compile_commands.json is used by clangd[1] to provide code navigation
and completion functionality to editors. See [2] for an example
configuration that includes this functionality for VSCode.
It can currently be built manually when using kunit.py, by running:
./scripts/clang-tools/gen_compile_commands.py -d .kunit
With this change however, it's built automatically so you don't need to
manually keep it up to date.
Unlike the manual approach, having make build the compile_commands.json
means that it appears in the build output tree instead of at the root of
the source tree, so you'll need to add --compile-commands-dir= to your
clangd args for it to be found.
[1] https://clangd.llvm.org/
[2] https://github.com/FlorentRevest/linux-kernel-vscode
Signed-off-by: Brendan Jackman <jackmanb(a)google.com>
---
tools/testing/kunit/kunit_kernel.py | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/tools/testing/kunit/kunit_kernel.py b/tools/testing/kunit/kunit_kernel.py
index 7254c110ff23..61931c4926fd 100644
--- a/tools/testing/kunit/kunit_kernel.py
+++ b/tools/testing/kunit/kunit_kernel.py
@@ -72,7 +72,8 @@ class LinuxSourceTreeOperations:
raise ConfigError(e.output.decode())
def make(self, jobs: int, build_dir: str, make_options: Optional[List[str]]) -> None:
- command = ['make', 'ARCH=' + self._linux_arch, 'O=' + build_dir, '--jobs=' + str(jobs)]
+ command = ['make', 'all', 'compile_commands.json', 'ARCH=' + self._linux_arch,
+ 'O=' + build_dir, '--jobs=' + str(jobs)]
if make_options:
command.extend(make_options)
if self._cross_compile:
---
base-commit: 3c999d1ae3c75991902a1a7dad0cb62c2a3008b4
change-id: 20240516-kunit-compile-commands-d994074fc2be
Best regards,
--
Brendan Jackman <jackmanb(a)google.com>
This patch series adds a some not yet picked selftests to the kvm s390x
selftest suite.
The additional test cases are covering:
* Assert KVM_EXIT_S390_UCONTROL exit on not mapped memory access
* Assert functionality of storage keys in ucontrol VM
* Assert that memory region operations are rejected for ucontrol VMs
Running the test cases requires sys_admin capabilities to start the
ucontrol VM.
This can be achieved by running as root or with a command like:
sudo setpriv --reuid nobody --inh-caps -all,+sys_admin \
--ambient-caps -all,+sys_admin --bounding-set -all,+sys_admin \
./ucontrol_test
---
The patches in this series have been part of the previous patch series.
The test cases added here do depend on the fixture added in the earlier
patches.
From v5 PATCH 7-9 the segment and page table generation has been removed
and DAT
has been disabled. Since DAT is not necessary to validate the KVM code.
https://lore.kernel.org/kvm/20240807154512.316936-1-schlameuss@linux.ibm.co…
v2:
- Reenable KSS intercept and handle it within skey test.
- Modify the checked register between storing (sske) and reading (iske)
it within the test program to make sure the.
- Add an additional state assertion in the end of uc_skey
- Fix some typos and white spaces.
v1:
- Remove segment and page table generation and disable DAT. This is not
necessary to validate the KVM code.
Christoph Schlameuss (3):
selftests: kvm: s390: Add uc_map_unmap VM test case
selftests: kvm: s390: Add uc_skey VM test case
selftests: kvm: s390: Verify reject memory region operations for
ucontrol VMs
.../selftests/kvm/s390x/ucontrol_test.c | 232 +++++++++++++++++-
1 file changed, 230 insertions(+), 2 deletions(-)
--
2.46.0
Resent due to missing linux-kernel@ mailing list inclusion.
There were several attempts to resolve circular include dependency
after the addition of percpu.h: 1c9df907da83 ("random: fix circular
include dependency on arm64 after addition of percpu.h"), c0842fbc1b18
("random32: move the pseudo-random 32-bit definitions to prandom.h") and
finally d9f29deb7fe8 ("prandom: Remove unused include") that completely
removes the inclusion of <linux/percpu.h>.
Due to legacy reasons, <linux/random.h> includes <linux/prandom.h>, but
with the commit entry remark:
--quote--
A further cleanup step would be to remove this from <linux/random.h>
entirely, and make people who use the prandom infrastructure include
just the new header file. That's a bit of a churn patch, but grepping
for "prandom_" and "next_pseudo_random32" "struct rnd_state" should
catch most users.
But it turns out that that nice cleanup step is fairly painful, because
a _lot_ of code currently seems to depend on the implicit include of
<linux/random.h>, which can currently come in a lot of ways, including
such fairly core headfers as <linux/net.h>.
So the "nice cleanup" part may or may never happen.
--/quote--
We would like to include <linux/percpu.h> in <linux/prandom.h>.
In [1] we would like to repurpose __percpu tag as a named address space
qualifier, where __percpu macro uses defines from <linux/percpu.h>.
The major roadblock to inclusion of <linux/percpu.h> is the above
mentioned legacy inclusion of <linux/prandom.h> in <linux/random.h> that
causes circular include dependency that prevents <linux/percpu.h>
inclusion.
This patch series is the "nice cleanup" part that:
a) Substitutes the inclusion of <linux/random.h> with the
inclusion of <linux/prandom.h> where needed (patches 1 - 17).
b) Removes legacy inclusion of <linux/prandom.h> from
<linux/random.h> (patch 18).
c) Includes <linux/percpu.h> in <linux/prandom.h> (patch 19).
The whole series was tested by compiling the kernel for x86_64 allconfig
and some popular architectures, namely arm64 defconfig, powerpc defconfig
and loongarch defconfig.
[1] https://lore.kernel.org/lkml/20240812115945.484051-4-ubizjak@gmail.com/
Cc: Dave Hansen <dave.hansen(a)linux.intel.com>
Cc: Andy Lutomirski <luto(a)kernel.org>
Cc: Peter Zijlstra <peterz(a)infradead.org>
Cc: Thomas Gleixner <tglx(a)linutronix.de>
Cc: Ingo Molnar <mingo(a)redhat.com>
Cc: Borislav Petkov <bp(a)alien8.de>
Cc: x86(a)kernel.org
Cc: "H. Peter Anvin" <hpa(a)zytor.com>
Cc: Jani Nikula <jani.nikula(a)linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen(a)linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi(a)intel.com>
Cc: Tvrtko Ursulin <tursulin(a)ursulin.net>
Cc: David Airlie <airlied(a)gmail.com>
Cc: Daniel Vetter <daniel(a)ffwll.ch>
Cc: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com>
Cc: Maxime Ripard <mripard(a)kernel.org>
Cc: Thomas Zimmermann <tzimmermann(a)suse.de>
Cc: Hans Verkuil <hverkuil(a)xs4all.nl>
Cc: Mauro Carvalho Chehab <mchehab(a)kernel.org>
Cc: Miquel Raynal <miquel.raynal(a)bootlin.com>
Cc: Richard Weinberger <richard(a)nod.at>
Cc: Vignesh Raghavendra <vigneshr(a)ti.com>
Cc: Eric Biggers <ebiggers(a)kernel.org>
Cc: "Theodore Y. Ts'o" <tytso(a)mit.edu>
Cc: Jaegeuk Kim <jaegeuk(a)kernel.org>
Cc: "Jason A. Donenfeld" <Jason(a)zx2c4.com>
Cc: Linus Torvalds <torvalds(a)linux-foundation.org>
Cc: Hannes Reinecke <hare(a)suse.de>
Cc: "James E.J. Bottomley" <James.Bottomley(a)HansenPartnership.com>
Cc: "Martin K. Petersen" <martin.petersen(a)oracle.com>
Cc: Alexei Starovoitov <ast(a)kernel.org>
Cc: Daniel Borkmann <daniel(a)iogearbox.net>
Cc: John Fastabend <john.fastabend(a)gmail.com>
Cc: Andrii Nakryiko <andrii(a)kernel.org>
Cc: Martin KaFai Lau <martin.lau(a)linux.dev>
Cc: Eduard Zingerman <eddyz87(a)gmail.com>
Cc: Song Liu <song(a)kernel.org>
Cc: Yonghong Song <yonghong.song(a)linux.dev>
Cc: KP Singh <kpsingh(a)kernel.org>
Cc: Stanislav Fomichev <sdf(a)fomichev.me>
Cc: Hao Luo <haoluo(a)google.com>
Cc: Jiri Olsa <jolsa(a)kernel.org>
Cc: Andrew Morton <akpm(a)linux-foundation.org>
Cc: Brendan Higgins <brendan.higgins(a)linux.dev>
Cc: David Gow <davidgow(a)google.com>
Cc: Rae Moar <rmoar(a)google.com>
Cc: "David S. Miller" <davem(a)davemloft.net>
Cc: Eric Dumazet <edumazet(a)google.com>
Cc: Jakub Kicinski <kuba(a)kernel.org>
Cc: Paolo Abeni <pabeni(a)redhat.com>
Cc: Jiri Pirko <jiri(a)resnulli.us>
Cc: Petr Mladek <pmladek(a)suse.com>
Cc: Steven Rostedt <rostedt(a)goodmis.org>
Cc: Andy Shevchenko <andriy.shevchenko(a)linux.intel.com>
Cc: Rasmus Villemoes <linux(a)rasmusvillemoes.dk>
Cc: Sergey Senozhatsky <senozhatsky(a)chromium.org>
Cc: Stephen Hemminger <stephen(a)networkplumber.org>
Cc: Jamal Hadi Salim <jhs(a)mojatatu.com>
Cc: Cong Wang <xiyou.wangcong(a)gmail.com>
Cc: Kent Overstreet <kent.overstreet(a)linux.dev>
---
v2: - Reword commit messages to mention the removal of legacy inclusion
of <linux/prandom.h> from <linux/random.h>
- Add missing substitution in crypto/testmgr.c
(reported by kernel test robot)
- Add Acked-by:.
Uros Bizjak (19):
x86/kaslr: Include <linux/prandom.h> instead of <linux/random.h>
crypto: testmgr: Include <linux/prandom.h> instead of <linux/random.h>
drm/i915/selftests: Include <linux/prandom.h> instead of
<linux/random.h>
drm/lib: Include <linux/prandom.h> instead of <linux/random.h>
media: vivid: Include <linux/prandom.h> in vivid-vid-cap.c
mtd: tests: Include <linux/prandom.h> instead of <linux/random.h>
fscrypt: Include <linux/once.h> in fs/crypto/keyring.c
scsi: libfcoe: Include <linux/prandom.h> instead of <linux/random.h>
bpf: Include <linux/prandom.h> instead of <linux/random.h>
lib/interval_tree_test.c: Include <linux/prandom.h> instead of
<linux/random.h>
kunit: string-stream-test: Include <linux/prandom.h> instead of
<linux/random.h>
random32: Include <linux/prandom.h> instead of <linux/random.h>
lib/rbtree-test: Include <linux/prandom.h> instead of <linux/random.h>
bpf/tests: Include <linux/prandom.h> instead of <linux/random.h>
lib/test_parman: Include <linux/prandom.h> instead of <linux/random.h>
lib/test_scanf: Include <linux/prandom.h> instead of <linux/random.h>
netem: Include <linux/prandom.h> in sch_netem.c
random: Do not include <linux/prandom.h> in <linux/random.h>
prandom: Include <linux/percpu.h> in <linux/prandom.h>
arch/x86/mm/kaslr.c | 2 +-
crypto/testmgr.c | 2 +-
drivers/gpu/drm/i915/selftests/i915_gem.c | 2 +-
drivers/gpu/drm/i915/selftests/i915_random.h | 2 +-
drivers/gpu/drm/i915/selftests/scatterlist.c | 2 +-
drivers/gpu/drm/lib/drm_random.h | 2 +-
drivers/media/test-drivers/vivid/vivid-vid-cap.c | 1 +
drivers/mtd/tests/oobtest.c | 2 +-
drivers/mtd/tests/pagetest.c | 2 +-
drivers/mtd/tests/subpagetest.c | 2 +-
fs/crypto/keyring.c | 1 +
include/linux/prandom.h | 1 +
include/linux/random.h | 7 -------
include/scsi/libfcoe.h | 2 +-
kernel/bpf/core.c | 2 +-
lib/interval_tree_test.c | 2 +-
lib/kunit/string-stream-test.c | 1 +
lib/random32.c | 2 +-
lib/rbtree_test.c | 2 +-
lib/test_bpf.c | 2 +-
lib/test_parman.c | 2 +-
lib/test_scanf.c | 2 +-
net/sched/sch_netem.c | 1 +
23 files changed, 22 insertions(+), 24 deletions(-)
--
2.46.0
v26: https://patchwork.kernel.org/project/netdevbpf/list/?series=888227&state=*
====
No major changes. Only applied Reviewed-by tags from Jakub and addressed
reported nits.
v25: https://patchwork.kernel.org/project/netdevbpf/list/?series=885396&state=*
===
Major changes:
- Moved devmem.h and mp_dmabuf_devmem.h to internal header files.
- Changed the page_pool_params to take in a queue_idx rather than
a struct netdev_rx_queue.
- Added WARN_ON_ONCE around __skb_checksum readability check and added
check to skb_checksum_help().
Other more minor feedback addressed as well.
v24: https://patchwork.kernel.org/project/netdevbpf/list/?series=884556&state=*
====
No major changes. Mostly addressing issues in the error paths of dmabuf
binding, and code cleanups/improvements from reviewers:
Changes:
- Fix failing ynl regen error.
- Error path fixes & extack error messages in dmabuf binding.
- Code cleanup in introspection.
- gitignore ynl.d generated file.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v24/
v23: https://patchwork.kernel.org/project/netdevbpf/list/?series=882978&state=*
====
Fixing relatively minor issues called out in v22. (thanks again!)
Mostly code cleanups, extack error messages, and minor reworks. Nothing
major really changed, so the exact changes per commit is called in the
commit messages.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v23/
v22: https://patchwork.kernel.org/project/netdevbpf/list/?series=881158&state=*
====
v22 aims to resolve the pending issue pointed to in v21, which is the
interaction with xdp. In this series I rebase on top of the minor
refactor which refactors propagating xdp configuration to slave devices:
https://patchwork.kernel.org/project/netdevbpf/list/?series=881994&state=*
I then disable setting xdp on devices using memory providers, and
propagating xdp configuration to devices using memory providers.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v22/
v21: https://patchwork.kernel.org/project/netdevbpf/list/?series=880735&state=*
====
v20 addressed some comments and resolved a test failure, but introduced
an unfortunate build error with a config edge case I wasn't testing. v21
simply resolves that error.
Major Changes:
- Resolve build error with CONFIG_PAGE_POOL=n && CONFIG_NET=y
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v21/
v20: https://patchwork.kernel.org/project/netdevbpf/list/?series=879373&state=*
====
v20 aims to resolve a couple of bug reports against v19, and addresses
some review comments around the page_pool_check_memory_provider
mechanism.
Major changes:
- Test edge cases such as header split disabled in selftest.
- Change `offset = 0` back to `offset = offset - start` to resolve issue
found in RX path by Taehee (thanks!)
- Address a few comments around page_pool_check_memory_provider() from
Pavel & Jakub.
- Removed some unnecessary includes across various patches in the
series.
- Removed unnecessary EXPORT_SYMBOL(page_pool_mem_providers) (Jakub).
- Fix regression caused by incorrect dev_get_max_mp_channel check, along
with rename (Jakub).
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v20/
v19: https://patchwork.kernel.org/project/netdevbpf/list/?series=876852&state=*
====
v18 got a thorough review (thanks!), and this iteration addresses the
feedback.
Major changes:
- Prevent deactivating mp bound queues.
- Prevent installing xdp on mp bound netdevs, or installing mps on xdp
installed netdevs.
- Fix corner cases in netlink API vis-a-vis missing attributes.
- Iron out the unreadable netmem driver support story. To be honest, the
conversation with Jakub & Pavel got a bit confusing for me. I've
implemented an approach in this set that makes sense to me, and
AFAICT, addresses the requirements. It may be good as-is, or it
may be a conversation starter/continuer. To be honest IMO there
are many ways to skin this cat and I don't see an extremely strong
reason to go for one approach over another. Here is one approach you
may like.
- Don't reset niov dma_addr on allocation & free.
- Add some tests to the selftest that catches some of the issues around
missing netlink attributes or deactivating mp-bound queues.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v19/
v18: https://patchwork.kernel.org/project/netdevbpf/list/?series=874848&state=*
====
v17 got minor feedback: (a) to beef up the description on patch 1 and (b)
to remove the leading underscores in the header definition.
I applied (a). (b) seems to be against current conventions so I did not
apply before further discussion.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v17/
v17: https://patchwork.kernel.org/project/netdevbpf/list/?series=869900&state=*
====
v16 also got a very thorough review and some testing (thanks again!).
Thes version addresses all the concerns reported on v15, in terms of
feedback and issues reported.
