Patch series "use per-vma locks for /proc/pid/maps reads", v8.
Reading /proc/pid/maps requires read-locking mmap_lock which prevents any
other task from concurrently modifying the address space. This guarantees
coherent reporting of virtual address ranges, however it can block
important updates from happening. Oftentimes /proc/pid/maps readers are
low priority monitoring tasks and them blocking high priority tasks
results in priority inversion.
Locking the entire address space is required to present fully coherent
picture of the address space, however even current implementation does not
strictly guarantee that by outputting vmas in page-size chunks and
dropping mmap_lock in between each chunk. Address space modifications are
possible while mmap_lock is dropped and userspace reading the content is
expected to deal with possible concurrent address space modifications.
Considering these relaxed rules, holding mmap_lock is not strictly needed
as long as we can guarantee that a concurrently modified vma is reported
either in its original form or after it was modified.
This patchset switches from holding mmap_lock while reading /proc/pid/maps
to taking per-vma locks as we walk the vma tree. This reduces the
contention with tasks modifying the address space because they would have
to contend for the same vma as opposed to the entire address space.
Previous version of this patchset [1] tried to perform /proc/pid/maps
reading under RCU, however its implementation is quite complex and the
results are worse than the new version because it still relied on
mmap_lock speculation which retries if any part of the address space gets
modified. New implementaion is both simpler and results in less
contention. Note that similar approach would not work for /proc/pid/smaps
reading as it also walks the page table and that's not RCU-safe.
Paul McKenney's designed a test [2] to measure mmap/munmap latencies while
concurrently reading /proc/pid/maps. The test has a pair of processes
scanning /proc/PID/maps, and another process unmapping and remapping 4K
pages from a 128MB range of anonymous memory. At the end of each 10
second run, the latency of each mmap() or munmap() operation is measured,
and for each run the maximum and mean latency is printed. The map/unmap
process is started first, its PID is passed to the scanners, and then the
map/unmap process waits until both scanners are running before starting
its timed test. The scanners keep scanning until the specified
/proc/PID/maps file disappears.
The latest results from Paul:
Stock mm-unstable, all of the runs had maximum latencies in excess of 0.5
milliseconds, and with 80% of the runs' latencies exceeding a full
millisecond, and ranging up beyond 4 full milliseconds. In contrast, 99%
of the runs with this patch series applied had maximum latencies of less
than 0.5 milliseconds, with the single outlier at only 0.608 milliseconds.
From a median-performance (as opposed to maximum-latency) viewpoint, this
patch series also looks good, with stock mm weighing in at 11 microseconds
and patch series at 6 microseconds, better than a 2x improvement.
Before the change:
./run-proc-vs-map.sh --nsamples 100 --rawdata -- --busyduration 2
0.011 0.008 0.521
0.011 0.008 0.552
0.011 0.008 0.590
0.011 0.008 0.660
...
0.011 0.015 2.987
0.011 0.015 3.038
0.011 0.016 3.431
0.011 0.016 4.707
After the change:
./run-proc-vs-map.sh --nsamples 100 --rawdata -- --busyduration 2
0.006 0.005 0.026
0.006 0.005 0.029
0.006 0.005 0.034
0.006 0.005 0.035
...
0.006 0.006 0.421
0.006 0.006 0.423
0.006 0.006 0.439
0.006 0.006 0.608
The patchset also adds a number of tests to check for /proc/pid/maps data
coherency. They are designed to detect any unexpected data tearing while
performing some common address space modifications (vma split, resize and
remap). Even before these changes, reading /proc/pid/maps might have
inconsistent data because the file is read page-by-page with mmap_lock
being dropped between the pages. An example of user-visible inconsistency
can be that the same vma is printed twice: once before it was modified and
then after the modifications. For example if vma was extended, it might
be found and reported twice. What is not expected is to see a gap where
there should have been a vma both before and after modification. This
patchset increases the chances of such tearing, therefore it's even more
important now to test for unexpected inconsistencies.
In [3] Lorenzo identified the following possible vma merging/splitting
scenarios:
Merges with changes to existing vmas:
1 Merge both - mapping a vma over another one and between two vmas which
can be merged after this replacement;
2. Merge left full - mapping a vma at the end of an existing one and
completely over its right neighbor;
3. Merge left partial - mapping a vma at the end of an existing one and
partially over its right neighbor;
4. Merge right full - mapping a vma before the start of an existing one
and completely over its left neighbor;
5. Merge right partial - mapping a vma before the start of an existing one
and partially over its left neighbor;
Merges without changes to existing vmas:
6. Merge both - mapping a vma into a gap between two vmas which can be
merged after the insertion;
7. Merge left - mapping a vma at the end of an existing one;
8. Merge right - mapping a vma before the start end of an existing one;
Splits
9. Split with new vma at the lower address;
10. Split with new vma at the higher address;
If such merges or splits happen concurrently with the /proc/maps reading
we might report a vma twice, once before the modification and once after
it is modified:
Case 1 might report overwritten and previous vma along with the final
merged vma;
Case 2 might report previous and the final merged vma;
Case 3 might cause us to retry once we detect the temporary gap caused by
shrinking of the right neighbor;
Case 4 might report overritten and the final merged vma;
Case 5 might cause us to retry once we detect the temporary gap caused by
shrinking of the left neighbor;
Case 6 might report previous vma and the gap along with the final marged
vma;
Case 7 might report previous and the final merged vma;
Case 8 might report the original gap and the final merged vma covering the
gap;
Case 9 might cause us to retry once we detect the temporary gap caused by
shrinking of the original vma at the vma start;
Case 10 might cause us to retry once we detect the temporary gap caused by
shrinking of the original vma at the vma end;
In all these cases the retry mechanism prevents us from reporting possible
temporary gaps.
