bind_bhash.c passes (SO_REUSEADDR | SO_REUSEPORT) to setsockopt().
In the asm-generic definition, the value happens to match with the
bare SO_REUSEPORT, (2 | 15) == 15, but not on some arch.
arch/alpha/include/uapi/asm/socket.h:18:#define SO_REUSEADDR 0x0004
arch/alpha/include/uapi/asm/socket.h:24:#define SO_REUSEPORT 0x0200
arch/mips/include/uapi/asm/socket.h:24:#define SO_REUSEADDR 0x0004 /* Allow reuse of local addresses. */
arch/mips/include/uapi/asm/socket.h:33:#define SO_REUSEPORT 0x0200 /* Allow local address and port reuse. */
arch/parisc/include/uapi/asm/socket.h:12:#define SO_REUSEADDR 0x0004
arch/parisc/include/uapi/asm/socket.h:18:#define SO_REUSEPORT 0x0200
arch/sparc/include/uapi/asm/socket.h:13:#define SO_REUSEADDR 0x0004
arch/sparc/include/uapi/asm/socket.h:20:#define SO_REUSEPORT 0x0200
include/uapi/asm-generic/socket.h:12:#define SO_REUSEADDR 2
include/uapi/asm-generic/socket.h:27:#define SO_REUSEPORT 15
Let's pass SO_REUSEPORT only.
Fixes: c35ecb95c4 ("selftests/net: Add test for timing a bind request to a port with a populated bhash entry")
Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20250903222938.2601522-1-kuniyu@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Yi Chen reports that 'udpclash' loops forever depending on compiler
(and optimization level used); while (x == 1) gets optimized into
for (;;). Add volatile qualifier to avoid that.
While at it, also run it under timeout(1) and fix the resize script
to not ignore the timeout passed as second parameter to insert_flood.
Reported-by: Yi Chen <yiche@redhat.com>
Suggested-by: Yi Chen <yiche@redhat.com>
Fixes: 78a5883635 ("selftests: netfilter: add conntrack clash resolution test case")
Signed-off-by: Florian Westphal <fw@strlen.de>
Commit 4b30209255 ("selftests/xsk: Add tail adjustment tests and support
check") added a new global to xsk_xdp_progs.c, but left out the access in
the testapp_xdp_metadata_copy() function. Since bpf_map_update_elem() will
write to the whole bss section, it gets truncated. Fix by writing to
skel_rx->bss->count directly.
Fixes: 4b30209255 ("selftests/xsk: Add tail adjustment tests and support check")
Signed-off-by: Ricardo B. Marlière <rbm@suse.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20250829-selftests-bpf-xsk_regression_fix-v1-1-5f5acdb9fe6b@suse.com
Add test cases for VXLAN with FDB nexthop groups, testing both IPv4 and
IPv6. Test basic Tx functionality as well as some corner cases.
Example output:
# ./test_vxlan_nh.sh
TEST: VXLAN FDB nexthop: IPv4 basic Tx [ OK ]
TEST: VXLAN FDB nexthop: IPv6 basic Tx [ OK ]
TEST: VXLAN FDB nexthop: learning [ OK ]
TEST: VXLAN FDB nexthop: IPv4 proxy [ OK ]
TEST: VXLAN FDB nexthop: IPv6 proxy [ OK ]
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>
Link: https://patch.msgid.link/20250901065035.159644-4-idosch@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The rss_ctx test has gotten pretty flaky after I increased
the queue count in NIPA 2->3. Not 100% clear why. We get
a lot of failures in the rss_ctx.test_hitless_key_update case.
Looking closer it appears that the failures are mostly due
to startup costs. I measured the following timing for ethtool -X:
- python cmd(shell=True) : 150-250msec
- python cmd(shell=False) : 50- 70msec
- timed in bash : 45- 55msec
- YNL Netlink call : 2- 4msec
- .set_rxfh callback : 1- 2msec
The target in the test was set to 200msec. We were mostly measuring
ethtool startup cost it seems. Switch to YNL since it's 100x faster.
Lower the pass criteria to 150msec, no real science behind this number
but we removed some overhead, drivers which previously passed 200msec
should easily pass 150msec now.
Separately we should probably follow up on defaulting to shell=False,
when script doesn't explicitly ask for True, because the overhead
is rather significant.
