| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Each test currently has a "run test" function in per test file and
another resctrl_tests.c. The functions in resctrl_tests.c are almost
identical.
Generalize the one in resctrl_tests.c such that it can be shared
between all of the tests. It makes adding new tests easier and removes
the per test if () forests.
Also add comment to CPU vendor IDs that they must be defined as bits
for a bitmask.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
resctrl_tests reads a set of parameters and passes them individually
for each tests which causes variations in the call signature between
the tests.
Add struct input_params to hold all input parameters. It can be easily
passed to every test without varying the call signature.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
CAT test does not reset the CPU affinity after the benchmark.
This is relatively harmless as is because CAT test is the last
benchmark to run, however, more tests may be added later.
Store the CPU affinity the first time taskset_benchmark() is run and
add taskset_restore() which the test can call to reset the CPU mask to
its original value.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
CAT test spawns two processes into two different control groups with
exclusive schemata. Both the processes alloc a buffer from memory
matching their allocated LLC block size and flush the entire buffer out
of caches. Since the processes are reading through the buffer only once
during the measurement and initially all the buffer was flushed, the
test isn't testing CAT.
Rewrite the CAT test to allocate a buffer sized to half of LLC. Then
perform a sequence of tests with different LLC alloc sizes starting
from half of the CBM bits down to 1-bit CBM. Flush the buffer before
each test and read the buffer twice. Observe the LLC misses on the
second read through the buffer. As the allocated LLC block gets smaller
and smaller, the LLC misses will become larger and larger giving a
strong signal on CAT working properly.
The new CAT test is using only a single process because it relies on
measured effect against another run of itself rather than another
process adding noise. The rest of the system is set to use the CBM bits
not used by the CAT test to keep the test isolated.
Replace count_bits() with count_contiguous_bits() to get the first bit
position in order to be able to calculate masks based on it.
This change has been tested with a number of systems from different
generations.
Suggested-by: Reinette Chatre <[email protected]>
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
When reading memory in order, HW prefetching optimizations will
interfere with measuring how caches and memory are being accessed. This
adds noise into the results.
Change the fill_buf reading loop to not use an obvious in-order access
using multiply by a prime and modulo.
Using a prime multiplier with modulo ensures the entire buffer is
eventually read. 23 is small enough that the reads are spread out but
wrapping does not occur very frequently (wrapping too often can trigger
L2 hits more frequently which causes noise to the test because getting
the data from LLC is not required).
It was discovered that not all primes work equally well and some can
cause wildly unstable results (e.g., in an earlier version of this
patch, the reads were done in reversed order and 59 was used as the
prime resulting in unacceptably high and unstable results in MBA and
MBM test on some architectures).
Link: https://lore.kernel.org/linux-kselftest/TYAPR01MB6330025B5E6537F94DA49ACB8B499@TYAPR01MB6330.jpnprd01.prod.outlook.com/
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
The fill_buf code prevents compiler optimizating the entire read loop
away by writing the final value of the variable into a file. While it
achieves the goal, writing into a file requires significant amount of
work within the innermost test loop and also error handling.
A simpler approach is to take advantage of volatile. Writing through
a pointer to a volatile variable is enough to prevent compiler from
optimizing the write away, and therefore compiler cannot remove the
read loop either.
Add a volatile 'value_sink' into resctrl_tests.c and make fill_buf to
write into it. As a result, the error handling in fill_buf.c can be
simplified.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
Perf fd (pe_fd) is opened, reset, and enabled during every test the CAT
selftest runs. Also, ioctl(pe_fd, ...) calls are not error checked even
if ioctl() could return an error.
Open perf fd only once before the tests and only reset and enable the
counter within the test loop. Add error checking to pe_fd ioctl()
calls.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
The main CAT test function is called cat_val() and resides in cache.c
which is illogical.
Rename the function to cat_test() and move it into cat_test.c.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
Perf related variables pea_llc_miss, pe_read, and pe_fd are globals in
cache.c.
Convert them to locals for better scoping and make pea_llc_miss simpler
by renaming it to pea. Make close(pe_fd) handling easier to understand
by doing it inside cat_val().
Make also sizeof()s use safer way to determine the right struct.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
struct perf_event_attr initialization is spread into
perf_event_initialize() and perf_event_attr_initialize() and setting
->config is hardcoded by the deepest level.
perf_event_attr init belongs to perf_event_attr_initialize() so move it
entirely there. Rename the other function
perf_event_initialized_read_format().
Call each init function directly from the test as they will take
different parameters (especially true after the perf related global
variables are moved to local variables).
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
Naming for perf event related functions, types, and variables is
inconsistent.
Make struct read_format and all functions related to perf events start
with "perf_". Adjust variable names towards the same direction but use
shorter names for variables where appropriate (pe prefix).