[1] https://lore.kernel.org/all/20250418174959.1431962-1-surenb@google.com/
[2] https://github.com/paulmckrcu/proc-mmap_sem-test
[3] https://lore.kernel.org/all/e1863f40-39ab-4e5b-984a-c48765ffde1c@lucifer.local/
The /proc/pid/maps file is generated page by page, with the mmap_lock
released between pages. This can lead to inconsistent reads if the
underlying vmas are concurrently modified. For instance, if a vma split
or merge occurs at a page boundary while /proc/pid/maps is being read, the
same vma might be seen twice: once before and once after the change. This
duplication is considered acceptable for userspace handling. However,
observing a "hole" where a vma should be (e.g., due to a vma being
replaced and the space temporarily being empty) is unacceptable.
Implement a test that:
1. Forks a child process which continuously modifies its address
space, specifically targeting a vma at the boundary between two pages.
2. The parent process repeatedly reads the child's /proc/pid/maps.
3. The parent process checks the last vma of the first page and the
first vma of the second page for consistency, looking for the effects
of vma splits or merges.
The test duration is configurable via DURATION environment variable
expressed in seconds. The default test duration is 5 seconds.
Example Command: DURATION=10 ./proc-maps-race
Link: https://lore.kernel.org/all/20250418174959.1431962-1-surenb@google.com/ [1]
Link: https://github.com/paulmckrcu/proc-mmap_sem-test [2]
Link: https://lore.kernel.org/all/e1863f40-39ab-4e5b-984a-c48765ffde1c@lucifer.local/ [3]
Link: https://lkml.kernel.org/r/20250719182854.3166724-1-surenb@google.com
Link: https://lkml.kernel.org/r/20250719182854.3166724-2-surenb@google.com
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Jeongjun Park <aha310510@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Josef Bacik <josef@toxicpanda.com>
Cc: Kalesh Singh <kaleshsingh@google.com>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: "Paul E . McKenney" <paulmck@kernel.org>
Cc: Peter Xu <peterx@redhat.com>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Thomas Weißschuh <linux@weissschuh.net>
Cc: T.J. Mercier <tjmercier@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Ye Bin <yebin10@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The non-tunneled tso6 test case was showing up as:
ok 8 tso.ipv4
This is because of the way test_builder() uses the inner_ipver arg in
test naming, and how test_info is iterated over in main(). Given that
some tunnels not supported yet, e.g. ipip or sit, only support ipv4 or
ipv6 as the inner network protocol, I think the best fix here is to
call test_builder() in separate branches for tunneled and non-tunneled
tests, and to make supported inner l3 types an explicit attribute of
tunnel test cases.
# Detected qstat for LSO wire-packets
TAP version 13
1..14
ok 1 tso.ipv4
# Testing with mangleid enabled
ok 2 tso.vxlan4_ipv4
ok 3 tso.vxlan4_ipv6
# Testing with mangleid enabled
ok 4 tso.vxlan_csum4_ipv4
ok 5 tso.vxlan_csum4_ipv6
# Testing with mangleid enabled
ok 6 tso.gre4_ipv4
ok 7 tso.gre4_ipv6
ok 8 tso.ipv6
# Testing with mangleid enabled
ok 9 tso.vxlan6_ipv4
ok 10 tso.vxlan6_ipv6
# Testing with mangleid enabled
ok 11 tso.vxlan_csum6_ipv4
ok 12 tso.vxlan_csum6_ipv6
# Testing with mangleid enabled
ok 13 tso.gre6_ipv4
ok 14 tso.gre6_ipv6
# Totals: pass:14 fail:0 xfail:0 xpass:0 skip:0 error:0
Fixes: 0d0f4174f6 ("selftests: drv-net: add a simple TSO test")
Signed-off-by: Daniel Zahka <daniel.zahka@gmail.com>
Link: https://patch.msgid.link/20250723184740.4075410-4-daniel.zahka@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When vxlan is used with ipv6 as the outer network header, the correct
ip link parameters for acheiving the SKB_GSO_UDP_TUNNEL gso type is
"udp6zerocsumtx udp6zerocsumrx". Otherwise the gso type will be
SKB_GSO_UDP_TUNNEL_CSUM.
This bug was the reason for the second of the three possible
invocations of run_one_stream() invocations, so that can be deleted as
well. We only need to test with the feature off and on.
