue
code asks the scheduler to bring back the task within the pod if the worker
is outside. ie. work items start executing within its affinity scope but can
be migrated outside as the scheduler sees fit. This removes the hard cap on
utilization while maintaining the benefits of affinity scopes.
After the earlier ->__pod_cpumask changes, the implementation is pretty
simple. When non-strict which is the new default:
* pool_allowed_cpus() returns @pool->attrs->cpumask instead of
->__pod_cpumask so that the workers are allowed to run on any CPU that
the associated workqueues allow.
* If the idle worker task's ->wake_cpu is outside the pod, kick_pool() sets
the field to a CPU within the pod.
This would be the first use of task_struct->wake_cpu outside scheduler
proper, so it isn't clear whether this would be acceptable. However, other
methods of migrating tasks are significantly more expensive and are likely
prohibitively so if we want to do this on every work item. This needs
discussion with scheduler folks.
There is also a race window where setting ->wake_cpu wouldn't be effective
as the target task is still on CPU. However, the window is pretty small and
this being a best-effort optimization, it doesn't seem to warrant more
complexity at the moment.
While the non-strict cache affinity scopes seem to be the best option, the
performance picture interacts with the affinity scope and is a bit
complicated to fully discuss in this patch, so the behavior is made easily
selectable through wqattrs and sysfs and the next patch will add
documentation to discuss performance implications.
v2: pool->attrs->affn_strict is set to true for per-cpu worker_pools.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
workqueue_attrs has two uses:
* to specify the required unouned workqueue properties by users
* to match worker_pool's properties to workqueues by core code
For example, if the user wants to restrict a workqueue to run only CPUs 0
and 2, and the two CPUs are on different affinity scopes, the workqueue's
attrs->cpumask would contains CPUs 0 and 2, and the workqueue would be
associated with two worker_pools, one with attrs->cpumask containing just
CPU 0 and the other CPU 2.
Workqueue wants to support non-strict affinity scopes where work items are
started in their matching affinity scopes but the scheduler is free to
migrate them outside the starting scopes, which can enable utilizing the
whole machine while maintaining most of the locality benefits from affinity
scopes.
To enable that, worker_pools need to distinguish the strict affinity that it
has to follow (because that's the restriction coming from the user) and the
soft affinity that it wants to apply when dispatching work items. Note that
two worker_pools with different soft dispatching requirements have to be
separate; otherwise, for example, we'd be ping-ponging worker threads across
NUMA boundaries constantly.
This patch adds workqueue_attrs->__pod_cpumask. The new field is double
underscored as it's only used internally to distinguish worker_pools. A
worker_pool's ->cpumask is now always the same as the online subset of
allowed CPUs of the associated workqueues, and ->__pod_cpumask is the pod's
subset of that ->cpumask. Going back to the example above, both worker_pools
would have ->cpumask containing both CPUs 0 and 2 but one's ->__pod_cpumask
would contain 0 while the other's 2.
* pool_allowed_cpus() is added. It returns the worker_pool's strict cpumask
that the pool's workers must stay within. This is currently always
->__pod_cpumask as all boundaries are still strict.
* As a workqueue_attrs can now track both the associated workqueues' cpumask
and its per-pod subset, wq_calc_pod_cpumask() no longer needs an external
out-argument. Drop @cpumask and instead store the result in
->__pod_cpumask.
* The above also simplifies apply_wqattrs_prepare() as the same
workqueue_attrs can be used to create all pods associated with a
workqueue. tmp_attrs is dropped.
* wq_update_pod() is updated to use wqattrs_equal() to test whether a pwq
update is needed instead of only comparing ->cpumask so that
->__pod_cpumask is compared too. It can directly compare ->__pod_cpumaks
but the code is easier to understand and more robust this way.
The only user-visible behavior change is that two workqueues with different
cpumasks no longer can share worker_pools even when their pod subsets
coincide. Going back to the example, let's say there's another workqueue
with cpumask 0, 2, 3, where 2 and 3 are in the same pod. It would be mapped
to two worker_pools - one with CPU 0, the other with 2 and 3. The former has
the same cpumask as the first pod of the earlier example and would have
shared the same worker_pool but that's no longer the case after this patch.
The worker_pools would have the same ->__pod_cpumask but their ->cpumask's
wouldn't match.
While this is necessary to support non-strict affinity scopes, there can be
further optimizations to maintain sharing among strict affinity scopes.
However, non-strict affinity scopes are going to be preferable for most use
cases and we don't see very diverse mixture of unbound workqueue cpumasks
anyway, so the additional overhead doesn't seem to justify the extra
complexity.
v2: - wq_update_pod() was incorrectly comparing target_attrs->__pod_cpumask
to pool->attrs->cpumask instead of its ->__pod_cpumask. Fix it by
using wqattrs_equal() for comparison instead.