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
Perf event handling has functions that are the sole caller of another
perf event handling related function:
- reset_enable_llc_perf() calls perf_event_open_llc_miss()
- perf_event_measure() calls get_llc_perf()
Remove the extra layer of calls to make the code easier to follow by
moving the code into the calling function.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
Perf counters are __u64 but the code converts them to unsigned long
before printing them out.
Remove unnecessary type conversion and retain the perf originating
value as __u64.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
show_cache_info() calculates results and provides generic cache
information. This makes it hard to alter pass/fail conditions.
Separate the test specific checks into CAT and CMT test files and
leave only the generic information part into show_cache_info().
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
measure_cache_vals() does a different thing depending on the test case
that called it:
- For CAT, it measures LLC misses through perf.
- For CMT, it measures LLC occupancy through resctrl.
Split these two functionalities into own functions the CAT and CMT
tests can call directly. Replace passing the struct resctrl_val_param
parameter with the filename because it's more generic and all those
functions need out of resctrl_val.
Co-developed-by: Fenghua Yu <[email protected]>
Signed-off-by: Fenghua Yu <[email protected]>
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
CAT test doesn't take shareable bits into account, i.e., the test might
be sharing cache with some devices (e.g., graphics).
Introduce get_mask_no_shareable() and use it to provision an
environment for CAT test where the allocated LLC is isolated better.
Excluding shareable_bits may create hole(s) into the cbm_mask, thus add
a new helper count_contiguous_bits() to find the longest contiguous set
of CBM bits.
create_bit_mask() is needed by an upcoming CAT test rewrite so make it
available in resctrl.h right away.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
CAT and CMT tests calculate size of the cache portion for the n-bits
cache allocation on their own.
Add cache_portion_size() helper that calculates size of the cache
portion for the given number of bits and use it to replace the existing
span calculations. This also prepares for the new CAT test that will
need to determine the size of the cache portion also during results
processing.
Rename also 'cache_size' local variables to 'cache_total_size' to
prevent misinterpretations.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
get_cache_size() does not modify cache_type so it could be const.
Mark cache_type const so that const char * can be passed to it. This
prevents warnings once many of the test parameters are marked const.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
Callers of get_cbm_mask() are required to pass a string into which the
capacity bitmask (CBM) is read. Neither CAT nor CMT tests need the
bitmask as string but just convert it into an unsigned long value.
Another limitation is that the bit mask reader can only read
.../cbm_mask files.
Generalize the bit mask reading function into get_bit_mask() such that
it can be used to handle other files besides the .../cbm_mask and
handles the unsigned long conversion within get_bit_mask() using
fscanf(). Change get_cbm_mask() to use get_bit_mask() and rename it to
get_full_cbm() to better indicate what the function does.
Return error from get_full_cbm() if the bitmask is zero for some reason
because it makes the code more robust as the selftests naturally assume
the bitmask has some bits.
Also mark cache_type const while at it and remove useless comments that
are related to processing of CBM bits.
Co-developed-by: Fenghua Yu <[email protected]>
Signed-off-by: Fenghua Yu <[email protected]>
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
There are unnecessary nested calls in fill_buf.c:
- run_fill_buf() calls fill_cache()
- alloc_buffer() calls malloc_and_init_memory()
Simplify the code flow and remove those unnecessary call levels by
moving the called code inside the calling function and remove the
duplicated error print.
Resolve the difference in run_fill_buf() and fill_cache() parameter
name into 'buf_size' which is more descriptive than 'span'. Also, while
moving the allocation related code, rename 'p' into 'buf' to be
consistent in naming the variables.
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
MBM, MBA and CMT test cases call run_fill_buf() that in turn calls
fill_cache() to alloc and loop indefinitely around the buffer. This
binds buffer allocation and running the benchmark into a single bundle
so that a selftest cannot allocate a buffer once and reuse it. CAT test
doesn't want to loop around the buffer continuously and after rewrite
it needs the ability to allocate the buffer separately.
Split buffer allocation out of fill_cache() into alloc_buffer(). This
change is part of preparation for the new CAT test that allocates a
buffer and does multiple passes over the same buffer (but not in an
infinite loop).
Co-developed-by: Fenghua Yu <[email protected]>
Signed-off-by: Fenghua Yu <[email protected]>
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
A number function comments state the function return non-zero on
failure but in reality they can only return 0 on success and < 0 on
error.
Update the comments to say < 0 on error to match the behavior.
While at it, improve cat_val() comment to state that 0 means the test
was run (either pass or fail).
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
perf_event_open_llc_miss() calls ctrlc_handler() to cleanup if
perf_event_open() returns an error. Those cleanups, however, are not
the responsibility of perf_event_open_llc_miss() and it thus interferes
unnecessarily with the usual cleanup pattern. Worse yet,
ctrlc_handler() calls exit() in the end preventing the ordinary cleanup
done in the calling function from executing.
ctrlc_handler() should only be used as a signal handler, not during
normal error handling.