Fixes: 0d0f4174f6 ("selftests: drv-net: add a simple TSO test")
Signed-off-by: Daniel Zahka <daniel.zahka@gmail.com>
Link: https://patch.msgid.link/20250723184740.4075410-3-daniel.zahka@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
tso.py uses the active features at the time of test execution
as the set of available gso features to test. This means if a gso
feature is supported but toggled off at test start, the test will be
skipped with a "Device does not support {feature}" message.
Instead, we can enumerate the set of toggleable features by capturing
the driver's hw_features bitmap. To avoid configuration side-effects
from running the test, we also snapshot the wanted_features flag set
before making any feature changes, and then attempt to restore the
same set of wanted_features before test exit.
Fixes: 0d0f4174f6 ("selftests: drv-net: add a simple TSO test")
Signed-off-by: Daniel Zahka <daniel.zahka@gmail.com>
Link: https://patch.msgid.link/20250723184740.4075410-2-daniel.zahka@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Make require_cmd() calls explicit about whether commands are needed
locally, remotely, or both.
Since require_cmd() defaults to local=True, tests should explicitly set
local=False when commands are only needed remotely.
- socat: Set local=False since it's only needed on remote hosts.
- iperf3: Use single call with both local=True and remote=True since
it's needed on both hosts.
This avoids unnecessary test failures when commands are missing locally
but available remotely where actually needed, and consolidates a
duplicate require_cmd() call into single call that checks both hosts.
Fixes: 0d0f4174f6 ("selftests: drv-net: add a simple TSO test")
Fixes: f1e68a1a4a ("selftests: drv-net: add require_XYZ() helpers for validating env")
Fixes: c76bab22e9 ("selftests: drv-net: rss_input_xfrm: Check test prerequisites before running")
Reviewed-by: Nimrod Oren <noren@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Link: https://patch.msgid.link/20250723135454.649342-3-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The require_cmd() method was checking for command availability locally
even when remote=True was specified, due to a missing host parameter.
Fix by passing host=self.remote when checking remote command
availability, ensuring commands are verified on the correct host.
Fixes: f1e68a1a4a ("selftests: drv-net: add require_XYZ() helpers for validating env")
Reviewed-by: Nimrod Oren <noren@nvidia.com>
Signed-off-by: Gal Pressman <gal@nvidia.com>
Link: https://patch.msgid.link/20250723135454.649342-2-gal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Instead of using '0' and '1' for napi threaded state use an enum with
'disabled' and 'enabled' states.
Tested:
./tools/testing/selftests/net/nl_netdev.py
TAP version 13
1..7
ok 1 nl_netdev.empty_check
ok 2 nl_netdev.lo_check
ok 3 nl_netdev.page_pool_check
ok 4 nl_netdev.napi_list_check
ok 5 nl_netdev.dev_set_threaded
ok 6 nl_netdev.napi_set_threaded
ok 7 nl_netdev.nsim_rxq_reset_down
# Totals: pass:7 fail:0 xfail:0 xpass:0 skip:0 error:0
Signed-off-by: Samiullah Khawaja <skhawaja@google.com>
Link: https://patch.msgid.link/20250723013031.2911384-4-skhawaja@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Martin KaFai Lau says:
====================
pull-request: bpf-next 2025-07-24
We've added 3 non-merge commits during the last 3 day(s) which contain
a total of 4 files changed, 40 insertions(+), 15 deletions(-).
The main changes are:
1) Improved verifier error message for incorrect narrower load from
pointer field in ctx, from Paul Chaignon.
2) Disabled migration in nf_hook_run_bpf to address a syzbot report,
from Kuniyuki Iwashima.
* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next:
selftests/bpf: Test invalid narrower ctx load
bpf: Reject narrower access to pointer ctx fields
bpf: Disable migration in nf_hook_run_bpf().
====================
Link: https://patch.msgid.link/20250724173306.3578483-1-martin.lau@linux.dev
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
For basic types we "flatten" setters. If a request "a" has a simple
nest "b" with value "val" we print helpers like:
req_set_a_b(struct a *req, int val)
{
req->_present.a = 1;
req->b._present.val = 1;
req->b.val = ...
}
This is not possible for multi-attr because they have to be allocated
dynamically by the user. Print "object level" setters so that user
preparing the object doesn't have to futz with the presence bits
and other YNL internals.
Add the ability to pass in the variable name to generated setters.
Using "req" here doesn't feel right, while the attr is part of a request
it's not the request itself, so it seems cleaner to call it "obj".
Example:
static inline void
netdev_queue_id_set_id(struct netdev_queue_id *obj, __u32 id)
{
obj->_present.id = 1;
obj->id = id;
}
Reviewed-by: Donald Hunter <donald.hunter@gmail.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250723171046.4027470-5-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In general YNL provides allocation and free helpers for types.
For pure nested structs which are used as multi-attr (and therefore
have to be allocated dynamically) we already print a free helper
as it's needed by free of the containing struct.