Major changes:
- Use ASSERT_RTNL.
- Moved around some of the page_pool helpers definitions so I can hide
some netmem helpers in private files as Jakub suggested.
- Don't make every net_iov hold a ref on the binding as Jakub suggested.
- Fix issue reported by Taehee where we access queues after they have
been freed.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v17/
v16: https://patchwork.kernel.org/project/netdevbpf/list/?series=866353&state=*
====
v15 got a thorough review and some testing, and this version addresses almost
all the feedback. Some more minor comments where the authors said it
could be done later, I left out.
Major changes:
- Addition of dma-buf introspection to page-pool-get and queue-get.
- Fixes to selftests suggested by Taehee.
- Fixes to documentation suggested by Donald.
- A couple of suggestions and fixes to TCP patches by Eric and David.
- Fixes to number assignements suggested by Arnd.
- Use rtnl_lock()ing to guard against queue reconfiguration while the
page_pool initialization is happening. (Jakub).
- Fixes to a few warnings reproduced by Taehee.
- Fixes to dma-buf binding suggested by Taehee and Jakub.
- Fixes to netlink UAPI suggested by Jakub
- Applied a number of Reviewed-bys and Acked-bys (including ones I lost
from v13+).
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v16/
One caveat: Taehee reproduced a KASAN warning and reported it here:
https://lore.kernel.org/netdev/CAMArcTUdCxOBYGF3vpbq=eBvqZfnc44KBaQTN7H-wqd…
I estimate the issue to be minor and easily fixable:
https://lore.kernel.org/netdev/CAHS8izNgaqC--GGE2xd85QB=utUnOHmioCsDd1TNxJW…
I hope to be able to follow up with a fix to net tree as net-next closes
imminently, but if this iteration doesn't make it in, I will repost with
a fix squashed after net-next reopens, no problem.
v15: https://patchwork.kernel.org/project/netdevbpf/list/?series=865481&state=*
====
No material changes in this version, only a fix to linking against
libynl.a from the last version. Per Jakub's instructions I've pulled one
of his patches into this series, and now use the new libynl.a correctly,
I hope.
As usual, the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v15/
v14: https://patchwork.kernel.org/project/netdevbpf/list/?series=865135&archive=…
====
No material changes in this version. Only rebase and re-verification on
top of net-next. v13, I think, raced with commit ebad6d0334793
("net/ipv4: Use nested-BH locking for ipv4_tcp_sk.") being merged to
net-next that caused a patchwork failure to apply. This series should
apply cleanly on commit c4532232fa2a4 ("selftests: net: remove unneeded
IP_GRE config").
I did not wait the customary 24hr as Jakub said it's OK to repost as soon
as I build test the rebased version:
https://lore.kernel.org/netdev/20240625075926.146d769d@kernel.org/
v13: https://patchwork.kernel.org/project/netdevbpf/list/?series=861406&archive=…
====
Major changes:
--------------
This iteration addresses Pavel's review comments, applies his
reviewed-by's, and seeks to fix the patchwork build error (sorry!).
As usual, the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v13/
v12: https://patchwork.kernel.org/project/netdevbpf/list/?series=859747&state=*
====
Major changes:
--------------
This iteration only addresses one minor comment from Pavel with regards
to the trace printing of netmem, and the patchwork build error
introduced in v11 because I missed doing an allmodconfig build, sorry.
Other than that v11, AFAICT, received no feedback. There is one
discussion about how the specifics of plugging io uring memory through
the page pool, but not relevant to content in this particular patchset,
AFAICT.
As usual, the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v12/
v11: https://patchwork.kernel.org/project/netdevbpf/list/?series=857457&state=*
====
Major Changes:
--------------
v11 addresses feedback received in v10. The major change is the removal
of the memory provider ops as requested by Christoph. We still
accomplish the same thing, but utilizing direct function calls with if
statements rather than generic ops.
Additionally address sparse warnings, bugs and review comments from
folks that reviewed.
As usual, the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v11/
Detailed changelog:
-------------------
- Fixes in netdev_rx_queue_restart() from Pavel & David.
- Remove commit e650e8c3a36f5 ("net: page_pool: create hooks for
custom page providers") from the series to address Christoph's
feedback and rebased other patches on the series on this change.
- Fixed build errors with CONFIG_DMA_SHARED_BUFFER &&
!CONFIG_GENERIC_ALLOCATOR build.
- Fixed sparse warnings pointed out by Paolo.
- Drop unnecessary gro_pull_from_frag0 checks.
- Added Bagas reviewed-by to docs.
v10: https://patchwork.kernel.org/project/netdevbpf/list/?series=852422&state=*
====
Major Changes:
--------------
v9 was sent right before the merge window closed (sorry!). v10 is almost
a re-send of the series now that the merge window re-opened. Only
rebased to latest net-next and addressed some minor iterative comments
received on v9.
As usual, the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v10/
Detailed changelog:
-------------------
- Fixed tokens leaking in DONTNEED setsockopt (Nikolay).
- Moved net_iov_dma_addr() to devmem.c and made it a devmem specific
helpers (David).
- Rename hook alloc_pages to alloc_netmems as alloc_pages is now
preprocessor macro defined and causes a build error.
v9:
===
Major Changes:
--------------
GVE queue API has been merged. Submitting this version as non-RFC after
rebasing on top of the merged API, and dropped the out of tree queue API
I was carrying on github. Addressed the little feedback v8 has received.
Detailed changelog:
------------------
- Added new patch from David Wei to this series for
netdev_rx_queue_restart()
- Fixed sparse error.
- Removed CONFIG_ checks in netmem_is_net_iov()
- Flipped skb->readable to skb->unreadable
- Minor fixes to selftests & docs.
RFC v8:
=======
Major Changes:
--------------
- Fixed build error generated by patch-by-patch build.
- Applied docs suggestions from Randy.
RFC v7:
=======
Major Changes:
--------------
This revision largely rebases on top of net-next and addresses the feedback
RFCv6 received from folks, namely Jakub, Yunsheng, Arnd, David, & Pavel.
The series remains in RFC because the queue-API ndos defined in this
series are not yet implemented. I have a GVE implementation I carry out
of tree for my testing. A upstreamable GVE implementation is in the
works. Aside from that, in my estimation all the patches are ready for
review/merge. Please do take a look.
As usual the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v7/
Detailed changelog:
- Use admin-perm in netlink API.
- Addressed feedback from Jakub with regards to netlink API
implementation.
- Renamed devmem.c functions to something more appropriate for that
file.
- Improve the performance seen through the page_pool benchmark.
- Fix the value definition of all the SO_DEVMEM_* uapi.
- Various fixes to documentation.
Perf - page-pool benchmark:
---------------------------
Improved performance of bench_page_pool_simple.ko tests compared to v6:
https://pastebin.com/raw/v5dYRg8L
net-next base: 8 cycle fast path.
RFC v6: 10 cycle fast path.
RFC v7: 9 cycle fast path.
RFC v7 with CONFIG_DMA_SHARED_BUFFER disabled: 8 cycle fast path,
same as baseline.
Perf - Devmem TCP benchmark:
---------------------
Perf is about the same regardless of the changes in v7, namely the
removal of the static_branch_unlikely to improve the page_pool benchmark
performance:
189/200gbps bi-directional throughput with RX devmem TCP and regular TCP
TX i.e. ~95% line rate.
RFC v6:
=======
Major Changes:
--------------
This revision largely rebases on top of net-next and addresses the little
feedback RFCv5 received.
The series remains in RFC because the queue-API ndos defined in this
series are not yet implemented. I have a GVE implementation I carry out
of tree for my testing. A upstreamable GVE implementation is in the
works. Aside from that, in my estimation all the patches are ready for
review/merge. Please do take a look.
As usual the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v6/
This version also comes with some performance data recorded in the cover
letter (see below changelog).
Detailed changelog:
- Rebased on top of the merged netmem_ref changes.
- Converted skb->dmabuf to skb->readable (Pavel). Pavel's original
suggestion was to remove the skb->dmabuf flag entirely, but when I
looked into it closely, I found the issue that if we remove the flag
we have to dereference the shinfo(skb) pointer to obtain the first
frag to tell whether an skb is readable or not. This can cause a
performance regression if it dirties the cache line when the
shinfo(skb) was not really needed. Instead, I converted the skb->dmabuf
flag into a generic skb->readable flag which can be re-used by io_uring
0-copy RX.
- Squashed a few locking optimizations from Eric Dumazet in the RX path
and the DEVMEM_DONTNEED setsockopt.
- Expanded the tests a bit. Added validation for invalid scenarios and
added some more coverage.
Perf - page-pool benchmark:
---------------------------
bench_page_pool_simple.ko tests with and without these changes:
https://pastebin.com/raw/ncHDwAbn
AFAIK the number that really matters in the perf tests is the
'tasklet_page_pool01_fast_path Per elem'. This one measures at about 8
cycles without the changes but there is some 1 cycle noise in some
results.
With the patches this regresses to 9 cycles with the changes but there
is 1 cycle noise occasionally running this test repeatedly.
Lastly I tried disable the static_branch_unlikely() in
netmem_is_net_iov() check. To my surprise disabling the
static_branch_unlikely() check reduces the fast path back to 8 cycles,
but the 1 cycle noise remains.
Perf - Devmem TCP benchmark:
---------------------
189/200gbps bi-directional throughput with RX devmem TCP and regular TCP
TX i.e. ~95% line rate.
Major changes in RFC v5:
========================
1. Rebased on top of 'Abstract page from net stack' series and used the
new netmem type to refer to LSB set pointers instead of re-using
struct page.
2. Downgraded this series back to RFC and called it RFC v5. This is
because this series is now dependent on 'Abstract page from net
stack'[1] and the queue API. Both are removed from the series to
reduce the patch # and those bits are fairly independent or
pre-requisite work.
3. Reworked the page_pool devmem support to use netmem and for some
more unified handling.
4. Reworked the reference counting of net_iov (renamed from
page_pool_iov) to use pp_ref_count for refcounting.
The full changes including the dependent series and GVE page pool
support is here:
https://github.com/mina/linux/commits/tcpdevmem-rfcv5/
[1] https://patchwork.kernel.org/project/netdevbpf/list/?series=810774
Major changes in v1:
====================
1. Implemented MVP queue API ndos to remove the userspace-visible
driver reset.
2. Fixed issues in the napi_pp_put_page() devmem frag unref path.
3. Removed RFC tag.
Many smaller addressed comments across all the patches (patches have
individual change log).
Full tree including the rest of the GVE driver changes:
https://github.com/mina/linux/commits/tcpdevmem-v1
Changes in RFC v3:
==================
1. Pulled in the memory-provider dependency from Jakub's RFC[1] to make the
series reviewable and mergeable.
2. Implemented multi-rx-queue binding which was a todo in v2.
3. Fix to cmsg handling.
The sticking point in RFC v2[2] was the device reset required to refill
the device rx-queues after the dmabuf bind/unbind. The solution
suggested as I understand is a subset of the per-queue management ops
Jakub suggested or similar:
https://lore.kernel.org/netdev/20230815171638.4c057dcd@kernel.org/
This is not addressed in this revision, because:
1. This point was discussed at netconf & netdev and there is openness to
using the current approach of requiring a device reset.
2. Implementing individual queue resetting seems to be difficult for my
test bed with GVE. My prototype to test this ran into issues with the
rx-queues not coming back up properly if reset individually. At the
moment I'm unsure if it's a mistake in the POC or a genuine issue in
the virtualization stack behind GVE, which currently doesn't test
individual rx-queue restart.
3. Our usecases are not bothered by requiring a device reset to refill
the buffer queues, and we'd like to support NICs that run into this
limitation with resetting individual queues.
My thought is that drivers that have trouble with per-queue configs can
use the support in this series, while drivers that support new netdev
ops to reset individual queues can automatically reset the queue as
part of the dma-buf bind/unbind.
The same approach with device resets is presented again for consideration
with other sticking points addressed.
This proposal includes the rx devmem path only proposed for merge. For a
snapshot of my entire tree which includes the GVE POC page pool support &
device memory support:
https://github.com/torvalds/linux/compare/master...mina:linux:tcpdevmem-v3
[1] https://lore.kernel.org/netdev/f8270765-a27b-6ccf-33ea-cda097168d79@redhat.…
[2] https://lore.kernel.org/netdev/CAHS8izOVJGJH5WF68OsRWFKJid1_huzzUK+hpKbLcL4…
Changes in RFC v2:
==================
The sticking point in RFC v1[1] was the dma-buf pages approach we used to
deliver the device memory to the TCP stack. RFC v2 is a proof-of-concept
that attempts to resolve this by implementing scatterlist support in the
networking stack, such that we can import the dma-buf scatterlist
directly. This is the approach proposed at a high level here[2].
Detailed changes:
1. Replaced dma-buf pages approach with importing scatterlist into the
page pool.
2. Replace the dma-buf pages centric API with a netlink API.
3. Removed the TX path implementation - there is no issue with
implementing the TX path with scatterlist approach, but leaving
out the TX path makes it easier to review.
4. Functionality is tested with this proposal, but I have not conducted
perf testing yet. I'm not sure there are regressions, but I removed
perf claims from the cover letter until they can be re-confirmed.
5. Added Signed-off-by: contributors to the implementation.
6. Fixed some bugs with the RX path since RFC v1.
Any feedback welcome, but specifically the biggest pending questions
needing feedback IMO are:
1. Feedback on the scatterlist-based approach in general.
2. Netlink API (Patch 1 & 2).
3. Approach to handle all the drivers that expect to receive pages from
the page pool (Patch 6).
[1] https://lore.kernel.org/netdev/dfe4bae7-13a0-3c5d-d671-f61b375cb0b4@gmail.c…
[2] https://lore.kernel.org/netdev/CAHS8izPm6XRS54LdCDZVd0C75tA1zHSu6jLVO8nzTLX…
==================
* TL;DR:
Device memory TCP (devmem TCP) is a proposal for transferring data to and/or
from device memory efficiently, without bouncing the data to a host memory
buffer.
* Problem:
A large amount of data transfers have device memory as the source and/or
destination. Accelerators drastically increased the volume of such transfers.
Some examples include:
- ML accelerators transferring large amounts of training data from storage into
GPU/TPU memory. In some cases ML training setup time can be as long as 50% of
TPU compute time, improving data transfer throughput & efficiency can help
improving GPU/TPU utilization.
- Distributed training, where ML accelerators, such as GPUs on different hosts,
exchange data among them.
- Distributed raw block storage applications transfer large amounts of data with
remote SSDs, much of this data does not require host processing.
Today, the majority of the Device-to-Device data transfers the network are
implemented as the following low level operations: Device-to-Host copy,
Host-to-Host network transfer, and Host-to-Device copy.
The implementation is suboptimal, especially for bulk data transfers, and can
put significant strains on system resources, such as host memory bandwidth,
PCIe bandwidth, etc. One important reason behind the current state is the
kernel’s lack of semantics to express device to network transfers.
* Proposal:
In this patch series we attempt to optimize this use case by implementing
socket APIs that enable the user to:
1. send device memory across the network directly, and
2. receive incoming network packets directly into device memory.
Packet _payloads_ go directly from the NIC to device memory for receive and from
device memory to NIC for transmit.
Packet _headers_ go to/from host memory and are processed by the TCP/IP stack
normally. The NIC _must_ support header split to achieve this.
Advantages:
- Alleviate host memory bandwidth pressure, compared to existing
network-transfer + device-copy semantics.
- Alleviate PCIe BW pressure, by limiting data transfer to the lowest level
of the PCIe tree, compared to traditional path which sends data through the
root complex.
* Patch overview:
** Part 1: netlink API
Gives user ability to bind dma-buf to an RX queue.
** Part 2: scatterlist support
Currently the standard for device memory sharing is DMABUF, which doesn't
generate struct pages. On the other hand, networking stack (skbs, drivers, and
page pool) operate on pages. We have 2 options:
1. Generate struct pages for dmabuf device memory, or,
2. Modify the networking stack to process scatterlist.
Approach #1 was attempted in RFC v1. RFC v2 implements approach #2.
** part 3: page pool support
We piggy back on page pool memory providers proposal:
https://github.com/kuba-moo/linux/tree/pp-providers
It allows the page pool to define a memory provider that provides the
page allocation and freeing. It helps abstract most of the device memory
TCP changes from the driver.
** part 4: support for unreadable skb frags
Page pool iovs are not accessible by the host; we implement changes
throughput the networking stack to correctly handle skbs with unreadable
frags.
** Part 5: recvmsg() APIs
We define user APIs for the user to send and receive device memory.
Not included with this series is the GVE devmem TCP support, just to
simplify the review. Code available here if desired:
https://github.com/mina/linux/tree/tcpdevmem
This series is built on top of net-next with Jakub's pp-providers changes
cherry-picked.