Remove call to ctrlc_handler() from perf_event_open_llc_miss(). As
unmounting resctrlfs and test cleanup are already handled properly
by error rollbacks in the calling functions, no other changes are
necessary.
Suggested-by: Reinette Chatre <[email protected]>
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
A number of functions in the resctrl selftests return errno. It is
problematic because errno is positive which is often counterintuitive.
Also, every site returning errno prints the error message already with
ksft_perror() so there is not much added value in returning the precise
error code.
Simply convert all places returning errno to return -1 that is typical
userspace error code in case of failures.
While at it, improve resctrl_val() comment to state that 0 means the
test was run (either pass or fail).
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
The resctrl selftest code contains a number of perror() calls. Some of
them come with hash character and some don't. The kselftest framework
provides ksft_perror() that is compatible with test output formatting
so it should be used instead of adding custom hash signs.
Some perror() calls are too far away from anything that sets error.
For those call sites, ksft_print_msg() must be used instead.
Convert perror() to ksft_perror() or ksft_print_msg().
Other related changes:
- Remove hash signs
- Remove trailing stops & newlines from ksft_perror()
- Add terminating newlines for converted ksft_print_msg()
- Use consistent capitalization
- Small fixes/tweaks to typos & grammar of the messages
- Extract error printing out of PARENT_EXIT() to be able to
differentiate
Signed-off-by: Ilpo Järvinen <[email protected]>
Reviewed-by: Reinette Chatre <[email protected]>
Signed-off-by: Shuah Khan <[email protected]>
|
|
[Changes from V1:
- Avoid conflict by rebasing with latest master.]
Some BPF tests use loop unrolling compiler pragmas that are clang
specific and not supported by GCC. These pragmas, along with their
GCC equivalences are:
#pragma clang loop unroll_count(N)
#pragma GCC unroll N
#pragma clang loop unroll(full)
#pragma GCC unroll 65534
#pragma clang loop unroll(disable)
#pragma GCC unroll 1
#pragma unroll [aka #pragma clang loop unroll(enable)]
There is no GCC equivalence to this pragma. It enables unrolling on
loops that the compiler would not ordinarily unroll even with
-O2|-funroll-loops, but it is not equivalent to full unrolling
either.
This patch adds a new header progs/bpf_compiler.h that defines the
following macros, which correspond to each pair of compiler-specific
pragmas above:
__pragma_loop_unroll_count(N)
__pragma_loop_unroll_full
__pragma_loop_no_unroll
__pragma_loop_unroll
The selftests using loop unrolling pragmas are then changed to include
the header and use these macros in place of the explicit pragmas.
Tested in bpf-next master.
No regressions.
Signed-off-by: Jose E. Marchesi <[email protected]>
Signed-off-by: Andrii Nakryiko <[email protected]>
Acked-by: Yonghong Song <[email protected]>
Link: https://lore.kernel.org/bpf/[email protected]
|
|
Add two tests to ensure fentry programs cannot attach to
bpf_spin_{lock,unlock}() helpers. The tracing_failure.c files
can be used in the future for other tracing failure cases.
Signed-off-by: Yonghong Song <[email protected]>
Signed-off-by: Andrii Nakryiko <[email protected]>
Link: https://lore.kernel.org/bpf/[email protected]
|
|
The netdev CI is reporting failures for the pmtu test:
[ 115.929264] br0: port 2(vxlan_a) entered forwarding state
# 2024/02/08 17:33:22 socat[7871] E bind(7, {AF=10 [0000:0000:0000:0000:0000:0000:0000:0000]:50000}, 28): Address already in use
# 2024/02/08 17:33:22 socat[7877] E write(7, 0x5598fb6ff000, 8192): Connection refused
# TEST: IPv6, bridged vxlan4: PMTU exceptions [FAIL]
# File size 0 mismatches exepcted value in locally bridged vxlan test
The root cause is apparently a socket created by a previous iteration
of the relevant loop still lasting in LAST_ACK state.
Note that even the file size check is racy, the receiver process dumping
the file could still be running in background
Allow the listener to bound on the same local port via SO_REUSEADDR and
collect file output file size only after the listener completion.
Fixes: 136a1b434bbb ("selftests: net: test vxlan pmtu exceptions with tcp")
Signed-off-by: Paolo Abeni <[email protected]>
Link: https://lore.kernel.org/r/4f51c11a1ce7ca7a4dabd926cffff63dadac9ba1.1707731086.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
The helper waiting for a listener port can match any socket whose
hexadecimal representation of source or destination addresses
matches that of the given port.
Additionally, any socket state is accepted.
All the above can let the helper return successfully before the
relevant listener is actually ready, with unexpected results.
So far I could not find any related failure in the netdev CI, but
the next patch is going to make the critical event more easily
reproducible.