Add printing of the alloc helper for consistency. The helper
takes the number of entries to allocate as an argument, e.g.:
static inline struct netdev_queue_id *netdev_queue_id_alloc(unsigned int n)
{
return calloc(n, sizeof(struct netdev_queue_id));
}
Reviewed-by: Donald Hunter <donald.hunter@gmail.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250723171046.4027470-4-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The pidfd selftests run in userspace and include both userspace and kernel
header files. On some distros (for example, CentOS), this results in
duplicate-symbol warnings in allmodconfig builds, while on other distros
(for example, Ubuntu) it does not.
Therefore, use #undef to get rid of the userspace definitions in favor
of the kernel definitions.
Other ways of handling this include splitting up the selftest code so
that the userspace definitions go into one translation unit and the
kernel definitions into another (which might or might not be feasible)
or to adjust compiler command-line options to suppress the warnings
(which might or might not be desirable).
[ paulmck: Apply Shuah Khan feedback. ]
Link: https://lore.kernel.org/r/cc7e4fe7-299f-4bf3-af46-df6551d61997@paulmck-laptop
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Shuah Khan <skhan@linuxfoundation.org>
Cc: Christian Brauner <brauner@kernel.org>
Cc: <linux-kselftest@vger.kernel.org>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
The subsystem event test enables all "sched" events and makes sure there's
at least 3 different events in the output. It used to cat the entire trace
file to | wc -l, but on slow machines, that could last a very long time.
To solve that, it was changed to just read the first 100 lines of the
trace file. This can cause false failures as some events repeat so often,
that the 100 lines that are examined could possibly be of only one event.
Instead, create an awk script that looks for 3 different events and will
exit out after it finds them. This will find the 3 events the test looks
for (eventually if it works), and still exit out after the test is
satisfied and not cause slower machines to run forever.
Link: https://lore.kernel.org/r/20250721134212.53c3e140@batman.local.home
Reported-by: Tengda Wu <wutengda@huaweicloud.com>
Closes: https://lore.kernel.org/all/20250710130134.591066-1-wutengda@huaweicloud.com/
Fixes: 1a4ea83a6e ("selftests/ftrace: Limit length in subsystem-enable tests")
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Thresholds can increase the number of counters a metric needs. The NMI
watchdog can take away a counter (hopefully the buddy watchdog will
become the default and this will no longer be true). Add a new
constraint for the case that a metric and its thresholds would fit in
counters but only if the NMI watchdog isn't enabled. Either the
threshold or the NMI watchdog should be disabled to make the metric
fit. Wire this up into the metric__group_events logic.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250719030517.1990983-16-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Topdown metric events require grouping with a slots event. In perf
metrics this is currently achieved by metrics adding an unnecessary
"0 * tma_info_thread_slots". New TMA metrics trigger optimizations of
the metric expression that removes the event and breaks the metric due
to the missing but required event. Add a pass immediately before
sorting and fixing parsed events, that insert a slots event if one is
missing. Update test expectations to match this.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250719030517.1990983-15-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
The string comparisons were overly broad and could fire for the
incorrect PMU and events. Switch to using the config in the attribute
then add a perf test to confirm the attribute config values match
those of parsed events of that name and don't match others. This
exposed matches for slots events that shouldn't have matched as the
slots fixed counter event, such as topdown.slots_p.
Fixes: fbc798316b ("perf x86/topdown: Refine helper arch_is_topdown_metrics()")
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250719030517.1990983-14-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Counting events system-wide with a specified CPU prior to this change
worked:
```
$ perf stat -e 'msr/tsc/,msr/tsc,cpu=cpu_core/,msr/tsc,cpu=cpu_atom/' -a sleep 1
Performance counter stats for 'system wide':
59,393,419,099 msr/tsc/
33,927,965,927 msr/tsc,cpu=cpu_core/
25,465,608,044 msr/tsc,cpu=cpu_atom/
```
However, when counting with process the counts became system wide:
```
$ perf stat -e 'msr/tsc/,msr/tsc,cpu=cpu_core/,msr/tsc,cpu=cpu_atom/' perf test -F 10
10.1: Basic parsing test : Ok
10.2: Parsing without PMU name : Ok
10.3: Parsing with PMU name : Ok
Performance counter stats for 'perf test -F 10':
59,233,549 msr/tsc/
59,227,556 msr/tsc,cpu=cpu_core/
59,224,053 msr/tsc,cpu=cpu_atom/
```
Make the handling of CPU maps with event parsing clearer. When an
event is parsed creating an evsel the cpus should be either the PMU's
cpumask or user specified CPUs.
Update perf_evlist__propagate_maps so that it doesn't clobber the user
specified CPUs. Try to make the behavior clearer, firstly fix up
missing cpumasks. Next, perform sanity checks and adjustments from the
global evlist CPU requests and for the PMU including simplifying to
the "any CPU"(-1) value. Finally remove the event if the cpumask is
empty.
So that events are opened with a CPU and a thread change stat's
create_perf_stat_counter to give both.