* NIC dependencies:
1. (strict) Devmem TCP require the NIC to support header split, i.e. the
capability to split incoming packets into a header + payload and to put
each into a separate buffer. Devmem TCP works by using device memory
for the packet payload, and host memory for the packet headers.
2. (optional) Devmem TCP works better with flow steering support & RSS support,
i.e. the NIC's ability to steer flows into certain rx queues. This allows the
sysadmin to enable devmem TCP on a subset of the rx queues, and steer
devmem TCP traffic onto these queues and non devmem TCP elsewhere.
The NIC I have access to with these properties is the GVE with DQO support
running in Google Cloud, but any NIC that supports these features would suffice.
I may be able to help reviewers bring up devmem TCP on their NICs.
* Testing:
The series includes a udmabuf kselftest that show a simple use case of
devmem TCP and validates the entire data path end to end without
a dependency on a specific dmabuf provider.
** Test Setup
Kernel: net-next with this series and memory provider API cherry-picked
locally.
Hardware: Google Cloud A3 VMs.
NIC: GVE with header split & RSS & flow steering support.
Cc: Pavel Begunkov <asml.silence(a)gmail.com>
Cc: David Wei <dw(a)davidwei.uk>
Cc: Jason Gunthorpe <jgg(a)ziepe.ca>
Cc: Yunsheng Lin <linyunsheng(a)huawei.com>
Cc: Shailend Chand <shailend(a)google.com>
Cc: Harshitha Ramamurthy <hramamurthy(a)google.com>
Cc: Shakeel Butt <shakeel.butt(a)linux.dev>
Cc: Jeroen de Borst <jeroendb(a)google.com>
Cc: Praveen Kaligineedi <pkaligineedi(a)google.com>
Cc: Bagas Sanjaya <bagasdotme(a)gmail.com>
Cc: Steven Rostedt <rostedt(a)goodmis.org>
Cc: Christoph Hellwig <hch(a)infradead.org>
Cc: Nikolay Aleksandrov <razor(a)blackwall.org>
Cc: Taehee Yoo <ap420073(a)gmail.com>
Cc: Donald Hunter <donald.hunter(a)gmail.com>
Mina Almasry (13):
netdev: add netdev_rx_queue_restart()
net: netdev netlink api to bind dma-buf to a net device
netdev: support binding dma-buf to netdevice
netdev: netdevice devmem allocator
page_pool: devmem support
memory-provider: dmabuf devmem memory provider
net: support non paged skb frags
net: add support for skbs with unreadable frags
tcp: RX path for devmem TCP
net: add SO_DEVMEM_DONTNEED setsockopt to release RX frags
net: add devmem TCP documentation
selftests: add ncdevmem, netcat for devmem TCP
netdev: add dmabuf introspection
Documentation/netlink/specs/netdev.yaml | 61 +++
Documentation/networking/devmem.rst | 269 +++++++++++
Documentation/networking/index.rst | 1 +
arch/alpha/include/uapi/asm/socket.h | 6 +
arch/mips/include/uapi/asm/socket.h | 6 +
arch/parisc/include/uapi/asm/socket.h | 6 +
arch/sparc/include/uapi/asm/socket.h | 6 +
include/linux/netdevice.h | 2 +
include/linux/skbuff.h | 61 ++-
include/linux/skbuff_ref.h | 9 +-
include/linux/socket.h | 1 +
include/net/netdev_rx_queue.h | 5 +
include/net/netmem.h | 132 +++++-
include/net/page_pool/helpers.h | 39 +-
include/net/page_pool/types.h | 23 +-
include/net/sock.h | 2 +
include/net/tcp.h | 3 +-
include/trace/events/page_pool.h | 12 +-
include/uapi/asm-generic/socket.h | 6 +
include/uapi/linux/netdev.h | 13 +
include/uapi/linux/uio.h | 18 +
net/Kconfig | 5 +
net/core/Makefile | 2 +
net/core/datagram.c | 6 +
net/core/dev.c | 33 +-
net/core/devmem.c | 389 ++++++++++++++++
net/core/devmem.h | 180 ++++++++
net/core/gro.c | 3 +-
net/core/mp_dmabuf_devmem.h | 44 ++
net/core/netdev-genl-gen.c | 23 +
net/core/netdev-genl-gen.h | 6 +
net/core/netdev-genl.c | 139 +++++-
net/core/netdev_rx_queue.c | 81 ++++
net/core/netmem_priv.h | 31 ++
net/core/page_pool.c | 119 +++--
net/core/page_pool_priv.h | 46 ++
net/core/page_pool_user.c | 32 +-
net/core/skbuff.c | 77 +++-
net/core/sock.c | 68 +++
net/ethtool/common.c | 8 +
net/ipv4/esp4.c | 3 +-
net/ipv4/tcp.c | 263 ++++++++++-
net/ipv4/tcp_input.c | 13 +-
net/ipv4/tcp_ipv4.c | 16 +
net/ipv4/tcp_minisocks.c | 2 +
net/ipv4/tcp_output.c | 5 +-
net/ipv6/esp6.c | 3 +-
net/packet/af_packet.c | 4 +-
net/xdp/xsk_buff_pool.c | 5 +
tools/include/uapi/linux/netdev.h | 13 +
tools/net/ynl/lib/.gitignore | 1 +
tools/testing/selftests/net/.gitignore | 1 +
tools/testing/selftests/net/Makefile | 9 +
tools/testing/selftests/net/ncdevmem.c | 570 ++++++++++++++++++++++++
54 files changed, 2757 insertions(+), 124 deletions(-)
create mode 100644 Documentation/networking/devmem.rst
create mode 100644 net/core/devmem.c
create mode 100644 net/core/devmem.h
create mode 100644 net/core/mp_dmabuf_devmem.h
create mode 100644 net/core/netdev_rx_queue.c
create mode 100644 net/core/netmem_priv.h
create mode 100644 tools/testing/selftests/net/ncdevmem.c
--
2.46.0.598.g6f2099f65c-goog
Here are some various fixes for the MPTCP selftests.
Patch 1 fixes a recently modified test to continue to work as expected
on older kernels. This is a fix for a recent fix that can be backported
up to v5.15.
Patch 2 and 3 include dependences when exporting or installing the
tests. Two fixes for v6.11-rc1.
Signed-off-by: Matthieu Baerts (NGI0) <matttbe(a)kernel.org>
---
Matthieu Baerts (NGI0) (3):
selftests: mptcp: join: restrict fullmesh endp on 1st sf
selftests: mptcp: include lib.sh file
selftests: mptcp: include net_helper.sh file
tools/testing/selftests/net/mptcp/Makefile | 2 ++
tools/testing/selftests/net/mptcp/mptcp_join.sh | 4 +++-
2 files changed, 5 insertions(+), 1 deletion(-)
---
base-commit: 48aa361c5db0b380c2b75c24984c0d3e7c1e8c09
change-id: 20240910-net-selftests-mptcp-fix-install-2b82ae5a99c8
Best regards,
--
Matthieu Baerts (NGI0) <matttbe(a)kernel.org>
Handle the case where the meta-page content is bigger than the system
page-size. This prepares the ground for extending features covered by
the meta-page.
Cc: Shuah Khan <skhan(a)linuxfoundation.org>
Cc: linux-kselftest(a)vger.kernel.org
Signed-off-by: Vincent Donnefort <vdonnefort(a)google.com>
diff --git a/tools/testing/selftests/ring-buffer/map_test.c b/tools/testing/selftests/ring-buffer/map_test.c
index ba12fd31de87..d10a847130fb 100644
--- a/tools/testing/selftests/ring-buffer/map_test.c
+++ b/tools/testing/selftests/ring-buffer/map_test.c
@@ -92,12 +92,22 @@ int tracefs_cpu_map(struct tracefs_cpu_map_desc *desc, int cpu)
if (desc->cpu_fd < 0)
return -ENODEV;
+again:
map = mmap(NULL, page_size, PROT_READ, MAP_SHARED, desc->cpu_fd, 0);
if (map == MAP_FAILED)
return -errno;
desc->meta = (struct trace_buffer_meta *)map;
+ /* the meta-page is bigger than the original mapping */
+ if (page_size < desc->meta->meta_struct_len) {
+ int meta_page_size = desc->meta->meta_page_size;
+
+ munmap(desc->meta, page_size);
+ page_size = meta_page_size;
+ goto again;
+ }
+
return 0;
}
--
2.46.0.598.g6f2099f65c-goog
Two s390 fixes to make vdso selftests running on s390.
Jason, given that you carry already a lot of changes for vdso
selftests I guess these should be routed via the random tree.
Patches apply on top of current random.git master branch.
Thanks,
Heiko
Heiko Carstens (1):
selftests: vDSO: fix vdso_config for s390
Jens Remus (1):
selftests: vDSO: fix ELF hash table entry size for s390x
tools/testing/selftests/vDSO/parse_vdso.c | 14 ++++++++++----
tools/testing/selftests/vDSO/vdso_config.h | 4 ++--
2 files changed, 12 insertions(+), 6 deletions(-)
--
2.43.0
Padding is not included in UDP and TCP checksums. Therefore, reduce the
length of the checksummed data to include only the data in the IP
payload. This fixes spurious reported checksum failures like
rx: pkt: sport=33000 len=26 csum=0xc850 verify=0xf9fe
pkt: bad csum
Technically it is possible for there to be trailing bytes after the UDP
data but before the Ethernet padding (e.g. if sizeof(ip) + sizeof(udp) +
udp.len < ip.len). However, we don't generate such packets.
Fixes: 91a7de85600d ("selftests/net: add csum offload test")
Signed-off-by: Sean Anderson <sean.anderson(a)linux.dev>
---
Found while testing for this very bug in hardware checksum offloads.
tools/testing/selftests/net/lib/csum.c | 16 ++++++++++++++--
1 file changed, 14 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/net/lib/csum.c b/tools/testing/selftests/net/lib/csum.c
index b9f3fc3c3426..e0a34e5e8dd5 100644
--- a/tools/testing/selftests/net/lib/csum.c
+++ b/tools/testing/selftests/net/lib/csum.c
@@ -654,10 +654,16 @@ static int recv_verify_packet_ipv4(void *nh, int len)
{
struct iphdr *iph = nh;
uint16_t proto = cfg_encap ? IPPROTO_UDP : cfg_proto;
+ uint16_t ip_len;
if (len < sizeof(*iph) || iph->protocol != proto)
return -1;
+ ip_len = ntohs(iph->tot_len);
+ if (ip_len > len || ip_len < sizeof(*iph))
+ return -1;
+
+ len = ip_len;
iph_addr_p = &iph->saddr;
if (proto == IPPROTO_TCP)
return recv_verify_packet_tcp(iph + 1, len - sizeof(*iph));
@@ -669,16 +675,22 @@ static int recv_verify_packet_ipv6(void *nh, int len)
{
struct ipv6hdr *ip6h = nh;
uint16_t proto = cfg_encap ? IPPROTO_UDP : cfg_proto;
+ uint16_t ip_len;
if (len < sizeof(*ip6h) || ip6h->nexthdr != proto)
return -1;
+ ip_len = ntohs(ip6h->payload_len);
+ if (ip_len > len - sizeof(*ip6h))
+ return -1;
+
+ len = ip_len;
iph_addr_p = &ip6h->saddr;
if (proto == IPPROTO_TCP)
- return recv_verify_packet_tcp(ip6h + 1, len - sizeof(*ip6h));
+ return recv_verify_packet_tcp(ip6h + 1, len);
else
- return recv_verify_packet_udp(ip6h + 1, len - sizeof(*ip6h));
+ return recv_verify_packet_udp(ip6h + 1, len);
}
/* return whether auxdata includes TP_STATUS_CSUM_VALID */
--
2.35.1.1320.gc452695387.dirty
From: Willem de Bruijn <willemb(a)google.com>
Lay the groundwork to import into kselftests the over 150 packetdrill
TCP/IP conformance tests on github.com/google/packetdrill.
1/2: add kselftest infra for TEST_PROGS that need an interpreter
2/2: add the specific packetdrill tests
Both can go through net-next, I imagine. But let me know if the
core infra should go through linux-kselftest.
Willem de Bruijn (2):
selftests: support interpreted scripts with ksft_runner.sh
selftests/net: integrate packetdrill with ksft
tools/testing/selftests/Makefile | 5 +-
tools/testing/selftests/kselftest/runner.sh | 7 ++-
.../selftests/net/packetdrill/Makefile | 9 +++
.../testing/selftests/net/packetdrill/config | 5 ++
.../selftests/net/packetdrill/defaults.sh | 63 +++++++++++++++++++
.../selftests/net/packetdrill/ksft_runner.sh | 41 ++++++++++++
.../net/packetdrill/tcp_inq_client.pkt | 51 +++++++++++++++
.../net/packetdrill/tcp_inq_server.pkt | 51 +++++++++++++++
.../tcp_md5_md5-only-on-client-ack.pkt | 28 +++++++++
9 files changed, 256 insertions(+), 4 deletions(-)
create mode 100644 tools/testing/selftests/net/packetdrill/Makefile
create mode 100644 tools/testing/selftests/net/packetdrill/config
create mode 100755 tools/testing/selftests/net/packetdrill/defaults.sh
create mode 100755 tools/testing/selftests/net/packetdrill/ksft_runner.sh
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_inq_client.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_inq_server.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_md5_md5-only-on-client-ack.pkt
--
2.46.0.469.g59c65b2a67-goog
From: Joshua Hahn <joshua.hahn6(a)gmail.com>
v1 -> v2: Edited commit messages for clarity.
Niced CPU usage is a metric reported in host-level /prot/stat, but is
not reported in cgroup-level statistics in cpu.stat. However, when a
host contains multiple tasks across different workloads, it becomes
difficult to gauge how much of the task is being spent on niced
processes based on /proc/stat alone, since host-level metrics do not
provide this cgroup-level granularity.
Exposing this metric will allow users to accurately probe the niced CPU
metric for each workload, and make more informed decisions when
directing higher priority tasks.
Joshua Hahn (2):
Tracking cgroup-level niced CPU time
Selftests for niced CPU statistics
include/linux/cgroup-defs.h | 1 +
kernel/cgroup/rstat.c | 16 ++++-
tools/testing/selftests/cgroup/test_cpu.c | 72 +++++++++++++++++++++++
3 files changed, 86 insertions(+), 3 deletions(-)
--
2.43.5
In this series from Geliang, modifying MPTCP BPF selftests, we have:
- A new MPTCP subflow BPF program setting socket options per subflow: it
looks better to have this old test program in the BPF selftests to
track regressions and to serve as example.
Note: Nicolas is no longer working at Tessares, but he did this work
while working for them, and his email address is no longer available.
- A new hook in the same BPF program to do the verification step.
- A new MPTCP BPF subtest validating the new BPF program added in the
first patch, with the help of the new hook added in the second patch.
Signed-off-by: Matthieu Baerts (NGI0) <matttbe(a)kernel.org>
---
Changes in v5:
- See the individual changelog for more details about them
- Patch 1/3: set TCP on the 2nd subflow
- Patch 2/3: new
- Patch 3/3: use the BPF program from patch 2/3 to do the validation
instead of using ss.
- Link to v4: https://lore.kernel.org/r/20240805-upstream-bpf-next-20240506-mptcp-subflow…
Changes in v4:
- Drop former patch 2/3: MPTCP's pm_nl_ctl requires a new header file:
- I will check later if it is possible to avoid having duplicated
header files in tools/include/uapi, but no need to block this series
for that. Patch 2/3 can be added later if needed.
- Patch 2/2: skip the test if 'ip mptcp' is not available.
- Link to v3: https://lore.kernel.org/r/20240703-upstream-bpf-next-20240506-mptcp-subflow…
Changes in v3:
- Sorry for the delay between v2 and v3, this series was conflicting
with the "add netns helpers", but it looks like it is on hold:
https://lore.kernel.org/cover.1715821541.git.tanggeliang@kylinos.cn
- Patch 1/3 includes "bpf_tracing_net.h", introduced in between.
- New patch 2/3: "selftests/bpf: Add mptcp pm_nl_ctl link".
- Patch 3/3: use the tool introduced in patch 2/3 + SYS_NOFAIL() helper.
- Link to v2: https://lore.kernel.org/r/20240509-upstream-bpf-next-20240506-mptcp-subflow…
Changes in v2:
- Previous patches 1/4 and 2/4 have been dropped from this series:
- 1/4: "selftests/bpf: Handle SIGINT when creating netns":
- A new version, more generic and no longer specific to MPTCP BPF
selftest will be sent later, as part of a new series. (Alexei)
- 2/4: "selftests/bpf: Add RUN_MPTCP_TEST macro":
- Removed, not to hide helper functions in macros. (Alexei)
- The commit message of patch 1/2 has been clarified to avoid some
possible confusions spot by Alexei.