Address the issue matching the port hex only vs the relevant socket
field and additionally checking the socket state for TCP sockets.
Fixes: 3bdd9fd29cb0 ("selftests/net: synchronize udpgro tests' tx and rx connection")
Signed-off-by: Paolo Abeni <[email protected]>
Link: https://lore.kernel.org/r/192b3dbc443d953be32991d1b0ca432bd4c65008.1707731086.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
The mentioned test is failing in slow environments:
# SO_TXTIME ipv4 clock monotonic
# ./so_txtime: recv: timeout: Resource temporarily unavailable
not ok 1 selftests: net: so_txtime.sh # exit=1
Tuning the tolerance in the test binary is error-prone and doomed
to failures is slow-enough environment.
Just resort to suppress any error in such cases. Note to suppress
them we need first to refactor a bit the code moving it to explicit
error handling.
Fixes: af5136f95045 ("selftests/net: SO_TXTIME with ETF and FQ")
Signed-off-by: Paolo Abeni <[email protected]>
Link: https://lore.kernel.org/r/2142d9ed4b5c5aa07dd1b455779625d91b175373.1707730902.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
The gro self-tests sends the packets to be aggregated with
multiple write operations.
When running is slow environment, it's hard to guarantee that
the GRO engine will wait for the last packet in an intended
train.
The above causes almost deterministic failures in our CI for
the 'large' test-case.
Address the issue explicitly ignoring failures for such case
in slow environments (KSFT_MACHINE_SLOW==true).
Fixes: 7d1575014a63 ("selftests/net: GRO coalesce test")
Reviewed-by: Willem de Bruijn <[email protected]>
Signed-off-by: Paolo Abeni <[email protected]>
Link: https://lore.kernel.org/r/97d3ba83f5a2bfeb36f6bc0fb76724eb3dafb608.1707729403.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
Zenghui noted that the test assertion for the ISTATUS bit is printing
the current timer value instead of the control register in the case of
failure. While the assertion is sound, printing CNT isn't informative.
Change things around to actually print the CTL register value instead.
Reported-by: Zenghui Yu <[email protected]>
Closes: https://lore.kernel.org/kvmarm/[email protected]/
Reviewed-by: Zenghui Yu <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Oliver Upton <[email protected]>
|
|
Older glibc's netinet/in.h may leave IPPROTO_MPTCP undefined when
building ip_local_port_range.c, that leads to "error: use of undeclared
identifier 'IPPROTO_MPTCP'".
Define IPPROTO_MPTCP in such cases, just like in other MPTCP selftests.
Fixes: 122db5e3634b ("selftests/net: add MPTCP coverage for IP_LOCAL_PORT_RANGE")
Reported-by: Linux Kernel Functional Testing <[email protected]>
Closes: https://lore.kernel.org/netdev/CA+G9fYvGO5q4o_Td_kyQgYieXWKw6ktMa-Q0sBu6S-0y3w2aEQ@mail.gmail.com/
Signed-off-by: Maxim Galaganov <[email protected]>
Tested-by: Linux Kernel Functional Testing <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
Fix a pile of -Wformat warnings in the KVM ARM selftests code, almost all
of which are benign "long" versus "long long" issues (selftests are 64-bit
only, and the guest printf code treats "ll" the same as "l"). The code
itself isn't problematic, but the warnings make it impossible to build ARM
selftests with -Werror, which does detect real issues from time to time.
Opportunistically have GUEST_ASSERT_BITMAP_REG() interpret set_expected,
which is a bool, as an unsigned decimal value, i.e. have it print '0' or
'1' instead of '0x0' or '0x1'.
Signed-off-by: Sean Christopherson <[email protected]>
Tested-by: Zenghui Yu <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Oliver Upton <[email protected]>
|
|
Add tests of changing permanent routes to temporary routes and the reversed
case to make sure GC working correctly in these cases. Add tests for the
temporary routes from RA.
The existing device will be deleted between tests to remove all routes
associated with it, so that the earlier tests don't mess up the later ones.
Reviewed-by: Hangbin Liu <[email protected]>
Reviewed-by: David Ahern <[email protected]>
Tested-by: Hangbin Liu <[email protected]>
Signed-off-by: Kui-Feng Lee <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
This test is time sensitive. It may fail on virtual machines and for
debug builds.
Similar to commit c41dfb0dfbec ("selftests/net: ignore timing errors in
so_txtime if KSFT_MACHINE_SLOW"), optionally suppress failure for timing
errors (only).