With the change things are fixed:
```
$ perf stat --no-scale -e 'msr/tsc/,msr/tsc,cpu=cpu_core/,msr/tsc,cpu=cpu_atom/' perf test -F 10
10.1: Basic parsing test : Ok
10.2: Parsing without PMU name : Ok
10.3: Parsing with PMU name : Ok
Performance counter stats for 'perf test -F 10':
63,704,975 msr/tsc/
47,060,704 msr/tsc,cpu=cpu_core/ (4.62%)
16,640,591 msr/tsc,cpu=cpu_atom/ (2.18%)
```
However, note the "--no-scale" option is used. This is necessary as
the running time for the event on the counter isn't the same as the
enabled time because the thread doesn't necessarily run on the CPUs
specified for the counter. All counter values are scaled with:
scaled_value = value * time_enabled / time_running
and so without --no-scale the scaled_value becomes very large. This
problem already exists on hybrid systems for the same reason. Here are
2 runs of the same code with an instructions event that counts the
same on both types of core, there is no real multiplexing happening on
the event:
```
$ perf stat -e instructions perf test -F 10
...
Performance counter stats for 'perf test -F 10':
87,896,447 cpu_atom/instructions/ (14.37%)
98,171,964 cpu_core/instructions/ (85.63%)
...
$ perf stat --no-scale -e instructions perf test -F 10
...
Performance counter stats for 'perf test -F 10':
13,069,890 cpu_atom/instructions/ (19.32%)
83,460,274 cpu_core/instructions/ (80.68%)
...
```
The scaling has inflated per-PMU instruction counts and the overall
count by 2x.
To fix this the kernel needs changing when a task+CPU event (or just
task event on hybrid) is scheduled out. A fix could be that the state
isn't inactive but off for such events, so that time_enabled counts
don't accumulate on them.
Reviewed-by: Thomas Falcon <thomas.falcon@intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250719030517.1990983-13-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Add evsel__open_per_cpu_and_thread that combines the operation of
evsel__open_per_cpu and evsel__open_per_thread so that an event
without the "any" cpumask can be opened with its cpumask and with
threads it specifies. Change the implementation of evsel__open_per_cpu
and evsel__open_per_thread to use evsel__open_per_cpu_and_thread to
make the implementation of those functions clearer.
Reviewed-by: Thomas Falcon <thomas.falcon@intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: James Clark <james.clark@linaro.org>
Link: https://lore.kernel.org/r/20250719030517.1990983-12-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
For hybrid metrics it is useful to know the number of p-core or e-core
CPUs. If a cpumask is specified for the num_cpus or num_cpus_online
tool events, compute the value relative to the given mask rather than
for the full system.
```
$ sudo /tmp/perf/perf stat -e 'tool/num_cpus/,tool/num_cpus,cpu=cpu_core/,
tool/num_cpus,cpu=cpu_atom/,tool/num_cpus_online/,tool/num_cpus_online,
cpu=cpu_core/,tool/num_cpus_online,cpu=cpu_atom/' true
Performance counter stats for 'true':
28 tool/num_cpus/
16 tool/num_cpus,cpu=cpu_core/
12 tool/num_cpus,cpu=cpu_atom/
28 tool/num_cpus_online/
16 tool/num_cpus_online,cpu=cpu_core/
12 tool/num_cpus_online,cpu=cpu_atom/
0.000767205 seconds time elapsed
0.000938000 seconds user
0.000000000 seconds sys
```
Reviewed-by: Thomas Falcon <thomas.falcon@intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: James Clark <james.clark@linaro.org>
Link: https://lore.kernel.org/r/20250719030517.1990983-6-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
On hybrid systems, events like msr/tsc/ will aggregate counts across
all CPUs. Often metrics only want a value like msr/tsc/ for the cores
on which the metric is being computed. Listing each CPU with terms
cpu=0,cpu=1.. is laborious and would need to be encoded for all
variations of a CPU model.
Allow the cpumask from a PMU to be an argument to the cpu term. For
example in the following the cpumask of the cstate_pkg PMU selects the
CPUs to count msr/tsc/ counter upon:
```
$ cat /sys/bus/event_source/devices/cstate_pkg/cpumask
0
$ perf stat -A -e 'msr/tsc,cpu=cstate_pkg/' -a sleep 0.1
Performance counter stats for 'system wide':
CPU0 252,621,253 msr/tsc,cpu=cstate_pkg/
0.101184092 seconds time elapsed
```
As the cpu term is now also allowed to be a string, allow it to encode
a range of CPUs (a list can't be supported as ',' is already a special
token).
The "event qualifiers" section of the `perf list` man page is updated
to detail the additional behavior. The man page formatting is tidied
up in this section, as it was incorrectly appearing within the
"parameterized events" section.
Reviewed-by: Thomas Falcon <thomas.falcon@intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250719030517.1990983-5-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Factor requested CPU warning out of evlist and into evsel. At the end
of adding an event, perform the warning check. To avoid repeatedly
testing if the cpu_list is empty, add a local variable.