Switch from _rss_key_rand() to random.randbytes(), YNL takes a binary
array rather than array of ints.
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20250901173139.881070-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Overhead of using shell=True is quite significant.
Micro-benchmark of running ethtool --help shows that
non-shell run is 2x faster.
Runtime of the XDP tests also shows improvement:
this patch: 2m34s 2m21s 2m18s 2m18s
before: 2m54s 2m36s 2m34s
Reviewed-by: Breno Leitao <leitao@debian.org>
Link: https://patch.msgid.link/20250830184317.696121-2-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch improves the utils.py module by removing unused imports
(errno, random), simplifying the fd_read_timeout() function by
eliminating unnecessary else clause, and cleaning up code style in the
defer class constructor.
Additionally, it renames the parameter in rand_port() from 'type' to
'stype' to avoid shadowing the built-in Python name 'type', improving
code clarity and preventing potential issues.
These changes enhance code readability and maintainability without
affecting functionality.
Signed-off-by: Breno Leitao <leitao@debian.org>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20250901-fix-v1-1-df0abb67481e@debian.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pull misc fixes from Andrew Morton:
"17 hotfixes. 13 are cc:stable and the remainder address post-6.16
issues or aren't considered necessary for -stable kernels. 11 of these
fixes are for MM.
This includes a three-patch series from Harry Yoo which fixes an
intermittent boot failure which can occur on x86 systems. And a
two-patch series from Alexander Gordeev which fixes a KASAN crash on
S390 systems"
* tag 'mm-hotfixes-stable-2025-09-01-17-20' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:
mm: fix possible deadlock in kmemleak
x86/mm/64: define ARCH_PAGE_TABLE_SYNC_MASK and arch_sync_kernel_mappings()
mm: introduce and use {pgd,p4d}_populate_kernel()
mm: move page table sync declarations to linux/pgtable.h
proc: fix missing pde_set_flags() for net proc files
mm: fix accounting of memmap pages
mm/damon/core: prevent unnecessary overflow in damos_set_effective_quota()
kexec: add KEXEC_FILE_NO_CMA as a legal flag
kasan: fix GCC mem-intrinsic prefix with sw tags
mm/kasan: avoid lazy MMU mode hazards
mm/kasan: fix vmalloc shadow memory (de-)population races
kunit: kasan_test: disable fortify string checker on kasan_strings() test
selftests/mm: fix FORCE_READ to read input value correctly
mm/userfaultfd: fix kmap_local LIFO ordering for CONFIG_HIGHPTE
ocfs2: prevent release journal inode after journal shutdown
rust: mm: mark VmaNew as transparent
of_numa: fix uninitialized memory nodes causing kernel panic
On 32bit ARM systems gcc-12 will use 32bit timestamps while gcc-13 and later
will use 64bit timestamps. The problem is that SYS_futex will continue
pointing at the 32bit system call. This makes the futex_wait test fail like
this:
waiter failed errno 110
not ok 1 futex_wake private returned: 0 Success
waiter failed errno 110
not ok 2 futex_wake shared (page anon) returned: 0 Success
waiter failed errno 110
not ok 3 futex_wake shared (file backed) returned: 0 Success
Instead of compiling differently depending on the gcc version, use the
-D_FILE_OFFSET_BITS=64 -D_TIME_BITS=64 options to ensure that 64bit timestamps
are used. Then use ifdefs to make SYS_futex point to the 64bit system call.
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: André Almeida <andrealmeid@igalia.com>
Tested-by: Anders Roxell <anders.roxell@linaro.org>
Link: https://lore.kernel.org/20250827130011.677600-6-bigeasy@linutronix.de
Fix format-security warnings by using proper format strings when passing
message variables to ksft_exit_fail_msg(), ksft_test_result_pass(), and
ksft_test_result_skip() function.
Thus prevent potential security issues and eliminate compiler warnings when
building with -Wformat-security.
Signed-off-by: Nai-Chen Cheng <bleach1827@gmail.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/20250827130011.677600-4-bigeasy@linutronix.de
The "Memory out of range" subtest of futex_numa_mpol assumes that memory
access outside of the mmap'ed area is invalid. That may not be the case
depending on the actual memory layout of the test application. When that
subtest was run on an x86-64 system with latest upstream kernel, the test
passed as an error was returned from futex_wake(). On another PowerPC system,
the same subtest failed because futex_wake() returned 0.