Signed-off-by: Paolo Abeni <[email protected]>
Reviewed-by: Willem de Bruijn <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
In various performance profiles of kernels with BPF programs attached,
bpf_local_storage_lookup() appears as a significant portion of CPU
cycles spent. To enable the compiler generate more optimal code, turn
bpf_local_storage_lookup() into a static inline function, where only the
cache insertion code path is outlined
Notably, outlining cache insertion helps avoid bloating callers by
duplicating setting up calls to raw_spin_{lock,unlock}_irqsave() (on
architectures which do not inline spin_lock/unlock, such as x86), which
would cause the compiler produce worse code by deciding to outline
otherwise inlinable functions. The call overhead is neutral, because we
make 2 calls either way: either calling raw_spin_lock_irqsave() and
raw_spin_unlock_irqsave(); or call __bpf_local_storage_insert_cache(),
which calls raw_spin_lock_irqsave(), followed by a tail-call to
raw_spin_unlock_irqsave() where the compiler can perform TCO and (in
optimized uninstrumented builds) turns it into a plain jump. The call to
__bpf_local_storage_insert_cache() can be elided entirely if
cacheit_lockit is a false constant expression.
Based on results from './benchs/run_bench_local_storage.sh' (21 trials,
reboot between each trial; x86 defconfig + BPF, clang 16) this produces
improvements in throughput and latency in the majority of cases, with an
average (geomean) improvement of 8%:
+---- Hashmap Control --------------------
|
| + num keys: 10
| : <before> | <after>
| +-+ hashmap (control) sequential get +----------------------+----------------------
| +- hits throughput | 14.789 M ops/s | 14.745 M ops/s ( ~ )
| +- hits latency | 67.679 ns/op | 67.879 ns/op ( ~ )
| +- important_hits throughput | 14.789 M ops/s | 14.745 M ops/s ( ~ )
|
| + num keys: 1000
| : <before> | <after>
| +-+ hashmap (control) sequential get +----------------------+----------------------
| +- hits throughput | 12.233 M ops/s | 12.170 M ops/s ( ~ )
| +- hits latency | 81.754 ns/op | 82.185 ns/op ( ~ )
| +- important_hits throughput | 12.233 M ops/s | 12.170 M ops/s ( ~ )
|
| + num keys: 10000
| : <before> | <after>
| +-+ hashmap (control) sequential get +----------------------+----------------------
| +- hits throughput | 7.220 M ops/s | 7.204 M ops/s ( ~ )
| +- hits latency | 138.522 ns/op | 138.842 ns/op ( ~ )
| +- important_hits throughput | 7.220 M ops/s | 7.204 M ops/s ( ~ )
|
| + num keys: 100000
| : <before> | <after>
| +-+ hashmap (control) sequential get +----------------------+----------------------
| +- hits throughput | 5.061 M ops/s | 5.165 M ops/s (+2.1%)
| +- hits latency | 198.483 ns/op | 194.270 ns/op (-2.1%)
| +- important_hits throughput | 5.061 M ops/s | 5.165 M ops/s (+2.1%)
|
| + num keys: 4194304
| : <before> | <after>
| +-+ hashmap (control) sequential get +----------------------+----------------------
| +- hits throughput | 2.864 M ops/s | 2.882 M ops/s ( ~ )
| +- hits latency | 365.220 ns/op | 361.418 ns/op (-1.0%)
| +- important_hits throughput | 2.864 M ops/s | 2.882 M ops/s ( ~ )
|
+---- Local Storage ----------------------
|
| + num_maps: 1
| : <before> | <after>
| +-+ local_storage cache sequential get +----------------------+----------------------
| +- hits throughput | 33.005 M ops/s | 39.068 M ops/s (+18.4%)
| +- hits latency | 30.300 ns/op | 25.598 ns/op (-15.5%)
| +- important_hits throughput | 33.005 M ops/s | 39.068 M ops/s (+18.4%)
| :
| : <before> | <after>
| +-+ local_storage cache interleaved get +----------------------+----------------------
| +- hits throughput | 37.151 M ops/s | 44.926 M ops/s (+20.9%)
| +- hits latency | 26.919 ns/op | 22.259 ns/op (-17.3%)
| +- important_hits throughput | 37.151 M ops/s | 44.926 M ops/s (+20.9%)
|
| + num_maps: 10
| : <before> | <after>
| +-+ local_storage cache sequential get +----------------------+----------------------
| +- hits throughput | 32.288 M ops/s | 38.099 M ops/s (+18.0%)
| +- hits latency | 30.972 ns/op | 26.248 ns/op (-15.3%)
| +- important_hits throughput | 3.229 M ops/s | 3.810 M ops/s (+18.0%)
| :
| : <before> | <after>
| +-+ local_storage cache interleaved get +----------------------+----------------------
| +- hits throughput | 34.473 M ops/s | 41.145 M ops/s (+19.4%)
| +- hits latency | 29.010 ns/op | 24.307 ns/op (-16.2%)
| +- important_hits throughput | 12.312 M ops/s | 14.695 M ops/s (+19.4%)
|
| + num_maps: 16
| : <before> | <after>