```
$ perf stat -e cpu_atom/cycles,cpu=1/ -a true
WARNING: A requested CPU in '1' is not supported by PMU 'cpu_atom' (CPUs 16-27) for event 'cpu_atom/cycles/'
Performance counter stats for 'system wide':
<not supported> cpu_atom/cycles/
0.000781511 seconds time elapsed
```
Reviewed-by: Thomas Falcon <thomas.falcon@intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: James Clark <james.clark@linaro.org>
Link: https://lore.kernel.org/r/20250719030517.1990983-2-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Force loading all PMUs adds significant cost because DRM and other
PMUs are loaded, it should also not be required if the pmus__
functions are used.
Tested by run perf test, in particular the pfm related tests. Also
`perf list` is identical before and after.
Before:
$ time ./perf test pfm
54: Test libpfm4 support :
54.1: test of individual --pfm-events : Ok
54.2: test groups of --pfm-events : Ok
103: perf all libpfm4 events test : Ok
real 0m8.933s
user 0m1.824s
sys 0m7.122s
After:
$ time ./perf test pfm
54: Test libpfm4 support :
54.1: test of individual --pfm-events : Ok
54.2: test groups of --pfm-events : Ok
103: perf all libpfm4 events test : Ok
real 0m5.259s
user 0m1.793s
sys 0m3.570s
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20250722013449.146233-1-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Pull networking fixes from Paolo Abeni:
"Including fixes from can and xfrm.
The TI regression notified last week is actually on our net-next tree,
it does not affect 6.16.
We are investigating a virtio regression which is quite hard to
reproduce - currently only our CI sporadically hits it. Hopefully it
should not be critical, and I'm not sure that an additional week would
be enough to solve it.
Current release - fix to a fix:
- sched: sch_qfq: avoid sleeping in atomic context in qfq_delete_class
Previous releases - regressions:
- xfrm:
- set transport header to fix UDP GRO handling
- delete x->tunnel as we delete x
- eth:
- mlx5: fix memory leak in cmd_exec()
- i40e: when removing VF MAC filters, avoid losing PF-set MAC
- gve: fix stuck TX queue for DQ queue format
Previous releases - always broken:
- can: fix NULL pointer deref of struct can_priv::do_set_mode
- eth:
- ice: fix a null pointer dereference in ice_copy_and_init_pkg()
- ism: fix concurrency management in ism_cmd()
- dpaa2: fix device reference count leak in MAC endpoint handling
- icssg-prueth: fix buffer allocation for ICSSG
Misc:
- selftests: mptcp: increase code coverage"
* tag 'net-6.16-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (34 commits)
net: hns3: default enable tx bounce buffer when smmu enabled
net: hns3: fixed vf get max channels bug
net: hns3: disable interrupt when ptp init failed
net: hns3: fix concurrent setting vlan filter issue
s390/ism: fix concurrency management in ism_cmd()
selftests: drv-net: wait for iperf client to stop sending
MAINTAINERS: Add in6.h to MAINTAINERS
selftests: netfilter: tone-down conntrack clash test
can: netlink: can_changelink(): fix NULL pointer deref of struct can_priv::do_set_mode
net/sched: sch_qfq: Avoid triggering might_sleep in atomic context in qfq_delete_class
gve: Fix stuck TX queue for DQ queue format
net: appletalk: Fix use-after-free in AARP proxy probe
net: bcmasp: Restore programming of TX map vector register
selftests: mptcp: connect: also cover checksum
selftests: mptcp: connect: also cover alt modes
e1000e: ignore uninitialized checksum word on tgp
e1000e: disregard NVM checksum on tgp when valid checksum bit is not set
ice: Fix a null pointer dereference in ice_copy_and_init_pkg()
i40e: When removing VF MAC filters, only check PF-set MAC
i40e: report VF tx_dropped with tx_errors instead of tx_discards
...
* for-next/feat_mte_store_only:
: MTE feature to restrict tag checking to store only operations
kselftest/arm64/mte: Add MTE_STORE_ONLY testcases
kselftest/arm64/mte: Preparation for mte store only test
kselftest/arm64/abi: Add MTE_STORE_ONLY feature hwcap test
KVM: arm64: Expose MTE_STORE_ONLY feature to guest
arm64/hwcaps: Add MTE_STORE_ONLY hwcaps
arm64/kernel: Support store-only mte tag check
prctl: Introduce PR_MTE_STORE_ONLY
arm64/cpufeature: Add MTE_STORE_ONLY feature
* arm64/for-next/perf: (23 commits)
drivers/perf: hisi: Support PMUs with no interrupt
drivers/perf: hisi: Relax the event number check of v2 PMUs
drivers/perf: hisi: Add support for HiSilicon SLLC v3 PMU driver
drivers/perf: hisi: Use ACPI driver_data to retrieve SLLC PMU information
drivers/perf: hisi: Add support for HiSilicon DDRC v3 PMU driver
drivers/perf: hisi: Simplify the probe process for each DDRC version
perf/arm-ni: Support sharing IRQs within an NI instance
perf/arm-ni: Consolidate CPU affinity handling
perf/cxlpmu: Fix typos in cxl_pmu.c comments and documentation
perf/cxlpmu: Remove unintended newline from IRQ name format string
perf/cxlpmu: Fix devm_kcalloc() argument order in cxl_pmu_probe()
perf: arm_spe: Relax period restriction
perf: arm_pmuv3: Add support for the Branch Record Buffer Extension (BRBE)
KVM: arm64: nvhe: Disable branch generation in nVHE guests
arm64: Handle BRBE booting requirements
arm64/sysreg: Add BRBE registers and fields
perf/arm: Add missing .suppress_bind_attrs
perf/arm-cmn: Reduce stack usage during discovery
perf: imx9_perf: make the read-only array mask static const
perf/arm-cmn: Broaden module description for wider interconnect support
...