Bail out! futex2_wake(64, 0x86) should fail, but didn't
Looking further into the passed subtest on x86-64, it was found that an
-EINVAL was returned instead of -EFAULT. The -EINVAL error was returned
because the node value test with FLAGS_NUMA set failed with a node value
of 0x7f7f. IOW, the futex memory was accessible and futex_wake() failed
because the supposed node number wasn't valid. If that memory location
happens to have a very small value (e.g. 0), the test will pass and no
error will be returned.
Since this subtest is non-deterministic, drop it unless a guard page beyond
the mmap region is explicitly set.
The other problematic test is the "Memory too small" test. The futex_wake()
function returns the -EINVAL error code because the given futex address isn't
8-byte aligned, not because only 4 of the 8 bytes are valid and the other
4 bytes are not. So change the name of this subtest to "Mis-aligned futex" to
reflect the reality.
[ bp: Massage commit message. ]
Fixes: 3163369407 ("selftests/futex: Add futex_numa_mpol")
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: https://lore.kernel.org/20250827130011.677600-3-bigeasy@linutronix.de
While the GCC and Clang compilers already define __ASSEMBLER__
automatically when compiling assembler code, __ASSEMBLY__ is a
macro that only gets defined by the Makefiles in the kernel.
This is bad since macros starting with two underscores are names
that are reserved by the C language. It can also be very confusing
for the developers when switching between userspace and kernelspace
coding, or when dealing with uapi headers that rather should use
__ASSEMBLER__ instead. So let's standardize now on the __ASSEMBLER__
macro that is provided by the compilers.
This is almost a completely mechanical patch (done with a simple
"sed -i" statement), apart from tweaking two comments manually in
arch/powerpc/include/asm/bug.h and arch/powerpc/include/asm/kasan.h
(which did not have proper underscores at the end) and fixing a
checkpatch error about spaces in arch/powerpc/include/asm/spu_csa.h.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Madhavan Srinivasan <maddy@linux.ibm.com>
Link: https://patch.msgid.link/20250801082007.32904-3-thuth@redhat.com
Pull arm64 fixes from Catalin Marinas:
- CFI failure due to kpti_ng_pgd_alloc() signature mismatch
- Underallocation bug in the SVE ptrace kselftest
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
kselftest/arm64: Don't open code SVE_PT_SIZE() in fp-ptrace
arm64: mm: Fix CFI failure due to kpti_ng_pgd_alloc function signature
In fp-trace when allocating a buffer to write SVE register data we open
code the addition of the header size to the VL depeendent register data
size, which lead to an underallocation bug when we cut'n'pasted the code
for FPSIMD format writes. Use the SVE_PT_SIZE() macro that the kernel
UAPI provides for this.
Fixes: b84d2b2795 ("kselftest/arm64: Test FPSIMD format data writes via NT_ARM_SVE in fp-ptrace")
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20250812-arm64-fp-trace-macro-v1-1-317cfff986a5@kernel.org
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Jakub says:
nft_flowtable.sh is one of the most flake-atious test for netdev CI currently :(
The root cause is two-fold:
1. the failing part of the test is supposed to make sure that ip
fragments are forwarded for offloaded flows.
(flowtable has to pass them to classic forward path).
path mtu discovery for these subtests is disabled.
2. nft_flowtable.sh has two passes. One with fixed mtus/file size and
one where link mtus and file sizes are random.
The CI failures all have same pattern:
re-run with random mtus and file size: -o 27663 -l 4117 -r 10089 -s 54384840
[..]
PASS: dscp_egress: dscp packet counters match
FAIL: file mismatch for ns1 -> ns2
In some cases this error triggers a bit ealier, sometimes in a later
subtest:
re-run with random mtus and file size: -o 20201 -l 4555 -r 12657 -s 9405856
[..]
PASS: dscp_egress: dscp packet counters match
PASS: dscp_fwd: dscp packet counters match
2025/08/17 20:37:52 socat[18954] E write(7, 0x560716b96000, 8192): Broken pipe
FAIL: file mismatch for ns1 -> ns2
-rw------- 1 root root 9405856 Aug 17 20:36 /tmp/tmp.2n63vlTrQe
But all logs I saw show same scenario:
1. Failing tests have pmtu discovery off (i.e., ip fragmentation)
2. The test file is much larger than first-pass default (2M Byte)
3. peers have much larger MTUs compared to the 'network'.