- Link to v1: https://lore.kernel.org/r/20240507-upstream-bpf-next-20240506-mptcp-subflow…
---
Geliang Tang (2):
selftests/bpf: Add getsockopt to inspect mptcp subflow
selftests/bpf: Add mptcp subflow subtest
Nicolas Rybowski (1):
selftests/bpf: Add mptcp subflow example
MAINTAINERS | 2 +-
tools/testing/selftests/bpf/prog_tests/mptcp.c | 126 +++++++++++++++++++++
tools/testing/selftests/bpf/progs/mptcp_bpf.h | 42 +++++++
tools/testing/selftests/bpf/progs/mptcp_subflow.c | 128 ++++++++++++++++++++++
4 files changed, 297 insertions(+), 1 deletion(-)
---
base-commit: 6b083650a37318112fb60c65fbb6070584f53d93
change-id: 20240506-upstream-bpf-next-20240506-mptcp-subflow-test-faef6654bfa3
Best regards,
--
Matthieu Baerts (NGI0) <matttbe(a)kernel.org>
rxq contains a pointer to the device from where
the redirect happened. Currently, the BPF program
that was executed after a redirect via BPF_MAP_TYPE_DEVMAP*
does not have it set.
Add bugfix and related selftest.
Signed-off-by: Florian Kauer <florian.kauer(a)linutronix.de>
---
Changes in v3:
- initialize skel to NULL, thanks Stanislav
- Link to v2: https://lore.kernel.org/r/20240906-devel-koalo-fix-ingress-ifindex-v2-0-4ca…
Changes in v2:
- changed fixes tag
- added selftest
- Link to v1: https://lore.kernel.org/r/20240905-devel-koalo-fix-ingress-ifindex-v1-1-d12…
---
Florian Kauer (2):
bpf: devmap: provide rxq after redirect
bpf: selftests: send packet to devmap redirect XDP
kernel/bpf/devmap.c | 11 +-
.../selftests/bpf/prog_tests/xdp_devmap_attach.c | 114 +++++++++++++++++++--
2 files changed, 115 insertions(+), 10 deletions(-)
---
base-commit: 8e69c96df771ab469cec278edb47009351de4da6
change-id: 20240905-devel-koalo-fix-ingress-ifindex-b9293d471db6
Best regards,
--
Florian Kauer <florian.kauer(a)linutronix.de>
Mending test for list_cut_position*() for the missing check of integer
"i" after the second loop. The variable should be checked for second
time to make sure both lists after the cut operation are formed as
expected.
Signed-off-by: I Hsin Cheng <richard120310(a)gmail.com>
---
lib/list-test.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/lib/list-test.c b/lib/list-test.c
index 37cbc33e9fdb..f59188fc2aca 100644
--- a/lib/list-test.c
+++ b/lib/list-test.c
@@ -404,10 +404,13 @@ static void list_test_list_cut_position(struct kunit *test)
KUNIT_EXPECT_EQ(test, i, 2);
+ i = 0;
list_for_each(cur, &list1) {
KUNIT_EXPECT_PTR_EQ(test, cur, &entries[i]);
i++;
}
+
+ KUNIT_EXPECT_EQ(test, i, 1);
}
static void list_test_list_cut_before(struct kunit *test)
@@ -432,10 +435,13 @@ static void list_test_list_cut_before(struct kunit *test)
KUNIT_EXPECT_EQ(test, i, 1);
+ i = 0;
list_for_each(cur, &list1) {
KUNIT_EXPECT_PTR_EQ(test, cur, &entries[i]);
i++;
}
+
+ KUNIT_EXPECT_EQ(i, 2);
}
static void list_test_list_splice(struct kunit *test)
--
2.43.0
From: Jason Xing <kernelxing(a)tencent.com>
When I was trying to modify the tx timestamping feature, I found that
running "./txtimestamp -4 -C -L 127.0.0.1" didn't reflect the error:
I succeeded to generate timestamp stored in the skb but later failed
to report it to the userspace (which means failed to put css into cmsg).
It can happen when someone writes buggy codes in __sock_recv_timestamp(),
for example.
After adding the check so that running ./txtimestamp will reflect the
result correctly like this if there is a bug in the reporting phase:
protocol: TCP
payload: 10
server port: 9000
family: INET
test SND
USR: 1725458477 s 667997 us (seq=0, len=0)
Failed to report timestamps
USR: 1725458477 s 718128 us (seq=0, len=0)
Failed to report timestamps
USR: 1725458477 s 768273 us (seq=0, len=0)
Failed to report timestamps
USR: 1725458477 s 818416 us (seq=0, len=0)
Failed to report timestamps
...
In the future, it will help us detect whether the new coming patch has
bugs or not.
Signed-off-by: Jason Xing <kernelxing(a)tencent.com>
---
v2
Link: https://lore.kernel.org/all/20240904144446.41274-1-kerneljasonxing@gmail.co…
1. mainly change from "parse" to "report", update the commit message.
---
tools/testing/selftests/net/txtimestamp.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/tools/testing/selftests/net/txtimestamp.c b/tools/testing/selftests/net/txtimestamp.c
index ec60a16c9307..d626f22f9550 100644
--- a/tools/testing/selftests/net/txtimestamp.c
+++ b/tools/testing/selftests/net/txtimestamp.c
@@ -356,8 +356,12 @@ static void __recv_errmsg_cmsg(struct msghdr *msg, int payload_len)
}
}
- if (batch > 1)
+ if (batch > 1) {
fprintf(stderr, "batched %d timestamps\n", batch);
+ } else if (!batch) {
+ fprintf(stderr, "Failed to report timestamps\n");
+ test_failed = true;
+ }
}
static int recv_errmsg(int fd)
--
2.37.3
Patches here add 'time=<N>ms' in the diagnostic data of the TAP output,
e.g.
ok 1 - pm_netlink: defaults addr list # time=9ms
This addition is useful to quickly identify which subtests are taking a
longer time than the others, or more than expected.
Note that there are no specific formats to follow to show this time
according to the TAP 13, TAP 14 and KTAP specifications, but we follow
the format being parsed by NIPA [1].
Patch 1 modifies mptcp_lib.sh to add this support to all MPTCP
selftests.
Patch 2 removes the now duplicated info in mptcp_connect.sh
Patch 3 slightly improves the precision of the first subtests in all
MPTCP subtests.
Patches 4 and 5 remove duplicated spaces in TAP output, for the TAP
parsers that cannot handle them properly.
Link: https://github.com/linux-netdev/nipa/pull/36
Signed-off-by: Matthieu Baerts (NGI0) <matttbe(a)kernel.org>
---
Changes in v2:
- Typo in the commit message of patch 2 (Jakub)
- Two additional patches to remove duplicated spaces in TAP output
- Link to v1: https://lore.kernel.org/r/20240902-net-next-mptcp-ksft-subtest-time-v1-0-f1…
---
Matthieu Baerts (NGI0) (5):
selftests: mptcp: lib: add time per subtests in TAP output
selftests: mptcp: connect: remote time in TAP output
selftests: mptcp: reset the last TS before the first test
selftests: mptcp: diag: remove trailing whitespace
selftests: mptcp: connect: remove duplicated spaces in TAP output
tools/testing/selftests/net/mptcp/diag.sh | 2 +-
tools/testing/selftests/net/mptcp/mptcp_connect.sh | 17 ++++++++++-------
tools/testing/selftests/net/mptcp/mptcp_join.sh | 3 ++-
tools/testing/selftests/net/mptcp/mptcp_lib.sh | 17 ++++++++++++++++-
tools/testing/selftests/net/mptcp/mptcp_sockopt.sh | 1 +
tools/testing/selftests/net/mptcp/pm_netlink.sh | 2 ++
tools/testing/selftests/net/mptcp/simult_flows.sh | 1 +
tools/testing/selftests/net/mptcp/userspace_pm.sh | 1 +
8 files changed, 34 insertions(+), 10 deletions(-)
---
base-commit: 52fc70a32573707f70d6b1b5c5fe85cc91457393
change-id: 20240902-net-next-mptcp-ksft-subtest-time-a83cec43d894
Best regards,
--
Matthieu Baerts (NGI0) <matttbe(a)kernel.org>
When "arg#%d expected pointer to ctx, but got %s" error is printed, both
template parts actually point to the type of the argument, therefore, it
will also say "but got PTR", regardless of what was the actual register
type.
Fix the message to print the register type in the second part of the
template, change the existing test to adapt to the new format, and add a
new test to test the case when arg is a pointer to context, but reg is a
scalar.
Fixes: 00b85860feb8 ("bpf: Rewrite kfunc argument handling")
Signed-off-by: Maxim Mikityanskiy <maxim(a)isovalent.com>
---
kernel/bpf/verifier.c | 3 ++-
tools/testing/selftests/bpf/prog_tests/kfunc_call.c | 1 +
tools/testing/selftests/bpf/progs/kfunc_call_fail.c | 7 +++++++
tools/testing/selftests/bpf/verifier/calls.c | 2 +-
4 files changed, 11 insertions(+), 2 deletions(-)
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index d8520095ca03..8b9f0a2981d4 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -11948,7 +11948,8 @@ static int check_kfunc_args(struct bpf_verifier_env *env, struct bpf_kfunc_call_
switch (kf_arg_type) {
case KF_ARG_PTR_TO_CTX:
if (reg->type != PTR_TO_CTX) {
- verbose(env, "arg#%d expected pointer to ctx, but got %s\n", i, btf_type_str(t));
+ verbose(env, "arg#%d expected pointer to ctx, but got %s\n",
+ i, reg_type_str(env, reg->type));
return -EINVAL;
}
diff --git a/tools/testing/selftests/bpf/prog_tests/kfunc_call.c b/tools/testing/selftests/bpf/prog_tests/kfunc_call.c
index 5b743212292f..f79c8e53cb3e 100644
--- a/tools/testing/selftests/bpf/prog_tests/kfunc_call.c
+++ b/tools/testing/selftests/bpf/prog_tests/kfunc_call.c
@@ -68,6 +68,7 @@ static struct kfunc_test_params kfunc_tests[] = {
TC_FAIL(kfunc_call_test_get_mem_fail_oob, 0, "min value is outside of the allowed memory range"),
TC_FAIL(kfunc_call_test_get_mem_fail_not_const, 0, "is not a const"),
TC_FAIL(kfunc_call_test_mem_acquire_fail, 0, "acquire kernel function does not return PTR_TO_BTF_ID"),
+ TC_FAIL(kfunc_call_test_pointer_arg_type_mismatch, 0, "arg#0 expected pointer to ctx, but got scalar"),
/* success cases */
TC_TEST(kfunc_call_test1, 12),
diff --git a/tools/testing/selftests/bpf/progs/kfunc_call_fail.c b/tools/testing/selftests/bpf/progs/kfunc_call_fail.c
index 4b0b7b79cdfb..08fae306539c 100644
--- a/tools/testing/selftests/bpf/progs/kfunc_call_fail.c
+++ b/tools/testing/selftests/bpf/progs/kfunc_call_fail.c
@@ -150,4 +150,11 @@ int kfunc_call_test_mem_acquire_fail(struct __sk_buff *skb)
return ret;
}
+SEC("?tc")
+int kfunc_call_test_pointer_arg_type_mismatch(struct __sk_buff *skb)
+{
+ bpf_kfunc_call_test_pass_ctx((void *)10);
+ return 0;
+}
+
char _license[] SEC("license") = "GPL";
diff --git a/tools/testing/selftests/bpf/verifier/calls.c b/tools/testing/selftests/bpf/verifier/calls.c
index d0cdd156cd55..7afc2619ab14 100644
--- a/tools/testing/selftests/bpf/verifier/calls.c
+++ b/tools/testing/selftests/bpf/verifier/calls.c
@@ -76,7 +76,7 @@
},
.prog_type = BPF_PROG_TYPE_SCHED_CLS,
.result = REJECT,
- .errstr = "arg#0 expected pointer to ctx, but got PTR",
+ .errstr = "arg#0 expected pointer to ctx, but got fp",
.fixup_kfunc_btf_id = {
{ "bpf_kfunc_call_test_pass_ctx", 2 },
},
--
2.46.0
Hi all,
This is part of a hackathon organized by LKCAMP[1], focused on writing
tests using KUnit. We reached out a while ago asking for advice on what
would be a useful contribution[2] and ended up choosing data structures
that did not yet have tests.
This patch adds tests for the llist data structure, defined in
include/linux/llist.h, and is inspired by the KUnit tests for the doubly
linked list in lib/list-test.c[3].
It is important to note that this patch depends on the patch referenced
in [4], as it utilizes the newly created lib/tests/ subdirectory.
[1] https://lkcamp.dev/about/
[2] https://lore.kernel.org/all/Zktnt7rjKryTh9-N@arch/
[3] https://elixir.bootlin.com/linux/latest/source/lib/list-test.c
[4] https://lore.kernel.org/all/20240720181025.work.002-kees@kernel.org/
---
Changes in v2:
- Add MODULE_DESCRIPTION()
- Move the tests from lib/llist_kunit.c to lib/tests/llist_kunit.c
- Change the license from "GPL v2" to "GPL"
Artur Alves (1):
lib/llist_kunit.c: add KUnit tests for llist
lib/Kconfig.debug | 11 ++
lib/tests/Makefile | 1 +
lib/tests/llist_kunit.c | 361 ++++++++++++++++++++++++++++++++++++++++
3 files changed, 373 insertions(+)
create mode 100644 lib/tests/llist_kunit.c
--
2.46.0
This series of patches are for testing the lan743x network driver.
Testing comprises autonegotiation, speed, duplex and throughput checks.
Tools such as ethtool, iperf3 are used in the testing process.
Performance test is done for TCP streams at different speeds.
Signed-off-by: Mohan Prasad J <mohan.prasad(a)microchip.com>
Mohan Prasad J (3):
selftests: lan743x: Add testfile for lan743x network driver
selftests: lan743x: Add testcase to check speed and duplex state of
lan743x
selftests: lan743x: Add testcase to check throughput of lan743x
MAINTAINERS | 2 +
tools/testing/selftests/Makefile | 2 +-
.../drivers/net/hw/microchip/lan743x/Makefile | 7 ++
.../net/hw/microchip/lan743x/lan743x.py | 117 ++++++++++++++++++
.../hw/microchip/lan743x/lib/py/__init__.py | 16 +++
5 files changed, 143 insertions(+), 1 deletion(-)
create mode 100644 tools/testing/selftests/drivers/net/hw/microchip/lan743x/Makefile
create mode 100755 tools/testing/selftests/drivers/net/hw/microchip/lan743x/lan743x.py
create mode 100644 tools/testing/selftests/drivers/net/hw/microchip/lan743x/lib/py/__init__.py
--
2.43.0
The resctrl selftests for Memory Bandwidth Allocation (MBA) and Memory
Bandwidth Monitoring (MBM) are failing on some (for example [1]) Emerald
Rapids systems. The test failures result from the following two
properties of these systems:
1) Emerald Rapids systems can have up to 320MB L3 cache. The resctrl
MBA and MBM selftests measure memory traffic for which a hardcoded
250MB buffer has been sufficient so far. On platforms with L3 cache
larger than the buffer, the buffer fits in the L3 cache and thus
no/very little memory traffic is generated during the "memory
bandwidth" tests.
2) Some platform features, for example RAS features or memory
performance features that generate memory traffic may drive accesses
that are counted differently by performance counters and MBM
respectively, for instance generating "overhead" traffic which is not
counted against any specific RMID. Until now these counting
differences have always been "in the noise". On Emerald Rapids
systems the maximum MBA throttling (10% memory bandwidth)
throttles memory bandwidth to where memory accesses by these other
platform features push the memory bandwidth difference between
memory controller performance counters and resctrl (MBM) beyond the
tests' hardcoded tolerance.
Make the tests more robust against platform variations:
1) Let the buffer used by memory bandwidth tests be guided by the size
of the L3 cache.
2) Larger buffers require longer initialization time before the buffer can
be used to measurement. Rework the tests to ensure that buffer
initialization is complete before measurements start.
3) Do not compare performance counters and MBM measurements at low
bandwidth. The value of "low" is hardcoded to 750MiB based on
measurements on Emerald Rapids, Sapphire Rapids, and Ice Lake
systems. This limit is not applicable to AMD systems since it
only applies to the MBA and MBM tests that are isolated to Intel.