* for-next/livepatch:
: Support for HAVE_LIVEPATCH on arm64
arm64: Kconfig: Keep selects somewhat alphabetically ordered
arm64: Implement HAVE_LIVEPATCH
arm64: stacktrace: Implement arch_stack_walk_reliable()
arm64: stacktrace: Check kretprobe_find_ret_addr() return value
arm64/module: Use text-poke API for late relocations.
* for-next/user-contig-bbml2:
: Optimise the TLBI when folding/unfolding contigous PTEs on hardware with BBML2 and no TLB conflict aborts
arm64/mm: Elide tlbi in contpte_convert() under BBML2
iommu/arm: Add BBM Level 2 smmu feature
arm64: Add BBM Level 2 cpu feature
arm64: cpufeature: Introduce MATCH_ALL_EARLY_CPUS capability type
* for-next/misc:
: Miscellaneous arm64 patches
arm64/gcs: task_gcs_el0_enable() should use passed task
arm64: signal: Remove ISB when resetting POR_EL0
arm64/mm: Drop redundant addr increment in set_huge_pte_at()
arm64: Mark kernel as tainted on SAE and SError panic
arm64/gcs: Don't call gcs_free() when releasing task_struct
arm64: fix unnecessary rebuilding when CONFIG_DEBUG_EFI=y
arm64/mm: Optimize loop to reduce redundant operations of contpte_ptep_get
arm64: pi: use 'targets' instead of extra-y in Makefile
* for-next/acpi:
: Various ACPI arm64 changes
ACPI: Suppress misleading SPCR console message when SPCR table is absent
ACPI: Return -ENODEV from acpi_parse_spcr() when SPCR support is disabled
* for-next/debug-entry:
: Simplify the debug exception entry path
arm64: debug: remove debug exception registration infrastructure
arm64: debug: split bkpt32 exception entry
arm64: debug: split brk64 exception entry
arm64: debug: split hardware watchpoint exception entry
arm64: debug: split single stepping exception entry
arm64: debug: refactor reinstall_suspended_bps()
arm64: debug: split hardware breakpoint exception entry
arm64: entry: Add entry and exit functions for debug exceptions
arm64: debug: remove break/step handler registration infrastructure
arm64: debug: call step handlers statically
arm64: debug: call software breakpoint handlers statically
arm64: refactor aarch32_break_handler()
arm64: debug: clean up single_step_handler logic
* for-next/feat_mte_tagged_far:
: Support for reporting the non-address bits during a synchronous MTE tag check fault
kselftest/arm64/mte: Add mtefar tests on check_mmap_options
kselftest/arm64/mte: Refactor check_mmap_option test
kselftest/arm64/mte: Add verification for address tag in signal handler
kselftest/arm64/mte: Add address tag related macro and function
kselftest/arm64/mte: Check MTE_FAR feature is supported
kselftest/arm64/mte: Register mte signal handler with SA_EXPOSE_TAGBITS
kselftest/arm64: Add MTE_FAR hwcap test
KVM: arm64: Expose FEAT_MTE_TAGGED_FAR feature to guest
arm64: Report address tag when FEAT_MTE_TAGGED_FAR is supported
arm64/cpufeature: Add FEAT_MTE_TAGGED_FAR feature
* for-next/kselftest:
: Kselftest updates for arm64
kselftest/arm64: Handle attempts to disable SM on SME only systems
kselftest/arm64: Fix SVE write data generation for SME only systems
kselftest/arm64: Test SME on SME only systems in fp-ptrace
kselftest/arm64: Test FPSIMD format data writes via NT_ARM_SVE in fp-ptrace
kselftest/arm64: Allow sve-ptrace to run on SME only systems
kselftest/arm4: Provide local defines for AT_HWCAP3
kselftest/arm64: Specify SVE data when testing VL set in sve-ptrace
kselftest/arm64: Fix test for streaming FPSIMD write in sve-ptrace
kselftest/arm64: Fix check for setting new VLs in sve-ptrace
kselftest/arm64: Convert tpidr2 test to use kselftest.h
* for-next/mdscr-cleanup:
: Drop redundant DBG_MDSCR_* macros
KVM: selftests: Change MDSCR_EL1 register holding variables as uint64_t
arm64/debug: Drop redundant DBG_MDSCR_* macros
* for-next/vmap-stack:
: Force VMAP_STACK on arm64
arm64: remove CONFIG_VMAP_STACK checks from entry code
arm64: remove CONFIG_VMAP_STACK checks from SDEI stack handling
arm64: remove CONFIG_VMAP_STACK checks from stacktrace overflow logic
arm64: remove CONFIG_VMAP_STACK conditionals from traps overflow stack
arm64: remove CONFIG_VMAP_STACK conditionals from irq stack setup
arm64: Remove CONFIG_VMAP_STACK conditionals from THREAD_SHIFT and THREAD_ALIGN
arm64: efi: Remove CONFIG_VMAP_STACK check
arm64: Mandate VMAP_STACK
arm64: efi: Fix KASAN false positive for EFI runtime stack
arm64/ptrace: Fix stack-out-of-bounds read in regs_get_kernel_stack_nth()
arm64/gcs: Don't call gcs_free() during flush_gcs()
arm64: Restrict pagetable teardown to avoid false warning
docs: arm64: Fix ICC_SRE_EL2 register typo in booting.rst
The current behaviour of rvgen when running with the -a option is to
append the necessary lines at the end of the configuration for Kconfig,
Makefile and tracepoints.