These errors are very reproducible when re-running the test with
the same commandline arguments.
The timeout became much more prominent with
1d2fbaad7c ("tcp: stronger sk_rcvbuf checks"): reassembled packets
typically have a skb->truesize more than double the skb length.
As that commit is intentional and pmtud-off with
large-tcp-packets-as-fragments is not normal adjust the test to use a
smaller file for the pmtu-off subtests.
While at it, add more information to pass/fail messages and
also run the dscp alteration subtest with pmtu discovery enabled.
Link: https://netdev.bots.linux.dev/contest.html?test=nft-flowtable-sh
Fixes: f84ab63490 ("selftests: netfilter: nft_flowtable.sh: re-run with random mtu sizes")
Reported-by: Jakub Kicinski <kuba@kernel.org>
Closes: https://lore.kernel.org/netdev/20250822071330.4168f0db@kernel.org/
Signed-off-by: Florian Westphal <fw@strlen.de>
Link: https://patch.msgid.link/20250828214918.3385-1-fw@strlen.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pull kvm fixes from Paolo Bonzini:
"ARM:
- Correctly handle 'invariant' system registers for protected VMs
- Improved handling of VNCR data aborts, including external aborts
- Fixes for handling of FEAT_RAS for NV guests, providing a sane
fault context during SEA injection and preventing the use of
RASv1p1 fault injection hardware
- Ensure that page table destruction when a VM is destroyed gives an
opportunity to reschedule
- Large fix to KVM's infrastructure for managing guest context loaded
on the CPU, addressing issues where the output of AT emulation
doesn't get reflected to the guest
- Fix AT S12 emulation to actually perform stage-2 translation when
necessary
- Avoid attempting vLPI irqbypass when GICv4 has been explicitly
disabled for a VM
- Minor KVM + selftest fixes
RISC-V:
- Fix pte settings within kvm_riscv_gstage_ioremap()
- Fix comments in kvm_riscv_check_vcpu_requests()
- Fix stack overrun when setting vlenb via ONE_REG
x86:
- Use array_index_nospec() to sanitize the target vCPU ID when
handling PV IPIs and yields as the ID is guest-controlled.
- Drop a superfluous cpumask_empty() check when reclaiming SEV
memory, as the common case, by far, is that at least one CPU will
have entered the VM, and wbnoinvd_on_cpus_mask() will naturally
handle the rare case where the set of have_run_cpus is empty.
Selftests (not KVM):
- Rename the is_signed_type() macro in kselftest_harness.h to
is_signed_var() to fix a collision with linux/overflow.h. The
collision generates compiler warnings due to the two macros having
different meaning"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (29 commits)
KVM: arm64: nv: Fix ATS12 handling of single-stage translation
KVM: arm64: Remove __vcpu_{read,write}_sys_reg_{from,to}_cpu()
KVM: arm64: Fix vcpu_{read,write}_sys_reg() accessors
KVM: arm64: Simplify sysreg access on exception delivery
KVM: arm64: Check for SYSREGS_ON_CPU before accessing the 32bit state
RISC-V: KVM: fix stack overrun when loading vlenb
RISC-V: KVM: Correct kvm_riscv_check_vcpu_requests() comment
RISC-V: KVM: Fix pte settings within kvm_riscv_gstage_ioremap()
KVM: arm64: selftests: Sync ID_AA64MMFR3_EL1 in set_id_regs
KVM: arm64: Get rid of ARM64_FEATURE_MASK()
KVM: arm64: Make ID_AA64PFR1_EL1.RAS_frac writable
KVM: arm64: Make ID_AA64PFR0_EL1.RAS writable
KVM: arm64: Ignore HCR_EL2.FIEN set by L1 guest's EL2
KVM: arm64: Handle RASv1p1 registers
arm64: Add capability denoting FEAT_RASv1p1
KVM: arm64: Reschedule as needed when destroying the stage-2 page-tables
KVM: arm64: Split kvm_pgtable_stage2_destroy()
selftests: harness: Rename is_signed_type() to avoid collision with overflow.h
KVM: SEV: don't check have_run_cpus in sev_writeback_caches()
KVM: arm64: Correctly populate FAR_EL2 on nested SEA injection
...