[1]
https://ark.intel.com/content/www/us/en/ark/products/237261/intel-xeon-plat…
Reinette Chatre (6):
selftests/resctrl: Fix sparse warnings
selftests/resctrl: Ensure measurements skip initialization of default
benchmark
selftests/resctrl: Simplify benchmark parameter passing
selftests/resctrl: Use cache size to determine "fill_buf" buffer size
selftests/resctrl: Do not compare performance counters and resctrl at
low bandwidth
selftests/resctrl: Keep results from first test run
tools/testing/selftests/resctrl/cmt_test.c | 33 +--
tools/testing/selftests/resctrl/fill_buf.c | 19 +-
tools/testing/selftests/resctrl/mba_test.c | 26 +-
tools/testing/selftests/resctrl/mbm_test.c | 25 +-
tools/testing/selftests/resctrl/resctrl.h | 57 +++--
.../testing/selftests/resctrl/resctrl_tests.c | 15 +-
tools/testing/selftests/resctrl/resctrl_val.c | 223 +++++-------------
7 files changed, 152 insertions(+), 246 deletions(-)
--
2.46.0
There were several attempts to resolve circular include dependency
after the addition of percpu.h: 1c9df907da83 ("random: fix circular
include dependency on arm64 after addition of percpu.h"), c0842fbc1b18
("random32: move the pseudo-random 32-bit definitions to prandom.h") and
finally d9f29deb7fe8 ("prandom: Remove unused include") that completely
removes the inclusion of <linux/percpu.h>.
Due to legacy reasons, <linux/random.h> includes <linux/prandom.h>, but
with the commit entry remark:
--quote--
A further cleanup step would be to remove this from <linux/random.h>
entirely, and make people who use the prandom infrastructure include
just the new header file. That's a bit of a churn patch, but grepping
for "prandom_" and "next_pseudo_random32" "struct rnd_state" should
catch most users.
But it turns out that that nice cleanup step is fairly painful, because
a _lot_ of code currently seems to depend on the implicit include of
<linux/random.h>, which can currently come in a lot of ways, including
such fairly core headfers as <linux/net.h>.
So the "nice cleanup" part may or may never happen.
--/quote--
We would like to include <linux/percpu.h> in <linux/prandom.h>.
In [1] we would like to repurpose __percpu tag as a named address space
qualifier, where __percpu macro uses defines from <linux/percpu.h>.
The major roadblock to inclusion of <linux/percpu.h> is the above
mentioned legacy inclusion of <linux/prandom.h> in <linux/random.h> that
causes circular include dependency that prevents <linux/percpu.h>
inclusion.
This patch series is the "nice cleanup" part that:
a) Substitutes the inclusion of <linux/random.h> with the
inclusion of <linux/prandom.h> where needed (patches 1 - 17).
b) Removes legacy inclusion of <linux/prandom.h> from
<linux/random.h> (patch 18).
c) Includes <linux/percpu.h> in <linux/prandom.h> (patch 19).
The whole series was tested by compiling the kernel for x86_64 allconfig
and some popular architectures, namely arm64 defconfig, powerpc defconfig
and loongarch defconfig.
[1] https://lore.kernel.org/lkml/20240812115945.484051-4-ubizjak@gmail.com/
Cc: Dave Hansen <dave.hansen(a)linux.intel.com>
Cc: Andy Lutomirski <luto(a)kernel.org>
Cc: Peter Zijlstra <peterz(a)infradead.org>
Cc: Thomas Gleixner <tglx(a)linutronix.de>
Cc: Ingo Molnar <mingo(a)redhat.com>
Cc: Borislav Petkov <bp(a)alien8.de>
Cc: x86(a)kernel.org
Cc: "H. Peter Anvin" <hpa(a)zytor.com>
Cc: Jani Nikula <jani.nikula(a)linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen(a)linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi(a)intel.com>
Cc: Tvrtko Ursulin <tursulin(a)ursulin.net>
Cc: David Airlie <airlied(a)gmail.com>
Cc: Daniel Vetter <daniel(a)ffwll.ch>
Cc: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com>
Cc: Maxime Ripard <mripard(a)kernel.org>
Cc: Thomas Zimmermann <tzimmermann(a)suse.de>
Cc: Hans Verkuil <hverkuil(a)xs4all.nl>
Cc: Mauro Carvalho Chehab <mchehab(a)kernel.org>
Cc: Miquel Raynal <miquel.raynal(a)bootlin.com>
Cc: Richard Weinberger <richard(a)nod.at>
Cc: Vignesh Raghavendra <vigneshr(a)ti.com>
Cc: Eric Biggers <ebiggers(a)kernel.org>
Cc: "Theodore Y. Ts'o" <tytso(a)mit.edu>
Cc: Jaegeuk Kim <jaegeuk(a)kernel.org>
Cc: "Jason A. Donenfeld" <Jason(a)zx2c4.com>
Cc: Linus Torvalds <torvalds(a)linux-foundation.org>
Cc: Hannes Reinecke <hare(a)suse.de>
Cc: "James E.J. Bottomley" <James.Bottomley(a)HansenPartnership.com>
Cc: "Martin K. Petersen" <martin.petersen(a)oracle.com>
Cc: Alexei Starovoitov <ast(a)kernel.org>
Cc: Daniel Borkmann <daniel(a)iogearbox.net>
Cc: John Fastabend <john.fastabend(a)gmail.com>
Cc: Andrii Nakryiko <andrii(a)kernel.org>
Cc: Martin KaFai Lau <martin.lau(a)linux.dev>
Cc: Eduard Zingerman <eddyz87(a)gmail.com>
Cc: Song Liu <song(a)kernel.org>
Cc: Yonghong Song <yonghong.song(a)linux.dev>
Cc: KP Singh <kpsingh(a)kernel.org>
Cc: Stanislav Fomichev <sdf(a)fomichev.me>
Cc: Hao Luo <haoluo(a)google.com>
Cc: Jiri Olsa <jolsa(a)kernel.org>
Cc: Andrew Morton <akpm(a)linux-foundation.org>
Cc: Brendan Higgins <brendan.higgins(a)linux.dev>
Cc: David Gow <davidgow(a)google.com>
Cc: Rae Moar <rmoar(a)google.com>
Cc: "David S. Miller" <davem(a)davemloft.net>
Cc: Eric Dumazet <edumazet(a)google.com>
Cc: Jakub Kicinski <kuba(a)kernel.org>
Cc: Paolo Abeni <pabeni(a)redhat.com>
Cc: Jiri Pirko <jiri(a)resnulli.us>
Cc: Petr Mladek <pmladek(a)suse.com>
Cc: Steven Rostedt <rostedt(a)goodmis.org>
Cc: Andy Shevchenko <andriy.shevchenko(a)linux.intel.com>
Cc: Rasmus Villemoes <linux(a)rasmusvillemoes.dk>
Cc: Sergey Senozhatsky <senozhatsky(a)chromium.org>
Cc: Stephen Hemminger <stephen(a)networkplumber.org>
Cc: Jamal Hadi Salim <jhs(a)mojatatu.com>
Cc: Cong Wang <xiyou.wangcong(a)gmail.com>
Cc: Kent Overstreet <kent.overstreet(a)linux.dev>
---
v2: - Reword commit messages to mention the removal of legacy inclusion
of <linux/prandom.h> from <linux/random.h>
- Add missing substitution in crypto/testmgr.c
(reported by kernel test robot)
- Add Acked-by:.
Uros Bizjak (19):
x86/kaslr: Include <linux/prandom.h> instead of <linux/random.h>
crypto: testmgr: Include <linux/prandom.h> instead of <linux/random.h>
drm/i915/selftests: Include <linux/prandom.h> instead of
<linux/random.h>
drm/lib: Include <linux/prandom.h> instead of <linux/random.h>
media: vivid: Include <linux/prandom.h> in vivid-vid-cap.c
mtd: tests: Include <linux/prandom.h> instead of <linux/random.h>
fscrypt: Include <linux/once.h> in fs/crypto/keyring.c
scsi: libfcoe: Include <linux/prandom.h> instead of <linux/random.h>
bpf: Include <linux/prandom.h> instead of <linux/random.h>
lib/interval_tree_test.c: Include <linux/prandom.h> instead of
<linux/random.h>
kunit: string-stream-test: Include <linux/prandom.h> instead of
<linux/random.h>
random32: Include <linux/prandom.h> instead of <linux/random.h>
lib/rbtree-test: Include <linux/prandom.h> instead of <linux/random.h>
bpf/tests: Include <linux/prandom.h> instead of <linux/random.h>
lib/test_parman: Include <linux/prandom.h> instead of <linux/random.h>
lib/test_scanf: Include <linux/prandom.h> instead of <linux/random.h>
netem: Include <linux/prandom.h> in sch_netem.c
random: Do not include <linux/prandom.h> in <linux/random.h>
prandom: Include <linux/percpu.h> in <linux/prandom.h>
arch/x86/mm/kaslr.c | 2 +-
crypto/testmgr.c | 2 +-
drivers/gpu/drm/i915/selftests/i915_gem.c | 2 +-
drivers/gpu/drm/i915/selftests/i915_random.h | 2 +-
drivers/gpu/drm/i915/selftests/scatterlist.c | 2 +-
drivers/gpu/drm/lib/drm_random.h | 2 +-
drivers/media/test-drivers/vivid/vivid-vid-cap.c | 1 +
drivers/mtd/tests/oobtest.c | 2 +-
drivers/mtd/tests/pagetest.c | 2 +-
drivers/mtd/tests/subpagetest.c | 2 +-
fs/crypto/keyring.c | 1 +
include/linux/prandom.h | 1 +
include/linux/random.h | 7 -------
include/scsi/libfcoe.h | 2 +-
kernel/bpf/core.c | 2 +-
lib/interval_tree_test.c | 2 +-
lib/kunit/string-stream-test.c | 1 +
lib/random32.c | 2 +-
lib/rbtree_test.c | 2 +-
lib/test_bpf.c | 2 +-
lib/test_parman.c | 2 +-
lib/test_scanf.c | 2 +-
net/sched/sch_netem.c | 1 +
23 files changed, 22 insertions(+), 24 deletions(-)
--
2.46.0
v24: https://patchwork.kernel.org/project/netdevbpf/list/?series=884556&state=*
====
Changes:
- Fix failing ynl regen error.
- Error path fixes & extack error messages in dmabuf binding.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v24/
v23: https://patchwork.kernel.org/project/netdevbpf/list/?series=882978&state=*
====
Fixing relatively minor issues called out in v22. (thanks again!)
Mostly code cleanups, extack error messages, and minor reworks. Nothing
major really changed, so the exact changes per commit is called in the
commit messages.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v23/
v22: https://patchwork.kernel.org/project/netdevbpf/list/?series=881158&state=*
====
v22 aims to resolve the pending issue pointed to in v21, which is the
interaction with xdp. In this series I rebase on top of the minor
refactor which refactors propagating xdp configuration to slave devices:
https://patchwork.kernel.org/project/netdevbpf/list/?series=881994&state=*
I then disable setting xdp on devices using memory providers, and
propagating xdp configuration to devices using memory providers.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v22/
v21: https://patchwork.kernel.org/project/netdevbpf/list/?series=880735&state=*
====
v20 addressed some comments and resolved a test failure, but introduced
an unfortunate build error with a config edge case I wasn't testing. v21
simply resolves that error.
Major Changes:
- Resolve build error with CONFIG_PAGE_POOL=n && CONFIG_NET=y
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v21/
v20: https://patchwork.kernel.org/project/netdevbpf/list/?series=879373&state=*
====
v20 aims to resolve a couple of bug reports against v19, and addresses
some review comments around the page_pool_check_memory_provider
mechanism.
Major changes:
- Test edge cases such as header split disabled in selftest.
- Change `offset = 0` back to `offset = offset - start` to resolve issue
found in RX path by Taehee (thanks!)
- Address a few comments around page_pool_check_memory_provider() from
Pavel & Jakub.
- Removed some unnecessary includes across various patches in the
series.
- Removed unnecessary EXPORT_SYMBOL(page_pool_mem_providers) (Jakub).
- Fix regression caused by incorrect dev_get_max_mp_channel check, along
with rename (Jakub).
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v20/
v19: https://patchwork.kernel.org/project/netdevbpf/list/?series=876852&state=*
====
v18 got a thorough review (thanks!), and this iteration addresses the
feedback.
Major changes:
- Prevent deactivating mp bound queues.
- Prevent installing xdp on mp bound netdevs, or installing mps on xdp
installed netdevs.
- Fix corner cases in netlink API vis-a-vis missing attributes.
- Iron out the unreadable netmem driver support story. To be honest, the
conversation with Jakub & Pavel got a bit confusing for me. I've
implemented an approach in this set that makes sense to me, and
AFAICT, addresses the requirements. It may be good as-is, or it
may be a conversation starter/continuer. To be honest IMO there
are many ways to skin this cat and I don't see an extremely strong
reason to go for one approach over another. Here is one approach you
may like.
- Don't reset niov dma_addr on allocation & free.
- Add some tests to the selftest that catches some of the issues around
missing netlink attributes or deactivating mp-bound queues.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v19/
v18: https://patchwork.kernel.org/project/netdevbpf/list/?series=874848&state=*
====
v17 got minor feedback: (a) to beef up the description on patch 1 and (b)
to remove the leading underscores in the header definition.
I applied (a). (b) seems to be against current conventions so I did not
apply before further discussion.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v17/
v17: https://patchwork.kernel.org/project/netdevbpf/list/?series=869900&state=*
====
v16 also got a very thorough review and some testing (thanks again!).
Thes version addresses all the concerns reported on v15, in terms of
feedback and issues reported.
Major changes:
- Use ASSERT_RTNL.
- Moved around some of the page_pool helpers definitions so I can hide
some netmem helpers in private files as Jakub suggested.
- Don't make every net_iov hold a ref on the binding as Jakub suggested.
- Fix issue reported by Taehee where we access queues after they have
been freed.
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v17/
v16: https://patchwork.kernel.org/project/netdevbpf/list/?series=866353&state=*
====
v15 got a thorough review and some testing, and this version addresses almost
all the feedback. Some more minor comments where the authors said it
could be done later, I left out.
Major changes:
- Addition of dma-buf introspection to page-pool-get and queue-get.
- Fixes to selftests suggested by Taehee.
- Fixes to documentation suggested by Donald.
- A couple of suggestions and fixes to TCP patches by Eric and David.
- Fixes to number assignements suggested by Arnd.
- Use rtnl_lock()ing to guard against queue reconfiguration while the
page_pool initialization is happening. (Jakub).
- Fixes to a few warnings reproduced by Taehee.
- Fixes to dma-buf binding suggested by Taehee and Jakub.
- Fixes to netlink UAPI suggested by Jakub
- Applied a number of Reviewed-bys and Acked-bys (including ones I lost
from v13+).
Full devmem TCP changes including the full GVE driver implementation is
here:
https://github.com/mina/linux/commits/tcpdevmem-v16/
One caveat: Taehee reproduced a KASAN warning and reported it here:
https://lore.kernel.org/netdev/CAMArcTUdCxOBYGF3vpbq=eBvqZfnc44KBaQTN7H-wqd…
I estimate the issue to be minor and easily fixable:
https://lore.kernel.org/netdev/CAHS8izNgaqC--GGE2xd85QB=utUnOHmioCsDd1TNxJW…
I hope to be able to follow up with a fix to net tree as net-next closes
imminently, but if this iteration doesn't make it in, I will repost with
a fix squashed after net-next reopens, no problem.
v15: https://patchwork.kernel.org/project/netdevbpf/list/?series=865481&state=*
====
No material changes in this version, only a fix to linking against
libynl.a from the last version. Per Jakub's instructions I've pulled one
of his patches into this series, and now use the new libynl.a correctly,
I hope.
As usual, the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v15/
v14: https://patchwork.kernel.org/project/netdevbpf/list/?series=865135&archive=…
====
No material changes in this version. Only rebase and re-verification on
top of net-next. v13, I think, raced with commit ebad6d0334793
("net/ipv4: Use nested-BH locking for ipv4_tcp_sk.") being merged to
net-next that caused a patchwork failure to apply. This series should
apply cleanly on commit c4532232fa2a4 ("selftests: net: remove unneeded
IP_GRE config").
I did not wait the customary 24hr as Jakub said it's OK to repost as soon
as I build test the rebased version:
https://lore.kernel.org/netdev/20240625075926.146d769d@kernel.org/
v13: https://patchwork.kernel.org/project/netdevbpf/list/?series=861406&archive=…
====
Major changes:
--------------
This iteration addresses Pavel's review comments, applies his
reviewed-by's, and seeks to fix the patchwork build error (sorry!).
As usual, the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v13/
v12: https://patchwork.kernel.org/project/netdevbpf/list/?series=859747&state=*
====
Major changes:
--------------
This iteration only addresses one minor comment from Pavel with regards
to the trace printing of netmem, and the patchwork build error
introduced in v11 because I missed doing an allmodconfig build, sorry.
Other than that v11, AFAICT, received no feedback. There is one
discussion about how the specifics of plugging io uring memory through
the page pool, but not relevant to content in this particular patchset,
AFAICT.
As usual, the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v12/
v11: https://patchwork.kernel.org/project/netdevbpf/list/?series=857457&state=*
====
Major Changes:
--------------
v11 addresses feedback received in v10. The major change is the removal
of the memory provider ops as requested by Christoph. We still
accomplish the same thing, but utilizing direct function calls with if
statements rather than generic ops.