| +-+ local_storage cache sequential get +----------------------+----------------------
| +- hits throughput | 32.524 M ops/s | 38.341 M ops/s (+17.9%)
| +- hits latency | 30.748 ns/op | 26.083 ns/op (-15.2%)
| +- important_hits throughput | 2.033 M ops/s | 2.396 M ops/s (+17.9%)
| :
| : <before> | <after>
| +-+ local_storage cache interleaved get +----------------------+----------------------
| +- hits throughput | 34.575 M ops/s | 41.338 M ops/s (+19.6%)
| +- hits latency | 28.925 ns/op | 24.193 ns/op (-16.4%)
| +- important_hits throughput | 11.001 M ops/s | 13.153 M ops/s (+19.6%)
|
| + num_maps: 17
| : <before> | <after>
| +-+ local_storage cache sequential get +----------------------+----------------------
| +- hits throughput | 28.861 M ops/s | 32.756 M ops/s (+13.5%)
| +- hits latency | 34.649 ns/op | 30.530 ns/op (-11.9%)
| +- important_hits throughput | 1.700 M ops/s | 1.929 M ops/s (+13.5%)
| :
| : <before> | <after>
| +-+ local_storage cache interleaved get +----------------------+----------------------
| +- hits throughput | 31.529 M ops/s | 36.110 M ops/s (+14.5%)
| +- hits latency | 31.719 ns/op | 27.697 ns/op (-12.7%)
| +- important_hits throughput | 9.598 M ops/s | 10.993 M ops/s (+14.5%)
|
| + num_maps: 24
| : <before> | <after>
| +-+ local_storage cache sequential get +----------------------+----------------------
| +- hits throughput | 18.602 M ops/s | 19.937 M ops/s (+7.2%)
| +- hits latency | 53.767 ns/op | 50.166 ns/op (-6.7%)
| +- important_hits throughput | 0.776 M ops/s | 0.831 M ops/s (+7.2%)
| :
| : <before> | <after>
| +-+ local_storage cache interleaved get +----------------------+----------------------
| +- hits throughput | 21.718 M ops/s | 23.332 M ops/s (+7.4%)
| +- hits latency | 46.047 ns/op | 42.865 ns/op (-6.9%)
| +- important_hits throughput | 6.110 M ops/s | 6.564 M ops/s (+7.4%)
|
| + num_maps: 32
| : <before> | <after>
| +-+ local_storage cache sequential get +----------------------+----------------------
| +- hits throughput | 14.118 M ops/s | 14.626 M ops/s (+3.6%)
| +- hits latency | 70.856 ns/op | 68.381 ns/op (-3.5%)
| +- important_hits throughput | 0.442 M ops/s | 0.458 M ops/s (+3.6%)
| :
| : <before> | <after>
| +-+ local_storage cache interleaved get +----------------------+----------------------
| +- hits throughput | 17.111 M ops/s | 17.906 M ops/s (+4.6%)
| +- hits latency | 58.451 ns/op | 55.865 ns/op (-4.4%)
| +- important_hits throughput | 4.776 M ops/s | 4.998 M ops/s (+4.6%)
|
| + num_maps: 100
| : <before> | <after>
| +-+ local_storage cache sequential get +----------------------+----------------------
| +- hits throughput | 5.281 M ops/s | 5.528 M ops/s (+4.7%)
| +- hits latency | 192.398 ns/op | 183.059 ns/op (-4.9%)
| +- important_hits throughput | 0.053 M ops/s | 0.055 M ops/s (+4.9%)
| :
| : <before> | <after>
| +-+ local_storage cache interleaved get +----------------------+----------------------
| +- hits throughput | 6.265 M ops/s | 6.498 M ops/s (+3.7%)
| +- hits latency | 161.436 ns/op | 152.877 ns/op (-5.3%)
| +- important_hits throughput | 1.636 M ops/s | 1.697 M ops/s (+3.7%)
|
| + num_maps: 1000
| : <before> | <after>
| +-+ local_storage cache sequential get +----------------------+----------------------
| +- hits throughput | 0.355 M ops/s | 0.354 M ops/s ( ~ )
| +- hits latency | 2826.538 ns/op | 2827.139 ns/op ( ~ )
| +- important_hits throughput | 0.000 M ops/s | 0.000 M ops/s ( ~ )
| :
| : <before> | <after>
| +-+ local_storage cache interleaved get +----------------------+----------------------
| +- hits throughput | 0.404 M ops/s | 0.403 M ops/s ( ~ )
| +- hits latency | 2481.190 ns/op | 2487.555 ns/op ( ~ )
| +- important_hits throughput | 0.102 M ops/s | 0.101 M ops/s ( ~ )
The on_lookup test in {cgrp,task}_ls_recursion.c is removed
because the bpf_local_storage_lookup is no longer traceable
and adding tracepoint will make the compiler generate worse
code: https://lore.kernel.org/bpf/[email protected]/
Signed-off-by: Marco Elver <[email protected]>
Cc: Martin KaFai Lau <[email protected]>
Acked-by: Yonghong Song <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Martin KaFai Lau <[email protected]>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
Pull misc fixes from Andrew Morton:
"21 hotfixes. 12 are cc:stable and the remainder pertain to post-6.7
issues or aren't considered to be needed in earlier kernel versions"
* tag 'mm-hotfixes-stable-2024-02-10-11-16' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (21 commits)
nilfs2: fix potential bug in end_buffer_async_write