Cross-merge networking fixes after downstream PR (net-6.17-rc4).
No conflicts.
Adjacent changes:
drivers/net/ethernet/intel/idpf/idpf_txrx.c
02614eee26 ("idpf: do not linearize big TSO packets")
6c4e684802 ("idpf: remove obsolete stashing code")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
KVM/arm64 changes for 6.17, take #2
- Correctly handle 'invariant' system registers for protected VMs
- Improved handling of VNCR data aborts, including external aborts
- Fixes for handling of FEAT_RAS for NV guests, providing a sane
fault context during SEA injection and preventing the use of
RASv1p1 fault injection hardware
- Ensure that page table destruction when a VM is destroyed gives an
opportunity to reschedule
- Large fix to KVM's infrastructure for managing guest context loaded
on the CPU, addressing issues where the output of AT emulation
doesn't get reflected to the guest
- Fix AT S12 emulation to actually perform stage-2 translation when
necessary
- Avoid attempting vLPI irqbypass when GICv4 has been explicitly
disabled for a VM
- Minor KVM + selftest fixes
The -g parameter was meant to the test the immutable global hash instead of the
private hash which has been made immutable. The global hash is tested as part
at the end of the regular test. The immutable private hash been removed.
Remove last traces of the immutable private hash.
Fixes: 16adc7f136 ("selftests/futex: Remove support for IMMUTABLE")
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: André Almeida <andrealmeid@igalia.com>
Link: https://lore.kernel.org/20250827130011.677600-2-bigeasy@linutronix.de
Pull block fixes from Jens Axboe:
- Fix a lockdep spotted issue on recursive locking for zoned writes, in
case of errors
- Update bcache MAINTAINERS entry address for Coly
- Fix for a ublk release issue, with selftests
- Fix for a regression introduced in this cycle, where it assumed
q->rq_qos was always set if the bio flag indicated that
- Fix for a regression introduced in this cycle, where loop retrieving
block device sizes got broken
* tag 'block-6.17-20250828' of git://git.kernel.dk/linux:
bcache: change maintainer's email address
ublk selftests: add --no_ublk_fixed_fd for not using registered ublk char device
ublk: avoid ublk_io_release() called after ublk char dev is closed
block: validate QoS before calling __rq_qos_done_bio()
blk-zoned: Fix a lockdep complaint about recursive locking
loop: fix zero sized loop for block special file
Commit 0d6ccfe6b3 ("selftests: drv-net: rss_ctx: check for all-zero keys")
added a skip exception if NIC has fewer than 3 queues enabled,
but it's just constructing the object, it's not actually rising
this exception.
Before:
# Exception| net.lib.py.utils.CmdExitFailure: Command failed: ethtool -X enp1s0 equal 3 hkey d1:cc:77:47:9d:ea:15:f2:b9:6c:ef:68:62:c0:45:d5:b0:99:7d:cf:29:53:40:06:3d:8e:b9:bc:d4:70:89:b8:8d:59:04:ea:a9:c2:21:b3:55:b8:ab:6b:d9:48:b4:bd:4c:ff:a5:f0:a8:c2
not ok 1 rss_ctx.test_rss_key_indir
After:
ok 1 rss_ctx.test_rss_key_indir # SKIP Device has fewer than 3 queues (or doesn't support queue stats)
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20250827173558.3259072-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Fix the rules in tools/testing/selftests/vfio/.gitignore to not ignore
some already tracked files (.gitignore, Makefile, lib/libvfio.mk).
This change should be a no-op, since these files are already tracked by git
and thus git will not ignore updates to them even though they match the
ignore rules in the VFIO selftests .gitignore file.
However, they do generate warnings with W=1, as reported by the kernel test
robot.
$ KBUILD_EXTRA_WARN=1 scripts/misc-check
tools/testing/selftests/vfio/.gitignore: warning: ignored by one of the .gitignore files
tools/testing/selftests/vfio/Makefile: warning: ignored by one of the .gitignore files
tools/testing/selftests/vfio/lib/libvfio.mk: warning: ignored by one of the .gitignore files
Fix this by explicitly un-ignoring the tracked files.