Additionally address sparse warnings, bugs and review comments from
folks that reviewed.
As usual, the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v11/
Detailed changelog:
-------------------
- Fixes in netdev_rx_queue_restart() from Pavel & David.
- Remove commit e650e8c3a36f5 ("net: page_pool: create hooks for
custom page providers") from the series to address Christoph's
feedback and rebased other patches on the series on this change.
- Fixed build errors with CONFIG_DMA_SHARED_BUFFER &&
!CONFIG_GENERIC_ALLOCATOR build.
- Fixed sparse warnings pointed out by Paolo.
- Drop unnecessary gro_pull_from_frag0 checks.
- Added Bagas reviewed-by to docs.
v10: https://patchwork.kernel.org/project/netdevbpf/list/?series=852422&state=*
====
Major Changes:
--------------
v9 was sent right before the merge window closed (sorry!). v10 is almost
a re-send of the series now that the merge window re-opened. Only
rebased to latest net-next and addressed some minor iterative comments
received on v9.
As usual, the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v10/
Detailed changelog:
-------------------
- Fixed tokens leaking in DONTNEED setsockopt (Nikolay).
- Moved net_iov_dma_addr() to devmem.c and made it a devmem specific
helpers (David).
- Rename hook alloc_pages to alloc_netmems as alloc_pages is now
preprocessor macro defined and causes a build error.
v9:
===
Major Changes:
--------------
GVE queue API has been merged. Submitting this version as non-RFC after
rebasing on top of the merged API, and dropped the out of tree queue API
I was carrying on github. Addressed the little feedback v8 has received.
Detailed changelog:
------------------
- Added new patch from David Wei to this series for
netdev_rx_queue_restart()
- Fixed sparse error.
- Removed CONFIG_ checks in netmem_is_net_iov()
- Flipped skb->readable to skb->unreadable
- Minor fixes to selftests & docs.
RFC v8:
=======
Major Changes:
--------------
- Fixed build error generated by patch-by-patch build.
- Applied docs suggestions from Randy.
RFC v7:
=======
Major Changes:
--------------
This revision largely rebases on top of net-next and addresses the feedback
RFCv6 received from folks, namely Jakub, Yunsheng, Arnd, David, & Pavel.
The series remains in RFC because the queue-API ndos defined in this
series are not yet implemented. I have a GVE implementation I carry out
of tree for my testing. A upstreamable GVE implementation is in the
works. Aside from that, in my estimation all the patches are ready for
review/merge. Please do take a look.
As usual the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v7/
Detailed changelog:
- Use admin-perm in netlink API.
- Addressed feedback from Jakub with regards to netlink API
implementation.
- Renamed devmem.c functions to something more appropriate for that
file.
- Improve the performance seen through the page_pool benchmark.
- Fix the value definition of all the SO_DEVMEM_* uapi.
- Various fixes to documentation.
Perf - page-pool benchmark:
---------------------------
Improved performance of bench_page_pool_simple.ko tests compared to v6:
https://pastebin.com/raw/v5dYRg8L
net-next base: 8 cycle fast path.
RFC v6: 10 cycle fast path.
RFC v7: 9 cycle fast path.
RFC v7 with CONFIG_DMA_SHARED_BUFFER disabled: 8 cycle fast path,
same as baseline.
Perf - Devmem TCP benchmark:
---------------------
Perf is about the same regardless of the changes in v7, namely the
removal of the static_branch_unlikely to improve the page_pool benchmark
performance:
189/200gbps bi-directional throughput with RX devmem TCP and regular TCP
TX i.e. ~95% line rate.
RFC v6:
=======
Major Changes:
--------------
This revision largely rebases on top of net-next and addresses the little
feedback RFCv5 received.
The series remains in RFC because the queue-API ndos defined in this
series are not yet implemented. I have a GVE implementation I carry out
of tree for my testing. A upstreamable GVE implementation is in the
works. Aside from that, in my estimation all the patches are ready for
review/merge. Please do take a look.
As usual the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v6/
This version also comes with some performance data recorded in the cover
letter (see below changelog).
Detailed changelog:
- Rebased on top of the merged netmem_ref changes.
- Converted skb->dmabuf to skb->readable (Pavel). Pavel's original
suggestion was to remove the skb->dmabuf flag entirely, but when I
looked into it closely, I found the issue that if we remove the flag
we have to dereference the shinfo(skb) pointer to obtain the first
frag to tell whether an skb is readable or not. This can cause a
performance regression if it dirties the cache line when the
shinfo(skb) was not really needed. Instead, I converted the skb->dmabuf
flag into a generic skb->readable flag which can be re-used by io_uring
0-copy RX.
- Squashed a few locking optimizations from Eric Dumazet in the RX path
and the DEVMEM_DONTNEED setsockopt.
- Expanded the tests a bit. Added validation for invalid scenarios and
added some more coverage.
Perf - page-pool benchmark:
---------------------------
bench_page_pool_simple.ko tests with and without these changes:
https://pastebin.com/raw/ncHDwAbn
AFAIK the number that really matters in the perf tests is the
'tasklet_page_pool01_fast_path Per elem'. This one measures at about 8
cycles without the changes but there is some 1 cycle noise in some
results.
With the patches this regresses to 9 cycles with the changes but there
is 1 cycle noise occasionally running this test repeatedly.
Lastly I tried disable the static_branch_unlikely() in
netmem_is_net_iov() check. To my surprise disabling the
static_branch_unlikely() check reduces the fast path back to 8 cycles,
but the 1 cycle noise remains.
Perf - Devmem TCP benchmark:
---------------------
189/200gbps bi-directional throughput with RX devmem TCP and regular TCP
TX i.e. ~95% line rate.
Major changes in RFC v5:
========================
1. Rebased on top of 'Abstract page from net stack' series and used the
new netmem type to refer to LSB set pointers instead of re-using
struct page.
2. Downgraded this series back to RFC and called it RFC v5. This is
because this series is now dependent on 'Abstract page from net
stack'[1] and the queue API. Both are removed from the series to
reduce the patch # and those bits are fairly independent or
pre-requisite work.
3. Reworked the page_pool devmem support to use netmem and for some
more unified handling.
4. Reworked the reference counting of net_iov (renamed from
page_pool_iov) to use pp_ref_count for refcounting.
The full changes including the dependent series and GVE page pool
support is here:
https://github.com/mina/linux/commits/tcpdevmem-rfcv5/
[1] https://patchwork.kernel.org/project/netdevbpf/list/?series=810774
Major changes in v1:
====================
1. Implemented MVP queue API ndos to remove the userspace-visible
driver reset.
2. Fixed issues in the napi_pp_put_page() devmem frag unref path.
3. Removed RFC tag.
Many smaller addressed comments across all the patches (patches have
individual change log).
Full tree including the rest of the GVE driver changes:
https://github.com/mina/linux/commits/tcpdevmem-v1
Changes in RFC v3:
==================
1. Pulled in the memory-provider dependency from Jakub's RFC[1] to make the
series reviewable and mergeable.
2. Implemented multi-rx-queue binding which was a todo in v2.
3. Fix to cmsg handling.
The sticking point in RFC v2[2] was the device reset required to refill
the device rx-queues after the dmabuf bind/unbind. The solution
suggested as I understand is a subset of the per-queue management ops
Jakub suggested or similar:
https://lore.kernel.org/netdev/20230815171638.4c057dcd@kernel.org/
This is not addressed in this revision, because:
1. This point was discussed at netconf & netdev and there is openness to
using the current approach of requiring a device reset.
2. Implementing individual queue resetting seems to be difficult for my
test bed with GVE. My prototype to test this ran into issues with the
rx-queues not coming back up properly if reset individually. At the
moment I'm unsure if it's a mistake in the POC or a genuine issue in
the virtualization stack behind GVE, which currently doesn't test
individual rx-queue restart.
3. Our usecases are not bothered by requiring a device reset to refill
the buffer queues, and we'd like to support NICs that run into this
limitation with resetting individual queues.
My thought is that drivers that have trouble with per-queue configs can
use the support in this series, while drivers that support new netdev
ops to reset individual queues can automatically reset the queue as
part of the dma-buf bind/unbind.
The same approach with device resets is presented again for consideration
with other sticking points addressed.
This proposal includes the rx devmem path only proposed for merge. For a
snapshot of my entire tree which includes the GVE POC page pool support &
device memory support:
https://github.com/torvalds/linux/compare/master...mina:linux:tcpdevmem-v3
[1] https://lore.kernel.org/netdev/f8270765-a27b-6ccf-33ea-cda097168d79@redhat.…
[2] https://lore.kernel.org/netdev/CAHS8izOVJGJH5WF68OsRWFKJid1_huzzUK+hpKbLcL4…
Changes in RFC v2:
==================
The sticking point in RFC v1[1] was the dma-buf pages approach we used to
deliver the device memory to the TCP stack. RFC v2 is a proof-of-concept
that attempts to resolve this by implementing scatterlist support in the
networking stack, such that we can import the dma-buf scatterlist
directly. This is the approach proposed at a high level here[2].
Detailed changes:
1. Replaced dma-buf pages approach with importing scatterlist into the
page pool.
2. Replace the dma-buf pages centric API with a netlink API.
3. Removed the TX path implementation - there is no issue with
implementing the TX path with scatterlist approach, but leaving
out the TX path makes it easier to review.
4. Functionality is tested with this proposal, but I have not conducted
perf testing yet. I'm not sure there are regressions, but I removed
perf claims from the cover letter until they can be re-confirmed.
5. Added Signed-off-by: contributors to the implementation.
6. Fixed some bugs with the RX path since RFC v1.
Any feedback welcome, but specifically the biggest pending questions
needing feedback IMO are:
1. Feedback on the scatterlist-based approach in general.
2. Netlink API (Patch 1 & 2).
3. Approach to handle all the drivers that expect to receive pages from
the page pool (Patch 6).
[1] https://lore.kernel.org/netdev/dfe4bae7-13a0-3c5d-d671-f61b375cb0b4@gmail.c…
[2] https://lore.kernel.org/netdev/CAHS8izPm6XRS54LdCDZVd0C75tA1zHSu6jLVO8nzTLX…
==================
* TL;DR:
Device memory TCP (devmem TCP) is a proposal for transferring data to and/or
from device memory efficiently, without bouncing the data to a host memory
buffer.
* Problem:
A large amount of data transfers have device memory as the source and/or
destination. Accelerators drastically increased the volume of such transfers.
Some examples include:
- ML accelerators transferring large amounts of training data from storage into
GPU/TPU memory. In some cases ML training setup time can be as long as 50% of
TPU compute time, improving data transfer throughput & efficiency can help
improving GPU/TPU utilization.
- Distributed training, where ML accelerators, such as GPUs on different hosts,
exchange data among them.
- Distributed raw block storage applications transfer large amounts of data with
remote SSDs, much of this data does not require host processing.
Today, the majority of the Device-to-Device data transfers the network are
implemented as the following low level operations: Device-to-Host copy,
Host-to-Host network transfer, and Host-to-Device copy.
The implementation is suboptimal, especially for bulk data transfers, and can
put significant strains on system resources, such as host memory bandwidth,
PCIe bandwidth, etc. One important reason behind the current state is the
kernel’s lack of semantics to express device to network transfers.
* Proposal:
In this patch series we attempt to optimize this use case by implementing
socket APIs that enable the user to:
1. send device memory across the network directly, and
2. receive incoming network packets directly into device memory.
Packet _payloads_ go directly from the NIC to device memory for receive and from
device memory to NIC for transmit.
Packet _headers_ go to/from host memory and are processed by the TCP/IP stack
normally. The NIC _must_ support header split to achieve this.
Advantages:
- Alleviate host memory bandwidth pressure, compared to existing
network-transfer + device-copy semantics.
- Alleviate PCIe BW pressure, by limiting data transfer to the lowest level
of the PCIe tree, compared to traditional path which sends data through the
root complex.
* Patch overview:
** Part 1: netlink API
Gives user ability to bind dma-buf to an RX queue.
** Part 2: scatterlist support
Currently the standard for device memory sharing is DMABUF, which doesn't
generate struct pages. On the other hand, networking stack (skbs, drivers, and
page pool) operate on pages. We have 2 options:
1. Generate struct pages for dmabuf device memory, or,
2. Modify the networking stack to process scatterlist.
Approach #1 was attempted in RFC v1. RFC v2 implements approach #2.
** part 3: page pool support
We piggy back on page pool memory providers proposal:
https://github.com/kuba-moo/linux/tree/pp-providers
It allows the page pool to define a memory provider that provides the
page allocation and freeing. It helps abstract most of the device memory
TCP changes from the driver.
** part 4: support for unreadable skb frags
Page pool iovs are not accessible by the host; we implement changes
throughput the networking stack to correctly handle skbs with unreadable
frags.
** Part 5: recvmsg() APIs
We define user APIs for the user to send and receive device memory.
Not included with this series is the GVE devmem TCP support, just to
simplify the review. Code available here if desired:
https://github.com/mina/linux/tree/tcpdevmem
This series is built on top of net-next with Jakub's pp-providers changes
cherry-picked.
* NIC dependencies:
1. (strict) Devmem TCP require the NIC to support header split, i.e. the
capability to split incoming packets into a header + payload and to put
each into a separate buffer. Devmem TCP works by using device memory
for the packet payload, and host memory for the packet headers.
2. (optional) Devmem TCP works better with flow steering support & RSS support,
i.e. the NIC's ability to steer flows into certain rx queues. This allows the
sysadmin to enable devmem TCP on a subset of the rx queues, and steer
devmem TCP traffic onto these queues and non devmem TCP elsewhere.
The NIC I have access to with these properties is the GVE with DQO support
running in Google Cloud, but any NIC that supports these features would suffice.
I may be able to help reviewers bring up devmem TCP on their NICs.
* Testing:
The series includes a udmabuf kselftest that show a simple use case of
devmem TCP and validates the entire data path end to end without
a dependency on a specific dmabuf provider.
** Test Setup
Kernel: net-next with this series and memory provider API cherry-picked
locally.
Hardware: Google Cloud A3 VMs.
NIC: GVE with header split & RSS & flow steering support.
Cc: Pavel Begunkov <asml.silence(a)gmail.com>
Cc: David Wei <dw(a)davidwei.uk>
Cc: Jason Gunthorpe <jgg(a)ziepe.ca>
Cc: Yunsheng Lin <linyunsheng(a)huawei.com>
Cc: Shailend Chand <shailend(a)google.com>
Cc: Harshitha Ramamurthy <hramamurthy(a)google.com>
Cc: Shakeel Butt <shakeel.butt(a)linux.dev>
Cc: Jeroen de Borst <jeroendb(a)google.com>
Cc: Praveen Kaligineedi <pkaligineedi(a)google.com>
Cc: Bagas Sanjaya <bagasdotme(a)gmail.com>
Cc: Steven Rostedt <rostedt(a)goodmis.org>
Cc: Christoph Hellwig <hch(a)infradead.org>
Cc: Nikolay Aleksandrov <razor(a)blackwall.org>
Cc: Taehee Yoo <ap420073(a)gmail.com>
Cc: Donald Hunter <donald.hunter(a)gmail.com>
Mina Almasry (13):
netdev: add netdev_rx_queue_restart()
net: netdev netlink api to bind dma-buf to a net device
netdev: support binding dma-buf to netdevice
netdev: netdevice devmem allocator
page_pool: devmem support
memory-provider: dmabuf devmem memory provider
net: support non paged skb frags
net: add support for skbs with unreadable frags
tcp: RX path for devmem TCP
net: add SO_DEVMEM_DONTNEED setsockopt to release RX frags
net: add devmem TCP documentation
selftests: add ncdevmem, netcat for devmem TCP
netdev: add dmabuf introspection
Documentation/netlink/specs/netdev.yaml | 61 +++
Documentation/networking/devmem.rst | 269 +++++++++++
Documentation/networking/index.rst | 1 +
arch/alpha/include/uapi/asm/socket.h | 6 +
arch/mips/include/uapi/asm/socket.h | 6 +
arch/parisc/include/uapi/asm/socket.h | 6 +
arch/sparc/include/uapi/asm/socket.h | 6 +
include/linux/netdevice.h | 2 +
include/linux/skbuff.h | 61 ++-
include/linux/skbuff_ref.h | 9 +-
include/linux/socket.h | 1 +
include/net/devmem.h | 136 ++++++
include/net/mp_dmabuf_devmem.h | 44 ++
include/net/netdev_rx_queue.h | 5 +
include/net/netmem.h | 163 ++++++-
include/net/page_pool/helpers.h | 39 +-
include/net/page_pool/types.h | 22 +-
include/net/sock.h | 2 +
include/net/tcp.h | 5 +-
include/trace/events/page_pool.h | 12 +-
include/uapi/asm-generic/socket.h | 6 +
include/uapi/linux/netdev.h | 13 +
include/uapi/linux/uio.h | 17 +
net/Kconfig | 5 +
net/core/Makefile | 2 +
net/core/datagram.c | 6 +
net/core/dev.c | 28 +-
net/core/devmem.c | 388 ++++++++++++++++
net/core/gro.c | 3 +-
net/core/netdev-genl-gen.c | 23 +
net/core/netdev-genl-gen.h | 6 +
net/core/netdev-genl.c | 137 +++++-
net/core/netdev_rx_queue.c | 81 ++++
net/core/netmem_priv.h | 31 ++
net/core/page_pool.c | 117 +++--
net/core/page_pool_priv.h | 46 ++
net/core/page_pool_user.c | 31 +-
net/core/skbuff.c | 77 +++-
net/core/sock.c | 68 +++
net/ethtool/common.c | 8 +
net/ipv4/esp4.c | 3 +-
net/ipv4/tcp.c | 261 ++++++++++-
net/ipv4/tcp_input.c | 13 +-
net/ipv4/tcp_ipv4.c | 16 +
net/ipv4/tcp_minisocks.c | 2 +
net/ipv4/tcp_output.c | 5 +-
net/ipv6/esp6.c | 3 +-
net/packet/af_packet.c | 4 +-
net/xdp/xsk_buff_pool.c | 5 +
tools/include/uapi/linux/netdev.h | 13 +
tools/net/ynl/lib/.gitignore | 1 +
tools/testing/selftests/net/.gitignore | 1 +
tools/testing/selftests/net/Makefile | 9 +
tools/testing/selftests/net/ncdevmem.c | 570 ++++++++++++++++++++++++
54 files changed, 2731 insertions(+), 124 deletions(-)
create mode 100644 Documentation/networking/devmem.rst
create mode 100644 include/net/devmem.h
create mode 100644 include/net/mp_dmabuf_devmem.h
create mode 100644 net/core/devmem.c
create mode 100644 net/core/netdev_rx_queue.c
create mode 100644 net/core/netmem_priv.h
create mode 100644 tools/testing/selftests/net/ncdevmem.c
--
2.46.0.469.g59c65b2a67-goog
From: Jason Xing <kernelxing(a)tencent.com>
When one socket is set SOF_TIMESTAMPING_RX_SOFTWARE which means the
whole system turns on the netstamp_needed_key button, other sockets
that only have SOF_TIMESTAMPING_SOFTWARE will be affected and then
print the rx timestamp information even without setting
SOF_TIMESTAMPING_RX_SOFTWARE generation flag.