This is not always the desired behaviour in case of nested monitors:
while tracepoints are not affected by nesting and the Makefile's only
requirement is that the parent monitor is built before its children, in
the Kconfig it is better to have children defined right after their
parent, otherwise the result has wrong indentation:
[*] foo_parent monitor
[*] foo_child1 monitor
[*] foo_child2 monitor
[*] bar_parent monitor
[*] bar_child1 monitor
[*] bar_child2 monitor
[*] foo_child3 monitor
[*] foo_child4 monitor
Adapt rvgen to look for a different marker for nested monitors in the
Kconfig file and append the line right after the last sibling, instead
of the last monitor.
Also add the marker when creating a new parent monitor.
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Tomas Glozar <tglozar@redhat.com>
Cc: Juri Lelli <jlelli@redhat.com>
Cc: Clark Williams <williams@redhat.com>
Cc: John Kacur <jkacur@redhat.com>
Link: https://lore.kernel.org/20250723161240.194860-5-gmonaco@redhat.com
Reviewed-by: Nam Cao <namcao@linutronix.de>
Signed-off-by: Gabriele Monaco <gmonaco@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Currently the userspace RV tool starts a monitor and waits for the user
to press Ctrl-C (SIGINT) to terminate and stop the monitor.
This doesn't account for a scenario where a user starts RV in background
and simply kills it (SIGTERM unless the user specifies differently).
E.g.:
# rv mon wip &
# kill %
Would terminate RV without stopping the monitor and next RV executions
won't start correctly.
Register the signal handler used for SIGINT also to SIGTERM.
Cc: Nam Cao <namcao@linutronix.de>
Cc: Tomas Glozar <tglozar@redhat.com>
Cc: Juri Lelli <jlelli@redhat.com>
Cc: Clark Williams <williams@redhat.com>
Cc: John Kacur <jkacur@redhat.com>
Link: https://lore.kernel.org/20250723161240.194860-3-gmonaco@redhat.com
Signed-off-by: Gabriele Monaco <gmonaco@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Both container generation and DA monitor generation is implemented in the
class dot2k. That requires some ugly "if is_container ... else ...". If
linear temporal logic support is added at the current state, the "if else"
chain is longer and uglier.
Furthermore, container generation is irrevelant to .dot files. It is
therefore illogical to be implemented in class "dot2k".
Clean it up, restructure the dot2k class into the following class
hierarchy:
(RVGenerator)
/\
/ \
/ \
/ \
/ \
(Container) (Monitor)
/\
/ \
/ \
/ \
(dot2k) [ltl2k] <- intended
This allows a simple and clean integration of LTL.
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://lore.kernel.org/692137a581ba6bee7a64d37fb7173ae137c47bbd.1751634289.git.namcao@linutronix.de
Reviewed-by: Gabriele Monaco <gmonaco@redhat.com>
Signed-off-by: Nam Cao <namcao@linutronix.de>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
To simply the scripts and to allow easy integration of new monitor types,
restructure the template files as followed:
1. Move the template files to be in the same directory as the rvgen
package. Furthermore, the installation will now only install the
templates to the package directory, not /usr/share/. This simplify
templates reading, as the scripts do not need to find the templates at
multiple places.
2. Move dot2k_templates/* to:
- templates/dot2k/
- templates/container/
This allows sharing templates reading code between DA monitor generation
and container generation (and any future generation type).
For template files which can be shared between different generation
types, support putting them in templates/
This restructure aligns with the recommendation from:
https://python-packaging.readthedocs.io/en/latest/non-code-files.html
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://lore.kernel.org/462d90273f96804d3ba850474877d5f727031258.1751634289.git.namcao@linutronix.de
Reviewed-by: Gabriele Monaco <gmonaco@redhat.com>
Signed-off-by: Nam Cao <namcao@linutronix.de>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