Fixes: 292e9ee22b ("selftests: Create tools/testing/selftests/vfio")
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202508280918.rFRyiLEU-lkp@intel.com/
Signed-off-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20250828185815.382215-1-dmatlack@google.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Add a new command line option --no_ublk_fixed_fd that excludes the ublk
control device (/dev/ublkcN) from io_uring's registered files array.
When this option is used, only backing files are registered starting
from index 1, while the ublk control device is accessed using its raw
file descriptor.
Add ublk_get_registered_fd() helper function that returns the appropriate
file descriptor for use with io_uring operations.
Key optimizations implemented:
- Cache UBLKS_Q_NO_UBLK_FIXED_FD flag in ublk_queue.flags to avoid
reading dev->no_ublk_fixed_fd in fast path
- Cache ublk char device fd in ublk_queue.ublk_fd for fast access
- Update ublk_get_registered_fd() to use ublk_queue * parameter
- Update io_uring_prep_buf_register/unregister() to use ublk_queue *
- Replace ublk_device * access with ublk_queue * access in fast paths
Also pass --no_ublk_fixed_fd to test_stress_04.sh for covering
plain ublk char device mode.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20250827121602.2619736-3-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
When a packet flood hits one or more UDP sockets, many cpus
have to update sk->sk_drops.
This slows down other cpus, because currently
sk_drops is in sock_write_rx group.
Add a socket_drop_counters structure to udp sockets.
Using dedicated cache lines to hold drop counters
makes sure that consumers no longer suffer from
false sharing if/when producers only change sk->sk_drops.
This adds 128 bytes per UDP socket.
Tested with the following stress test, sending about 11 Mpps
to a dual socket AMD EPYC 7B13 64-Core.
super_netperf 20 -t UDP_STREAM -H DUT -l10 -- -n -P,1000 -m 120
Note: due to socket lookup, only one UDP socket is receiving
packets on DUT.
Then measure receiver (DUT) behavior. We can see both
consumer and BH handlers can process more packets per second.
Before:
nstat -n ; sleep 1 ; nstat | grep Udp
Udp6InDatagrams 615091 0.0
Udp6InErrors 3904277 0.0
Udp6RcvbufErrors 3904277 0.0
After:
nstat -n ; sleep 1 ; nstat | grep Udp
Udp6InDatagrams 816281 0.0
Udp6InErrors 7497093 0.0
Udp6RcvbufErrors 7497093 0.0
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com>
Link: https://patch.msgid.link/20250826125031.1578842-5-edumazet@google.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Add benchmarks for the standard set of operations: LOOKUP, INSERT,
UPDATE, DELETE. Also include benchmarks to measure the overhead of the
bench framework itself (NOOP) as well as the overhead of generating keys
(BASELINE). Lastly, this includes a benchmark for FREE (trie_free())
which is known to have terrible performance for maps with many entries.
Benchmarks operate on tries without gaps in the key range, i.e. each
test begins or ends with a trie with valid keys in the range [0,
nr_entries). This is intended to cause maximum branching when traversing
the trie.
LOOKUP, UPDATE, DELETE, and FREE fill a BPF LPM trie from userspace
using bpf_map_update_batch() and run the corresponding benchmark
operation via bpf_loop(). INSERT starts with an empty map and fills it
kernel-side from bpf_loop(). FREE records the time to free a filled LPM
trie by attaching and destroying a BPF prog. NOOP measures the overhead
of the test harness by running an empty function with bpf_loop().
BASELINE is similar to NOOP except that the function generates a key.
Each operation runs 10,000 times using bpf_loop(). Note that this value
is intentionally independent of the number of entries in the LPM trie so
that the stability of the results isn't affected by the number of
entries.
For those benchmarks that need to reset the LPM trie once it's full
(INSERT) or empty (DELETE), throughput and latency results are scaled by
the fraction of a second the operation actually ran to ignore any time
spent reinitialising the trie.
By default, benchmarks run using sequential keys in the range [0,
nr_entries). BASELINE, LOOKUP, and UPDATE can use random keys via the
--random parameter but beware there is a runtime cost involved in
generating random keys. Other benchmarks are prohibited from using
random keys because it can skew the results, e.g. when inserting an
existing key or deleting a missing one.