mm/damon/sysfs-schemes: fix wrong DAMOS tried regions update timeout setup
nilfs2: fix hang in nilfs_lookup_dirty_data_buffers()
MAINTAINERS: Leo Yan has moved
mm/zswap: don't return LRU_SKIP if we have dropped lru lock
fs,hugetlb: fix NULL pointer dereference in hugetlbs_fill_super
mailmap: switch email address for John Moon
mm: zswap: fix objcg use-after-free in entry destruction
mm/madvise: don't forget to leave lazy MMU mode in madvise_cold_or_pageout_pte_range()
arch/arm/mm: fix major fault accounting when retrying under per-VMA lock
selftests: core: include linux/close_range.h for CLOSE_RANGE_* macros
mm/memory-failure: fix crash in split_huge_page_to_list from soft_offline_page
mm: memcg: optimize parent iteration in memcg_rstat_updated()
nilfs2: fix data corruption in dsync block recovery for small block sizes
mm/userfaultfd: UFFDIO_MOVE implementation should use ptep_get()
exit: wait_task_zombie: kill the no longer necessary spin_lock_irq(siglock)
fs/proc: do_task_stat: use sig->stats_lock to gather the threads/children stats
fs/proc: do_task_stat: move thread_group_cputime_adjusted() outside of lock_task_sighand()
getrusage: use sig->stats_lock rather than lock_task_sighand()
getrusage: move thread_group_cputime_adjusted() outside of lock_task_sighand()
...
|
|
This exact case was fail for async crypto and we weren't
catching it.
Signed-off-by: Jakub Kicinski <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
udpgso regression test configures routing and device MTU directly through
uAPI (Netlink, ioctl) to do its job. While there is nothing wrong with it,
it takes more effort than doing it from shell.
Looking forward, we would like to extend the udpgso regression tests to
cover the EIO corner case [1], once it gets addressed. That will require a
dummy device and device feature manipulation to set it up. Which means more
Netlink code.
So, in preparation, pull out network configuration into the shell script
part of the test, so it is easily extendable in the future.
Also, because it now easy to setup routing, add a second local IPv6
address. Because the second address is not managed by the kernel, we can
"replace" the corresponding local route with a reduced-MTU one. This
unblocks the disabled "ipv6 connected" test case. Add a similar setup for
IPv4 for symmetry.
[1] https://lore.kernel.org/netdev/[email protected]/
Reviewed-by: Willem de Bruijn <[email protected]>
Signed-off-by: Jakub Sitnicki <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
Add a test case into the netlink checks that will show the number of
nested action recursions won't exceed 16. Going to 17 on a small
clone call isn't enough to exhaust the stack on (most) systems, so
it should be safe to run even on systems that don't have the fix
applied.
Signed-off-by: Aaron Conole <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
The altnames test uses the forwarding/lib.sh and that dependency
currently causes failures when running the test after install:
make -C tools/testing/selftests/ TARGETS=net install
./tools/testing/selftests/kselftest_install/run_kselftest.sh \
-t net:altnames.sh
# ...
# ./altnames.sh: line 8: ./forwarding/lib.sh: No such file or directory
# RTNETLINK answers: Operation not permitted
# ./altnames.sh: line 73: tests_run: command not found
# ./altnames.sh: line 65: pre_cleanup: command not found
Address the issue leveraging the TEST_INCLUDES infrastructure
provided by commit 2a0683be5b4c ("selftests: Introduce Makefile variable
to list shared bash scripts")
Signed-off-by: Paolo Abeni <[email protected]>
Link: https://lore.kernel.org/r/f7b1e9d468224cbc136d304362315499fe39848f.1707298927.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
Add 8 new mirred tdc tests that target mirred to block:
- Add mirred mirror to egress block action
- Add mirred mirror to ingress block action
- Add mirred redirect to egress block action
- Add mirred redirect to ingress block action
- Try to add mirred action with both dev and block
- Try to add mirred action without specifying neither dev nor block
- Replace mirred redirect to dev action with redirect to block
- Replace mirred redirect to block action with mirror to dev
Signed-off-by: Victor Nogueira <[email protected]>
Acked-by: Jamal Hadi Salim <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
The redirection test case fails in the netdev CI on debug kernels
because an FDB entry is learned despite the presence of a tc filter that
redirects incoming traffic [1].