- Per-cpu worker pools weren't initializing ->__pod_cpumask which caused
a subtle problem later on. Set it to cpumask_of(cpu) like ->cpumask.
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Checking need_more_worker() and calling wake_up_worker() is a repeated
pattern. Let's add kick_pool(), which checks need_more_worker() and
open-code wake_up_worker(), and replace wake_up_worker() uses. The following
conversions aren't one-to-one:
* __queue_work() was using __need_more_work() because it knows that
pool->worklist isn't empty. Switching to kick_pool() adds an extra
list_empty() test.
* create_worker() always needs to wake up the newly minted worker whether
there's more work to do or not to avoid triggering hung task check on the
new task. Keep the current wake_up_process() and still add kick_pool().
This may lead to an extra wakeup which isn't harmful.
* pwq_adjust_max_active() was explicitly checking whether it needs to wake
up a worker or not to avoid spurious wakeups. As kick_pool() only wakes up
a worker when necessary, this explicit check is no longer necessary and
dropped.
* unbind_workers() now calls kick_pool() instead of wake_up_worker() adding
a need_more_worker() test. This avoids spurious wakeups and shouldn't
break anything.
wake_up_worker() is dropped as kick_pool() replaces all its users. After
this patch, all paths that wakes up a non-rescuer worker to initiate work
item execution use kick_pool(). This will enable future changes to improve
locality.
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The two work execution paths in worker_thread() and rescuer_thread() use
move_linked_works() to claim work items from @pool->worklist. Once claimed,
process_schedule_works() is called which invokes process_one_work() on each
work item. process_one_work() then uses find_worker_executing_work() to
detect and handle collisions - situations where the work item to be executed
is still running on another worker.
This works fine, but, to improve work execution locality, we want to
establish work to worker association earlier and know for sure that the
worker is going to excute the work once asssigned, which requires performing
collision handling earlier while trying to assign the work item to the
worker.
This patch introduces assign_work() which assigns a work item to a worker
using move_linked_works() and then performs collision handling. As collision
handling is handled earlier, process_one_work() no longer needs to worry
about them.
After the this patch, collision checks for linked work items are skipped,
which should be fine as they can't be queued multiple times concurrently.
For work items running from rescuers, the timing of collision handling may
change but the invariant that the work items go through collision handling
before starting execution does not.
This patch shouldn't cause noticeable behavior changes, especially given
that worker_thread() behavior remains the same.
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Add three more affinity scopes - WQ_AFFN_CPU, SMT and CACHE - and make CACHE
the default. The code changes to actually add the additional scopes are
trivial.
Also add module parameter "workqueue.default_affinity_scope" to override the
default scope and "affinity_scope" sysfs file to configure it per workqueue.
wq_dump.py and documentations are updated accordingly.
This enables significant flexibility in configuring how unbound workqueues
behave. If affinity scope is set to "cpu", it'll behave close to a per-cpu
workqueue. On the other hand, "system" removes all locality boundaries.
Many modern machines have multiple L3 caches often while being mostly
uniform in terms of memory access. Thus, workqueue's previous behavior of
spreading work items in each NUMA node had negative performance implications
from unncessarily crossing L3 boundaries between issue and execution.
However, picking a finer grained affinity scope also has a downside in that
an issuer in one group can't utilize CPUs in other groups.
While dependent on the specifics of workload, there's usually a noticeable
penalty in crossing L3 boundaries, so let's default to CACHE. This issue
will be further addressed and documented with examples in future patches.
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
While wq_pod_type[] can now group CPUs in any aribitrary way, WQ_AFFN_NUM
init is hard coded into workqueue_init_topology(). This patch modularizes
the init path by introducing init_pod_type() which takes a callback to
determine whether two CPUs should share a pod as an argument.
init_pod_type() first scans the CPU combinations testing for sharing to
assign consecutive pod IDs and initialize pod_type->cpu_pod[]. Once
->cpu_pod[] is determined, ->pod_cpus[] and ->pod_node[] are initialized
accordingly. WQ_AFFN_NUMA is now initialized by calling init_pod_type() with
cpus_share_numa() which tests whether the CPU belongs to the same NUMA node.
This patch may change the pod ID assigned to each NUMA node but that
shouldn't cause any behavior changes as the NUMA node to use for allocations
are tracked separately in pod_type->pod_node[]. This makes adding new
affinty types pretty easy.