How to solve it without breaking users?
We introduce a new flag named SOF_TIMESTAMPING_OPT_RX_FILTER. Using
it together with SOF_TIMESTAMPING_SOFTWARE can stop reporting the
rx software timestamp.
Similarly, we also filter out the hardware case where one process
enables the rx hardware generation flag, then another process only
passing SOF_TIMESTAMPING_RAW_HARDWARE gets the timestamp. So we can set
both SOF_TIMESTAMPING_RAW_HARDWARE and SOF_TIMESTAMPING_OPT_RX_FILTER
to stop reporting rx hardware timestamp after this patch applied.
v5
Link: https://lore.kernel.org/all/20240905071738.3725-1-kerneljasonxing@gmail.com/
1. squash the hardware case patch into this one (Willem)
2. update corresponding commit message and doc (Willem)
3. remove the limitation in sock_set_timestamping() and restore the
simplification branches. (Willem)
4. add missing type and another test in selftests
v4
Link: https://lore.kernel.org/all/20240830153751.86895-1-kerneljasonxing@gmail.co…
1. revise the doc and commit message (Willem)
2. add patch [2/4] to make the doc right (Willem)
3. add patch [3/4] to cover the hardware use (Willem)
4. add testcase for hardware use.
Note: the reason why I split into 4 patches is try to make each commit
clean, atomic, easy to review.
v3
Link: https://lore.kernel.org/all/20240828160145.68805-1-kerneljasonxing@gmail.co…
1. introduce a new flag to avoid application breakage, suggested by
Willem.
2. add it into the selftests.
v2
Link: https://lore.kernel.org/all/20240825152440.93054-1-kerneljasonxing@gmail.co…
Discussed with Willem
1. update the documentation accordingly
2. add more comments in each patch
3. remove the previous test statements in __sock_recv_timestamp()
Jason Xing (2):
net-timestamp: introduce SOF_TIMESTAMPING_OPT_RX_FILTER flag
net-timestamp: add selftests for SOF_TIMESTAMPING_OPT_RX_FILTER
Documentation/networking/timestamping.rst | 27 +++++++++++++++++++++++
include/uapi/linux/net_tstamp.h | 3 ++-
net/ethtool/common.c | 1 +
net/ipv4/tcp.c | 9 ++++++--
net/socket.c | 10 +++++++--
tools/testing/selftests/net/rxtimestamp.c | 18 +++++++++++++++
6 files changed, 63 insertions(+), 5 deletions(-)
--
2.37.3
Hello,
kernel test robot noticed "kernel-selftests.vDSO.vdso_standalone_test_x86.fail" on:
commit: f68b079b1d5ec46687a097347303b616927eb9ff ("selftests: vDSO: build tests with O2 optimization")
https://git.kernel.org/cgit/linux/kernel/git/crng/random.git jd/arm64-vdso
in testcase: kernel-selftests
version: kernel-selftests-x86_64-977d51cf-1_20240508
with following parameters:
group: group-03
compiler: gcc-12
test machine: 36 threads 1 sockets Intel(R) Core(TM) i9-10980XE CPU @ 3.00GHz (Cascade Lake) with 32G memory
(please refer to attached dmesg/kmsg for entire log/backtrace)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <oliver.sang(a)intel.com>
| Closes: https://lore.kernel.org/oe-lkp/202409082121.553d4c89-oliver.sang@intel.com
# timeout set to 300
# selftests: vDSO: vdso_standalone_test_x86
# Segmentation fault
not ok 5 selftests: vDSO: vdso_standalone_test_x86 # exit=139
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20240908/202409082121.553d4c89-oliv…
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
Create a test for the robust list mechanism.
Signed-off-by: André Almeida <andrealmeid(a)igalia.com>
---
Changes from v1:
- Change futex type from int to _Atomic(unsigned int)
- Use old futex(FUTEX_WAIT) instead of the new sys_futex_wait()
---
.../selftests/futex/functional/.gitignore | 1 +
.../selftests/futex/functional/Makefile | 3 +-
.../selftests/futex/functional/robust_list.c | 448 ++++++++++++++++++
3 files changed, 451 insertions(+), 1 deletion(-)
create mode 100644 tools/testing/selftests/futex/functional/robust_list.c
diff --git a/tools/testing/selftests/futex/functional/.gitignore b/tools/testing/selftests/futex/functional/.gitignore
index fbcbdb6963b3..4726e1be7497 100644
--- a/tools/testing/selftests/futex/functional/.gitignore
+++ b/tools/testing/selftests/futex/functional/.gitignore
@@ -9,3 +9,4 @@ futex_wait_wouldblock
futex_wait
futex_requeue
futex_waitv
+robust_list
diff --git a/tools/testing/selftests/futex/functional/Makefile b/tools/testing/selftests/futex/functional/Makefile
index f79f9bac7918..b8635a1ac7f6 100644
--- a/tools/testing/selftests/futex/functional/Makefile
+++ b/tools/testing/selftests/futex/functional/Makefile
@@ -17,7 +17,8 @@ TEST_GEN_PROGS := \
futex_wait_private_mapped_file \
futex_wait \
futex_requeue \
- futex_waitv
+ futex_waitv \
+ robust_list
TEST_PROGS := run.sh
diff --git a/tools/testing/selftests/futex/functional/robust_list.c b/tools/testing/selftests/futex/functional/robust_list.c
new file mode 100644
index 000000000000..9308eb189d48
--- /dev/null
+++ b/tools/testing/selftests/futex/functional/robust_list.c
@@ -0,0 +1,448 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C) 2024 Igalia S.L.
+ *
+ * Robust list test by André Almeida <andrealmeid(a)igalia.com>
+ *
+ * The robust list uAPI allows userspace to create "robust" locks, in the sense
+ * that if the lock holder thread dies, the remaining threads that are waiting
+ * for the lock won't block forever, waiting for a lock that will never be
+ * released.
+ *
+ * This is achieve by userspace setting a list where a thread can enter all the
+ * locks (futexes) that it is holding. The robust list is a linked list, and
+ * userspace register the start of the list with the syscall set_robust_list().
+ * If such thread eventually dies, the kernel will walk this list, waking up one
+ * thread waiting for each futex and marking the futex word with the flag
+ * FUTEX_OWNER_DIED.
+ *
+ * See also
+ * man set_robust_list
+ * Documententation/locking/robust-futex-ABI.rst
+ * Documententation/locking/robust-futexes.rst
+ */
+
+#define _GNU_SOURCE
+
+#include "../../kselftest_harness.h"
+
+#include "futextest.h"
+
+#include <pthread.h>
+#include <stdatomic.h>
+#include <stddef.h>
+
+#define STACK_SIZE (1024 * 1024)
+
+#define FUTEX_TIMEOUT 3
+
+static pthread_barrier_t barrier, barrier2;
+
+int set_robust_list(struct robust_list_head *head, size_t len)
+{
+ return syscall(SYS_set_robust_list, head, len);
+}
+
+int get_robust_list(int pid, struct robust_list_head **head, size_t *len_ptr)
+{
+ return syscall(SYS_get_robust_list, pid, head, len_ptr);
+}
+
+/*
+ * Basic lock struct, contains just the futex word and the robust list element
+ * Real implementations have also a *prev to easily walk in the list
+ */
+struct lock_struct {
+ _Atomic(unsigned int) futex;
+ struct robust_list list;
+};
+
+/*
+ * Helper function to spawn a child thread. Returns -1 on error, pid on success
+ */
+static int create_child(int (*fn)(void *arg), void *arg)
+{
+ char *stack;
+ pid_t pid;
+
+ stack = mmap(NULL, STACK_SIZE, PROT_READ | PROT_WRITE,
+ MAP_PRIVATE | MAP_ANONYMOUS | MAP_STACK, -1, 0);
+ if (stack == MAP_FAILED)
+ return -1;
+
+ stack += STACK_SIZE;
+
+ pid = clone(fn, stack, CLONE_VM | SIGCHLD, arg);
+
+ if (pid == -1)
+ return -1;
+
+ return pid;
+}
+
+/*
+ * Helper function to prepare and register a robust list
+ */
+static int set_list(struct robust_list_head *head)
+{
+ int ret;
+
+ ret = set_robust_list(head, sizeof(struct robust_list_head));
+ if (ret)
+ return ret;
+
+ head->futex_offset = (size_t) offsetof(struct lock_struct, futex) -
+ (size_t) offsetof(struct lock_struct, list);
+ head->list.next = &head->list;
+ head->list_op_pending = NULL;
+
+ return 0;
+}
+
+/*
+ * A basic (and incomplete) mutex lock function with robustness
+ */
+static int mutex_lock(struct lock_struct *lock, struct robust_list_head *head, bool error_inject)
+{
+ _Atomic(unsigned int) *futex = &lock->futex;
+ int zero = 0, ret = -1;
+ pid_t tid = gettid();
+
+ /*
+ * Set list_op_pending before starting the lock, so the kernel can catch
+ * the case where the thread died during the lock operation
+ */
+ head->list_op_pending = &lock->list;
+
+ if (atomic_compare_exchange_strong(futex, &zero, tid)) {
+ /*
+ * We took the lock, insert it in the robust list
+ */
+ struct robust_list *list = &head->list;
+
+ /* Error injection to test list_op_pending */
+ if (error_inject)
+ return 0;
+
+ while (list->next != &head->list)
+ list = list->next;
+
+ list->next = &lock->list;
+ lock->list.next = &head->list;
+
+ ret = 0;
+ } else {
+ /*
+ * We didn't take the lock, wait until the owner wakes (or dies)
+ */
+ struct timespec to;
+
+ clock_gettime(CLOCK_MONOTONIC, &to);
+ to.tv_sec = to.tv_sec + FUTEX_TIMEOUT;
+
+ tid = atomic_load(futex);
+ /* Kernel ignores futexes without the waiters flag */
+ tid |= FUTEX_WAITERS;
+ atomic_store(futex, tid);
+
+ ret = futex_wait((futex_t *) futex, tid, &to, 0);
+
+ /*
+ * A real mutex_lock() implementation would loop here to finally
+ * take the lock. We don't care about that, so we stop here.
+ */
+ }
+
+ head->list_op_pending = NULL;
+
+ return ret;
+}
+
+/*
+ * This child thread will succeed taking the lock, and then will exit holding it
+ */
+static int child_fn_lock(void *arg)
+{
+ struct lock_struct *lock = (struct lock_struct *) arg;
+ struct robust_list_head head;
+ int ret;
+
+ ret = set_list(&head);
+ if (ret)
+ ksft_test_result_fail("set_robust_list error\n");
+
+ ret = mutex_lock(lock, &head, false);
+ if (ret)
+ ksft_test_result_fail("mutex_lock error\n");
+
+ pthread_barrier_wait(&barrier);
+
+ /*
+ * There's a race here: the parent thread needs to be inside
+ * futex_wait() before the child thread dies, otherwise it will miss the
+ * wakeup from handle_futex_death() that this child will emit. We wait a
+ * little bit just to make sure that this happens.
+ */
+ sleep(1);
+
+ return 0;
+}
+
+/*
+ * Spawns a child thread that will set a robust list, take the lock, register it
+ * in the robust list and die. The parent thread will wait on this futex, and
+ * should be waken up when the child exits.