All measurements are recorded from within the kernel to eliminate
syscall overhead. Most benchmarks run an XDP program to generate stats
but FREE needs to collect latencies using fentry/fexit on
map_free_deferred() because it's not possible to use fentry directly on
lpm_trie.c since commit c83508da56 ("bpf: Avoid deadlock caused by
nested kprobe and fentry bpf programs") and there's no way to
create/destroy a map from within an XDP program.
Here is example output from an AMD EPYC 9684X 96-Core machine for each
of the benchmarks using a trie with 10K entries and a 32-bit prefix
length, e.g.
$ ./bench lpm-trie-$op \
--prefix_len=32 \
--producers=1 \
--nr_entries=10000
noop: throughput 74.417 ± 0.032 M ops/s ( 74.417M ops/prod), latency 13.438 ns/op
baseline: throughput 70.107 ± 0.171 M ops/s ( 70.107M ops/prod), latency 14.264 ns/op
lookup: throughput 8.467 ± 0.047 M ops/s ( 8.467M ops/prod), latency 118.109 ns/op
insert: throughput 2.440 ± 0.015 M ops/s ( 2.440M ops/prod), latency 409.290 ns/op
update: throughput 2.806 ± 0.042 M ops/s ( 2.806M ops/prod), latency 356.322 ns/op
delete: throughput 4.625 ± 0.011 M ops/s ( 4.625M ops/prod), latency 215.613 ns/op
free: throughput 0.578 ± 0.006 K ops/s ( 0.578K ops/prod), latency 1.730 ms/op
And the same benchmarks using random keys:
$ ./bench lpm-trie-$op \
--prefix_len=32 \
--producers=1 \
--nr_entries=10000 \
--random
noop: throughput 74.259 ± 0.335 M ops/s ( 74.259M ops/prod), latency 13.466 ns/op
baseline: throughput 35.150 ± 0.144 M ops/s ( 35.150M ops/prod), latency 28.450 ns/op
lookup: throughput 7.119 ± 0.048 M ops/s ( 7.119M ops/prod), latency 140.469 ns/op
insert: N/A
update: throughput 2.736 ± 0.012 M ops/s ( 2.736M ops/prod), latency 365.523 ns/op
delete: N/A
free: N/A
Signed-off-by: Matt Fleming <mfleming@cloudflare.com>
Signed-off-by: Jesper Dangaard Brouer <hawk@kernel.org>
Link: https://lore.kernel.org/r/20250827140149.1001557-1-matt@readmodwrite.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
When using GCC on x86-64 to compile an usdt prog with -O1 or higher
optimization, the compiler will generate SIB addressing mode for global
array, e.g. "1@-96(%rbp,%rax,8)".
In this patch:
- enrich subtest_basic_usdt test case to cover SIB addressing usdt argument spec
handling logic
Signed-off-by: Jiawei Zhao <phoenix500526@163.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20250827053128.1301287-3-phoenix500526@163.com
Introduce run.sh, a script to help with running VFIO selftests. The
script is intended to be used for both humans manually running VFIO
selftests, and to incorporate into test automation where VFIO selftests
may run alongside other tests. As such the script aims to be hermetic,
returning the system to the state it was before the test started.
The script takes as input the BDF of a device to use and a command to
run (typically the command would be a VFIO selftest). e.g.
$ ./run.sh -d 0000:6a:01.0 ./vfio_pci_device_test
or
$ ./run.sh -d 0000:6a:01.0 -- ./vfio_pci_device_test
The script then handles unbinding device 0000:6a:01.0 from its current
driver, binding it to vfio-pci, running the test, unbinding from
vfio-pci, and binding back to the original driver.
When run.sh runs the provided test, it does so by appending the BDF as
the last parameter. For example:
$ ./run.sh -d 0000:6a:01.0 -- echo hello
Results in the following being printed to stdout:
hello 0000:6a:01.0
The script also supports a mode where it can break out into a shell so
that multiple tests can be run manually.
$ ./run.sh -d 0000:6a:01.0 -s
$ echo $VFIO_SELFTESTS_BDF
$ ./vfio_pci_device_test
$ exit
Choosing which device to use is up to the user.
In the future this script should be extensible to tests that want to use
multiple devices. The script can support accepting -d BDF multiple times
and parse them into an array, setup all the devices, pass the list of
BDFs to the test, and then cleanup all the devices.
Acked-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20250822212518.4156428-31-dmatlack@google.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