I am unable to reproduce the failure locally, but I can see how it can
happen given that learning is first enabled and only then the ingress tc
filter is configured. On debug kernels the time window between these two
operations is longer compared to regular kernels, allowing random
packets to be transmitted and trigger learning.
Fix by reversing the order and configure the ingress tc filter before
enabling learning.
[1]
[...]
# TEST: Locked port MAB redirect [FAIL]
# Locked entry created for redirected traffic
Fixes: 38c43a1ce758 ("selftests: forwarding: Add test case for traffic redirection from a locked port")
Signed-off-by: Ido Schimmel <[email protected]>
Reviewed-by: Hangbin Liu <[email protected]>
Acked-by: Nikolay Aleksandrov <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
Suppress the following grep warnings:
[...]
INFO: # Port group entries configuration tests - (*, G)
TEST: Common port group entries configuration tests (IPv4 (*, G)) [ OK ]
TEST: Common port group entries configuration tests (IPv6 (*, G)) [ OK ]
grep: warning: stray \ before /
grep: warning: stray \ before /
grep: warning: stray \ before /
TEST: IPv4 (*, G) port group entries configuration tests [ OK ]
grep: warning: stray \ before /
grep: warning: stray \ before /
grep: warning: stray \ before /
TEST: IPv6 (*, G) port group entries configuration tests [ OK ]
[...]
They do not fail the test, but do clutter the output.
Fixes: b6d00da08610 ("selftests: forwarding: Add bridge MDB test")
Signed-off-by: Ido Schimmel <[email protected]>
Reviewed-by: Hangbin Liu <[email protected]>
Acked-by: Nikolay Aleksandrov <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
After enabling a multicast querier on the bridge (like the test is
doing), the bridge will wait for the Max Response Delay before starting
to forward according to its MDB in order to let Membership Reports
enough time to be received and processed.
Currently, the test is waiting for exactly the default Max Response
Delay (10 seconds) which is racy and leads to failures [1].
Fix by reducing the Max Response Delay to 1 second.
[1]
[...]
# TEST: IPv4 host entries forwarding tests [FAIL]
# Packet locally received after flood
Fixes: b6d00da08610 ("selftests: forwarding: Add bridge MDB test")
Signed-off-by: Ido Schimmel <[email protected]>
Reviewed-by: Hangbin Liu <[email protected]>
Acked-by: Nikolay Aleksandrov <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
After enabling a multicast querier on the bridge (like the test is
doing), the bridge will wait for the Max Response Delay before starting
to forward according to its MDB in order to let Membership Reports
enough time to be received and processed.
Currently, the test is waiting for exactly the default Max Response
Delay (10 seconds) which is racy and leads to failures [1].
Fix by reducing the Max Response Delay to 1 second.
[1]
[...]
# TEST: L2 miss - Multicast (IPv4) [FAIL]
# Unregistered multicast filter was hit after adding MDB entry
Fixes: 8c33266ae26a ("selftests: forwarding: Add layer 2 miss test cases")
Signed-off-by: Ido Schimmel <[email protected]>
Reviewed-by: Hangbin Liu <[email protected]>
Acked-by: Nikolay Aleksandrov <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
The test toggles the carrier of a bridge port in order to test the
bridge backup port feature.
Due to the linkwatch delayed work the carrier change is not always
reflected fast enough to the bridge driver and packets are not forwarded
as the test expects, resulting in failures [1].
Fix by busy waiting on the bridge port state until it changes to the
desired state following the carrier change.
[1]
# Backup port
# -----------
[...]
# TEST: swp1 carrier off [ OK ]
# TEST: No forwarding out of swp1 [FAIL]
[ 641.995910] br0: port 1(swp1) entered disabled state
# TEST: No forwarding out of vx0 [ OK ]
Fixes: b408453053fb ("selftests: net: Add bridge backup port and backup nexthop ID test")
Signed-off-by: Ido Schimmel <[email protected]>
Reviewed-by: Petr Machata <[email protected]>
Acked-by: Paolo Abeni <[email protected]>
Acked-by: Nikolay Aleksandrov <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
The two tests that make use of multicast routig (router.sh and
router_multicast.sh) are currently failing in the netdev CI because the
kernel is missing multicast routing support.
Fix by adding the required config entries.
Fixes: 6d4efada3b82 ("selftests: forwarding: Add multicast routing test")
Signed-off-by: Ido Schimmel <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
The reuseport_addr_any.sh is currently skipping DCCP tests and
pmtu.sh is skipping all the FOU/GUE related cases: add the missing
options.
Signed-off-by: Paolo Abeni <[email protected]>
Reviewed-by: Eric Dumazet <[email protected]>
Link: https://lore.kernel.org/r/38d3ca7f909736c1aef56e6244d67c82a9bba6ff.1707326987.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <[email protected]>
|