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
configuration
Lack of visibility has always been a pain point for workqueues. While the
recently added wq_monitor.py improved the situation, it's still difficult to
understand what worker pools are active in the system, how workqueues map to
them and why. The lack of visibility into how workqueues are configured is
going to become more noticeable as workqueue improves locality awareness and
provides more mechanisms to customize locality related behaviors.
Now that the basic framework for more flexible locality support is in place,
this is a good time to improve the situation. This patch adds
tools/workqueues/wq_dump.py which prints out the topology configuration,
worker pools and how workqueues are mapped to pools. Read the command's help
message for more details.
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
While renamed to pod, the code still assumes that the pods are defined by
NUMA boundaries. Let's generalize it:
* workqueue_attrs->affn_scope is added. Each enum represents the type of
boundaries that define the pods. There are currently two scopes -
WQ_AFFN_NUMA and WQ_AFFN_SYSTEM. The former is the same behavior as before
- one pod per NUMA node. The latter defines one global pod across the
whole system.
* struct wq_pod_type is added which describes how pods are configured for
each affnity scope. For each pod, it lists the member CPUs and the
preferred NUMA node for memory allocations. The reverse mapping from CPU
to pod is also available.
* wq_pod_enabled is dropped. Pod is now always enabled. The previously
disabled behavior is now implemented through WQ_AFFN_SYSTEM.
* get_unbound_pool() wants to determine the NUMA node to allocate memory
from for the new pool. The variables are renamed from node to pod but the
logic still assumes they're one and the same. Clearly distinguish them -
walk the WQ_AFFN_NUMA pods to find the matching pod and then use the pod's
NUMA node.
* wq_calc_pod_cpumask() was taking @pod but assumed that it was the NUMA
node. Take @cpu instead and determine the cpumask to use from the pod_type
matching @attrs.
* apply_wqattrs_prepare() is update to return ERR_PTR() on error instead of
NULL so that it can indicate -EINVAL on invalid affinity scopes.
This patch allows CPUs to be grouped into pods however desired per type.
While this patch causes some internal behavior changes, nothing material
should change for workqueue users.
v2: Trigger WARN_ON_ONCE() in wqattrs_pod_type() if affn_scope is
WQ_AFFN_NR_TYPES which indicates that the function is called with a
worker_pool's attrs instead of a workqueue's.
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
workqueue_attrs can be used for both workqueues and worker_pools. However,
some fields, currently only ->ordered, only apply to workqueues and should
be cleared to the default / invalid values.
Currently, an unbound workqueue explicitly clears attrs->ordered in
get_unbound_pool() after copying the source workqueue attrs, while per-cpu
workqueues rely on the fact that zeroing on allocation gives us the desired
default value for pool->attrs->ordered.
This is fragile. Let's add wqattrs_clear_for_pool() which clears
attrs->ordered and is called from both init_worker_pool() and
get_unbound_pool(). This will ease adding more workqueue-only attrs fields.
In get_unbound_pool(), pool->node initialization is moved upwards for
readability. This shouldn't cause any behavior changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
wq_update_pod()
For an unbound pool, multiple cpumasks are involved.
U: The user-specified cpumask (may be filtered with cpu_possible_mask).
A: The actual cpumask filtered by wq_unbound_cpumask. If the filtering
leaves no CPU, wq_unbound_cpumask is used.
P: Per-pod subsets of #A.
wq->attrs stores #U, wq->dfl_pwq->pool->attrs->cpumask #A, and
wq->cpu_pwq[CPU]->pool->attrs->cpumask #P.
wq_update_pod() is called to update per-pod pwq's during CPU hotplug. To
calculate the new #P for each workqueue, it needs to call
wq_calc_pod_cpumask() with @attrs that contains #A. Currently,
wq_update_pod() achieves this by calling wq_calc_pod_cpumask() with
wq->dfl_pwq->pool->attrs.
This is rather fragile because we're calling wq_calc_pod_cpumask() with
@attrs of a worker_pool rather than the workqueue's actual attrs when what
we want to calculate is the workqueue's cpumask on the pod. While this works
fine currently, future changes will add fields which are used differently
between workqueues and worker_pools and this subtlety will bite us.
This patch factors out #U -> #A calculation from apply_wqattrs_prepare()
into wqattrs_actualize_cpumask and updates wq_update_pod() to copy
wq->unbound_attrs and use the new helper to obtain #A freshly instead of
abusing wq->dfl_pwq->pool_attrs.
This shouldn't cause any behavior changes in the current code.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: K Prateek Nayak <kprateek.nayak@amd.com>
Reference: http://lkml.kernel.org/r/30625cdd-4d61-594b-8db9-6816b017dde3@amd.com
|