+ */
+TEST(robustness)
+{
+ struct lock_struct lock = { .futex = 0 };
+ struct robust_list_head head;
+ _Atomic(unsigned int) *futex = &lock.futex;
+ int ret;
+
+ ret = set_list(&head);
+ ASSERT_EQ(ret, 0);
+
+ /*
+ * Lets use a barrier to ensure that the child thread takes the lock
+ * before the parent
+ */
+ ret = pthread_barrier_init(&barrier, NULL, 2);
+ ASSERT_EQ(ret, 0);
+
+ ret = create_child(&child_fn_lock, &lock);
+ ASSERT_NE(ret, -1);
+
+ pthread_barrier_wait(&barrier);
+ ret = mutex_lock(&lock, &head, false);
+
+ /*
+ * futex_wait() should return 0 and the futex word should be marked with
+ * FUTEX_OWNER_DIED
+ */
+ ASSERT_EQ(ret, 0) TH_LOG("futex wait returned %d", errno);
+ ASSERT_TRUE(*futex | FUTEX_OWNER_DIED);
+
+ pthread_barrier_destroy(&barrier);
+}
+
+/*
+ * The only valid value for len is sizeof(*head)
+ */
+TEST(set_robust_list_invalid_size)
+{
+ struct robust_list_head head;
+ size_t head_size = sizeof(struct robust_list_head);
+ int ret;
+
+ ret = set_robust_list(&head, head_size);
+ ASSERT_EQ(ret, 0);
+
+ ret = set_robust_list(&head, head_size * 2);
+ ASSERT_EQ(ret, -1);
+ ASSERT_EQ(errno, EINVAL);
+
+ ret = set_robust_list(&head, head_size - 1);
+ ASSERT_EQ(ret, -1);
+ ASSERT_EQ(errno, EINVAL);
+
+ ret = set_robust_list(&head, 0);
+ ASSERT_EQ(ret, -1);
+ ASSERT_EQ(errno, EINVAL);
+}
+
+/*
+ * Test get_robust_list with pid = 0, getting the list of the running thread
+ */
+TEST(get_robust_list_self)
+{
+ struct robust_list_head head, head2, *get_head;
+ size_t head_size = sizeof(struct robust_list_head), len_ptr;
+ int ret;
+
+ ret = set_robust_list(&head, head_size);
+ ASSERT_EQ(ret, 0);
+
+ ret = get_robust_list(0, &get_head, &len_ptr);
+ ASSERT_EQ(ret, 0);
+ ASSERT_EQ(get_head, &head);
+ ASSERT_EQ(head_size, len_ptr);
+
+ ret = set_robust_list(&head2, head_size);
+ ASSERT_EQ(ret, 0);
+
+ ret = get_robust_list(0, &get_head, &len_ptr);
+ ASSERT_EQ(ret, 0);
+ ASSERT_EQ(get_head, &head2);
+ ASSERT_EQ(head_size, len_ptr);
+}
+
+static int child_list(void *arg)
+{
+ struct robust_list_head *head = (struct robust_list_head *) arg;
+ int ret;
+
+ ret = set_robust_list(head, sizeof(struct robust_list_head));
+ if (ret)
+ ksft_test_result_fail("set_robust_list error\n");
+
+ pthread_barrier_wait(&barrier);
+ pthread_barrier_wait(&barrier2);
+
+ return 0;
+}
+
+/*
+ * Test get_robust_list from another thread. We use two barriers here to ensure
+ * that:
+ * 1) the child thread set the list before we try to get it from the
+ * parent
+ * 2) the child thread still alive when we try to get the list from it
+ */
+TEST(get_robust_list_child)
+{
+ pid_t tid;
+ int ret;
+ struct robust_list_head head, *get_head;
+ size_t len_ptr;
+
+ ret = pthread_barrier_init(&barrier, NULL, 2);
+ ret = pthread_barrier_init(&barrier2, NULL, 2);
+ ASSERT_EQ(ret, 0);
+
+ tid = create_child(&child_list, &head);
+ ASSERT_NE(tid, -1);
+
+ pthread_barrier_wait(&barrier);
+
+ ret = get_robust_list(tid, &get_head, &len_ptr);
+ ASSERT_EQ(ret, 0);
+ ASSERT_EQ(&head, get_head);
+
+ pthread_barrier_wait(&barrier2);
+
+ pthread_barrier_destroy(&barrier);
+ pthread_barrier_destroy(&barrier2);
+}
+
+static int child_fn_lock_with_error(void *arg)
+{
+ struct lock_struct *lock = (struct lock_struct *) arg;
+ struct robust_list_head head;
+ int ret;
+
+ ret = set_list(&head);
+ if (ret)
+ ksft_test_result_fail("set_robust_list error\n");
+
+ ret = mutex_lock(lock, &head, true);
+ if (ret)
+ ksft_test_result_fail("mutex_lock error\n");
+
+ pthread_barrier_wait(&barrier);
+
+ sleep(1);
+
+ return 0;
+}
+
+/*
+ * Same as robustness test, but inject an error where the mutex_lock() exits
+ * earlier, just after setting list_op_pending and taking the lock, to test the
+ * list_op_pending mechanism
+ */
+TEST(set_list_op_pending)
+{
+ struct lock_struct lock = { .futex = 0 };
+ struct robust_list_head head;
+ _Atomic(unsigned int) *futex = &lock.futex;
+ int ret;
+
+ ret = set_list(&head);
+ ASSERT_EQ(ret, 0);
+
+ ret = pthread_barrier_init(&barrier, NULL, 2);
+ ASSERT_EQ(ret, 0);
+
+ ret = create_child(&child_fn_lock_with_error, &lock);
+ ASSERT_NE(ret, -1);
+
+ pthread_barrier_wait(&barrier);
+ ret = mutex_lock(&lock, &head, false);
+
+ ASSERT_EQ(ret, 0) TH_LOG("futex wait returned %d", errno);
+ ASSERT_TRUE(*futex | FUTEX_OWNER_DIED);
+
+ pthread_barrier_destroy(&barrier);
+}
+
+#define CHILD_NR 10
+
+static int child_lock_holder(void *arg)
+{
+ struct lock_struct *locks = (struct lock_struct *) arg;
+ struct robust_list_head head;
+ int i;
+
+ set_list(&head);
+
+ for (i = 0; i < CHILD_NR; i++) {
+ locks[i].futex = 0;
+ mutex_lock(&locks[i], &head, false);
+ }
+
+ pthread_barrier_wait(&barrier);
+ pthread_barrier_wait(&barrier2);
+
+ sleep(1);
+ return 0;
+}
+
+static int child_wait_lock(void *arg)
+{
+ struct lock_struct *lock = (struct lock_struct *) arg;
+ struct robust_list_head head;
+ int ret;
+
+ pthread_barrier_wait(&barrier2);
+ ret = mutex_lock(lock, &head, false);
+
+ if (ret)
+ ksft_test_result_fail("mutex_lock error\n");
+
+ if (!(lock->futex | FUTEX_OWNER_DIED))
+ ksft_test_result_fail("futex not marked with FUTEX_OWNER_DIED\n");
+
+ return 0;
+}
+
+/*
+ * Test a robust list of more than one element. All the waiters should wake when
+ * the holder dies
+ */
+TEST(robust_list_multiple_elements)
+{
+ struct lock_struct locks[CHILD_NR];
+ int i, ret;
+
+ ret = pthread_barrier_init(&barrier, NULL, 2);
+ ASSERT_EQ(ret, 0);
+ ret = pthread_barrier_init(&barrier2, NULL, CHILD_NR + 1);
+ ASSERT_EQ(ret, 0);
+
+ create_child(&child_lock_holder, &locks);
+
+ /* Wait until the locker thread takes the look */
+ pthread_barrier_wait(&barrier);
+
+ for (i = 0; i < CHILD_NR; i++)
+ create_child(&child_wait_lock, &locks[i]);
+
+ /* Wait for all children to return */
+ while (wait(NULL) > 0);
+
+ pthread_barrier_destroy(&barrier);
+ pthread_barrier_destroy(&barrier2);
+}
+
+TEST_HARNESS_MAIN
--
2.46.0
Thank you for the review, I have added the changelog as requested.
Changelog:
- Added missing newline to the `ksft_print_msg` in `test_zswap_writeback` function.
Signed-off-by: Mohammed Anees <pvmohammedanees2003(a)gmail.com>
---
tools/testing/selftests/cgroup/test_zswap.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/testing/selftests/cgroup/test_zswap.c b/tools/testing/selftests/cgroup/test_zswap.c
index 190096017..7c849d836 100644
--- a/tools/testing/selftests/cgroup/test_zswap.c
+++ b/tools/testing/selftests/cgroup/test_zswap.c
@@ -351,7 +351,7 @@ static int test_zswap_writeback(const char *root, bool wb)
goto out;
if (wb != !!zswpwb_after) {
- ksft_print_msg("zswpwb_after is %ld while wb is %s",
+ ksft_print_msg("zswpwb_after is %ld while wb is %s\n",
zswpwb_after, wb ? "enabled" : "disabled");
goto out;
}
--
2.43.0
Hi Linus,
Please pull the following KUnit fixes second update for Linux 6.11-rc7.
This KUnit fixes update for Linux 6.11-rc7 consists of a fix to
missing function parameter warning found during documentation
build in linux-next.
diff is attached.
thanks,
-- Shuah
----------------------------------------------------------------
The following changes since commit f2c6dbd220170c2396fb019ead67fbada1e23ebd:
kunit: Device wrappers should also manage driver name (2024-08-26 07:03:46 -0600)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest tags/linux_kselftest-kunit-fixes-6.11-rc7-2
for you to fetch changes up to 12cb32a52eb607dc4d0e45fe6f4cf946d08da0fd:
kunit: Fix missing kerneldoc comment (2024-09-05 14:29:10 -0600)
----------------------------------------------------------------
linux_kselftest-kunit-fixes-6.11-rc7-2
This KUnit fixes update for Linux 6.11-rc7 consists of a fix to
missing function parameter warning found during documentation
build in linux-next.
----------------------------------------------------------------
David Gow (1):
kunit: Fix missing kerneldoc comment
include/kunit/test.h | 1 +
1 file changed, 1 insertion(+)
----------------------------------------------------------------
rxq contains a pointer to the device from where
the redirect happened. Currently, the BPF program
that was executed after a redirect via BPF_MAP_TYPE_DEVMAP*
does not have it set.
Add bugfix and related selftest.
Signed-off-by: Florian Kauer <florian.kauer(a)linutronix.de>
---
Changes in v2:
- changed fixes tag
- added selftest
- Link to v1: https://lore.kernel.org/r/20240905-devel-koalo-fix-ingress-ifindex-v1-1-d12…
---
Florian Kauer (2):
bpf: devmap: provide rxq after redirect
bpf: selftests: send packet to devmap redirect XDP
kernel/bpf/devmap.c | 11 +-
.../selftests/bpf/prog_tests/xdp_devmap_attach.c | 114 +++++++++++++++++++--
2 files changed, 115 insertions(+), 10 deletions(-)
---
base-commit: 8e69c96df771ab469cec278edb47009351de4da6
change-id: 20240905-devel-koalo-fix-ingress-ifindex-b9293d471db6
Best regards,
--
Florian Kauer <florian.kauer(a)linutronix.de>
While kunit/visibility.h is today not included in any generated
kernel documentation, also likely due to the fact that none of the
existing comments are correctly recognized as kernel-doc, but once
we decide to add this header and fix the tool, there will be:
../include/kunit/visibility.h:61: warning: Function parameter or
struct member 'symbol' not described in 'EXPORT_SYMBOL_IF_KUNIT'
Signed-off-by: Michal Wajdeczko <michal.wajdeczko(a)intel.com>
---
Cc: Rae Moar <rmoar(a)google.com>
Cc: David Gow <davidgow(a)google.com>
---
include/kunit/visibility.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/include/kunit/visibility.h b/include/kunit/visibility.h
index 0dfe35feeec6..efff77b58dd6 100644
--- a/include/kunit/visibility.h
+++ b/include/kunit/visibility.h
@@ -22,6 +22,7 @@
* EXPORTED_FOR_KUNIT_TESTING namespace only if CONFIG_KUNIT is
* enabled. Must use MODULE_IMPORT_NS(EXPORTED_FOR_KUNIT_TESTING)
* in test file in order to use symbols.
+ * @symbol: the symbol identifier to export
*/
#define EXPORT_SYMBOL_IF_KUNIT(symbol) EXPORT_SYMBOL_NS(symbol, \
EXPORTED_FOR_KUNIT_TESTING)
--
2.43.0
v1: https://groups.google.com/g/kunit-dev/c/f4LIMLyofj8
v2: make it more complex and attempt to be thread safe
s/FIXED_STUB/GLOBAL_STUB (David, Lucas)
make it little more thread safe (Rae, David)
wait until stub call finishes before test end (David)
wait until stub call finishes before changing stub (David)
allow stub deactivation (Rae)
prefer kunit log (David)
add simple selftest (Michal)
also introduce ONLY_IF_KUNIT macro (Michal)
v3: include example for DECLARE_IF_KUNIT (Lucas)
rename s/ONLY_IF_KUNIT/VALUE_IF_KUNIT (Michal)
and add simple usage example for it (Rae)
fix s/fixed/global in comments (Lucas)
improve stub sanitize flow (Lucas, Michal)
reformat kernel-doc for better output (Michal)
Test outputs:
$ tools/testing/kunit/kunit.py run *example*.*global* \
--kunitconfig lib/kunit/.kunitconfig --raw_output
KTAP version 1
1..1
# example: initializing suite
KTAP version 1
# Subtest: example
# module: kunit_example_test
1..1
# example_global_stub_test: initializing
# example_global_stub_test: add_two: redirecting to subtract_one
# example_global_stub_test: add_two: redirecting to subtract_one
# example_global_stub_test: cleaning up
ok 1 example_global_stub_test
# example: exiting suite
ok 1 example
$ tools/testing/kunit/kunit.py run *global*.*global* \
--kunitconfig lib/kunit/.kunitconfig --raw_output
KTAP version 1
1..1
KTAP version 1
# Subtest: kunit_global_stub
# module: kunit_test
1..4
# kunit_global_stub_test_activate: real_void_func: redirecting to replacement_void_func
# kunit_global_stub_test_activate: real_func: redirecting to replacement_func
# kunit_global_stub_test_activate: real_func: redirecting to replacement_func
# kunit_global_stub_test_activate: real_func: redirecting to other_replacement_func
# kunit_global_stub_test_activate: real_func: redirecting to other_replacement_func
# kunit_global_stub_test_activate: real_func: redirecting to super_replacement_func
# kunit_global_stub_test_activate: real_func: redirecting to super_replacement_func
ok 1 kunit_global_stub_test_activate
ok 2 kunit_global_stub_test_deactivate
# kunit_global_stub_test_slow_deactivate: real_func: redirecting to slow_replacement_func
# kunit_global_stub_test_slow_deactivate: real_func: redirecting to slow_replacement_func
# kunit_global_stub_test_slow_deactivate: waiting for slow_replacement_func
# kunit_global_stub_test_slow_deactivate.speed: slow
ok 3 kunit_global_stub_test_slow_deactivate
# kunit_global_stub_test_slow_replace: real_func: redirecting to slow_replacement_func
# kunit_global_stub_test_slow_replace: real_func: redirecting to slow_replacement_func
# kunit_global_stub_test_slow_replace: waiting for slow_replacement_func
# kunit_global_stub_test_slow_replace: real_func: redirecting to other_replacement_func
# kunit_global_stub_test_slow_replace.speed: slow
ok 4 kunit_global_stub_test_slow_replace
# kunit_global_stub: pass:4 fail:0 skip:0 total:4
# Totals: pass:4 fail:0 skip:0 total:4
ok 1 kunit_global_stub
Cc: Rae Moar <rmoar(a)google.com>
Cc: David Gow <davidgow(a)google.com>
Cc: Lucas De Marchi <lucas.demarchi(a)intel.com>
Michal Wajdeczko (6):
kunit: Introduce kunit_is_running()
kunit: Add macro to conditionally expose declarations to tests
kunit: Add macro to conditionally expose expressions to tests
kunit: Allow function redirection outside of the KUnit thread
kunit: Add example with alternate function redirection method
kunit: Add some selftests for global stub redirection macros
include/kunit/static_stub.h | 158 ++++++++++++++++++++
include/kunit/test-bug.h | 12 +-
include/kunit/visibility.h | 40 ++++++
lib/kunit/kunit-example-test.c | 67 +++++++++
lib/kunit/kunit-test.c | 254 ++++++++++++++++++++++++++++++++-
lib/kunit/static_stub.c | 50 +++++++
6 files changed, 578 insertions(+), 3 deletions(-)
--
2.43.0
From: Yuan Chen <chenyuan(a)kylinos.cn>
This patch identifies whether a test item is valid by adding a valid flag to res.
When we test the bpf_cookies/perf_event sub-test item of test_progs, there is a
probability failure of the test item. In fact, this is not a problem, because
the corresponding perf event is not collected. This should not output the test
failure, and it is more reasonable to output SKIP. Therefore, add a valid
identifier to res to distinguish whether the test item is valid, and skip the
test item if it is invalid.
Signed-off-by: Yuan Chen <chenyuan(a)kylinos.cn>
---
.../testing/selftests/bpf/prog_tests/bpf_cookie.c | 15 +++++++++++++++
.../testing/selftests/bpf/progs/test_bpf_cookie.c | 2 ++
2 files changed, 17 insertions(+)
diff --git a/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c b/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c
index 070c52c312e5..e5bf4b385501 100644
--- a/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c
+++ b/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c
@@ -456,6 +456,7 @@ static void pe_subtest(struct test_bpf_cookie *skel)
if (!ASSERT_GE(pfd, 0, "perf_fd"))
goto cleanup;
+ skel->bss->res_valid = false;
opts.bpf_cookie = 0x100000;
link = bpf_program__attach_perf_event_opts(skel->progs.handle_pe, pfd, &opts);
if (!ASSERT_OK_PTR(link, "link1"))
@@ -463,6 +464,12 @@ static void pe_subtest(struct test_bpf_cookie *skel)
burn_cpu(); /* trigger BPF prog */
+ if (!skel->bss->res_valid) {
+ printf("%s:SKIP:the corresponding perf event was not sampled.\n",
+ __func__);
+ test__skip();
+ goto cleanup;
+ }
ASSERT_EQ(skel->bss->pe_res, 0x100000, "pe_res1");
/* prevent bpf_link__destroy() closing pfd itself */
@@ -474,6 +481,7 @@ static void pe_subtest(struct test_bpf_cookie *skel)
link = NULL;
kern_sync_rcu();
skel->bss->pe_res = 0;
+ skel->bss->res_valid = false;
opts.bpf_cookie = 0x200000;
link = bpf_program__attach_perf_event_opts(skel->progs.handle_pe, pfd, &opts);
@@ -482,6 +490,13 @@ static void pe_subtest(struct test_bpf_cookie *skel)
burn_cpu(); /* trigger BPF prog */
+ if (!skel->bss->res_valid) {
+ printf("%s:SKIP:the corresponding perf event was not sampled.\n",
+ __func__);
+ test__skip();
+ goto cleanup;
+ }
+
ASSERT_EQ(skel->bss->pe_res, 0x200000, "pe_res2");
cleanup:
diff --git a/tools/testing/selftests/bpf/progs/test_bpf_cookie.c b/tools/testing/selftests/bpf/progs/test_bpf_cookie.c
index c83142b55f47..28d0ae6810d9 100644
--- a/tools/testing/selftests/bpf/progs/test_bpf_cookie.c
+++ b/tools/testing/selftests/bpf/progs/test_bpf_cookie.c
@@ -7,6 +7,7 @@
#include <errno.h>
int my_tid;
+bool res_valid;
__u64 kprobe_res;
__u64 kprobe_multi_res;
@@ -27,6 +28,7 @@ static void update(void *ctx, __u64 *res)
if (my_tid != (u32)bpf_get_current_pid_tgid())
return;
+ res_valid = true;
*res |= bpf_get_attach_cookie(ctx);
}
--
2.46.0