June 2023 - Linux-kselftest-mirror

[PATCH v2 0/3] drivers: base: Add tests showing devm handling inconsistencies

by Maxime Ripard

Hi, This follows the discussion here: https://lore.kernel.org/linux-kselftest/20230324123157.bbwvfq4gsxnlnfwb@hou… This shows a couple of inconsistencies with regard to how device-managed resources are cleaned up. Basically, devm resources will only be cleaned up if the device is attached to a bus and bound to a driver. Failing any of these cases, a call to device_unregister will not end up in the devm resources being released. We had to work around it in DRM to provide helpers to create a device for kunit tests, but the current discussion around creating similar, generic, helpers for kunit resumed interest in fixing this. This can be tested using the command: ./tools/testing/kunit/kunit.py run --kunitconfig=drivers/base/test/ I added the fix David suggested back in that discussion which does fix the tests. The SoB is missing, since David didn't provide it back then. Let me know what you think, Maxime Signed-off-by: Maxime Ripard <maxime(a)cerno.tech> --- Changes in v2: - Use an init function - Document the tests - Add a fix for the bugs - Link to v1: https://lore.kernel.org/r/20230329-kunit-devm-inconsistencies-test-v1-0-c33… --- David Gow (1): drivers: base: Free devm resources when unregistering a device Maxime Ripard (2): drivers: base: Add basic devm tests for root devices drivers: base: Add basic devm tests for platform devices drivers/base/core.c | 11 ++ drivers/base/test/.kunitconfig | 2 + drivers/base/test/Kconfig | 4 + drivers/base/test/Makefile | 3 + drivers/base/test/platform-device-test.c | 220 +++++++++++++++++++++++++++++++ drivers/base/test/root-device-test.c | 108 +++++++++++++++ 6 files changed, 348 insertions(+) --- base-commit: 53cdf865f90ba922a854c65ed05b519f9d728424 change-id: 20230329-kunit-devm-inconsistencies-test-5e5a7d01e60d Best regards, -- Maxime Ripard <mripard(a)kernel.org>

9 months, 1 week

2
11
0 0

[PATCH v2 1/3] bpf: Allow NULL buffers in bpf_dynptr_slice(_rw)

by Daniel Rosenberg

bpf_dynptr_slice(_rw) uses a user provided buffer if it can not provide a pointer to a block of contiguous memory. This buffer is unused in the case of local dynptrs, and may be unused in other cases as well. There is no need to require the buffer, as the kfunc can just return NULL if it was needed and not provided. This adds another kfunc annotation, __opt, which combines with __sz and __szk to allow the buffer associated with the size to be NULL. If the buffer is NULL, the verifier does not check that the buffer is of sufficient size. Signed-off-by: Daniel Rosenberg <drosen(a)google.com> --- Documentation/bpf/kfuncs.rst | 23 ++++++++++++++++++++++- include/linux/skbuff.h | 2 +- kernel/bpf/helpers.c | 30 ++++++++++++++++++------------ kernel/bpf/verifier.c | 17 +++++++++++++---- 4 files changed, 54 insertions(+), 18 deletions(-) diff --git a/Documentation/bpf/kfuncs.rst b/Documentation/bpf/kfuncs.rst index ea2516374d92..7a3d9de5f315 100644 --- a/Documentation/bpf/kfuncs.rst +++ b/Documentation/bpf/kfuncs.rst @@ -100,7 +100,7 @@ Hence, whenever a constant scalar argument is accepted by a kfunc which is not a size parameter, and the value of the constant matters for program safety, __k suffix should be used. -2.2.2 __uninit Annotation +2.2.3 __uninit Annotation ------------------------- This annotation is used to indicate that the argument will be treated as @@ -117,6 +117,27 @@ Here, the dynptr will be treated as an uninitialized dynptr. Without this annotation, the verifier will reject the program if the dynptr passed in is not initialized. +2.2.4 __opt Annotation +------------------------- + +This annotation is used to indicate that the buffer associated with an __sz or __szk +argument may be null. If the function is passed a nullptr in place of the buffer, +the verifier will not check that length is appropriate for the buffer. The kfunc is +responsible for checking if this buffer is null before using it. + +An example is given below:: + + __bpf_kfunc void *bpf_dynptr_slice(..., void *buffer__opt, u32 buffer__szk) + { + ... + } + +Here, the buffer may be null. If buffer is not null, it at least of size buffer_szk. +Either way, the returned buffer is either NULL, or of size buffer_szk. Without this +annotation, the verifier will reject the program if a null pointer is passed in with +a nonzero size. + + .. _BPF_kfunc_nodef: 2.3 Using an existing kernel function diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 738776ab8838..8ddb4af1a501 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -4033,7 +4033,7 @@ __skb_header_pointer(const struct sk_buff *skb, int offset, int len, if (likely(hlen - offset >= len)) return (void *)data + offset; - if (!skb || unlikely(skb_copy_bits(skb, offset, buffer, len) < 0)) + if (!skb || !buffer || unlikely(skb_copy_bits(skb, offset, buffer, len) < 0)) return NULL; return buffer; diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c index 8d368fa353f9..26efb6fbeab2 100644 --- a/kernel/bpf/helpers.c +++ b/kernel/bpf/helpers.c @@ -2167,13 +2167,15 @@ __bpf_kfunc struct task_struct *bpf_task_from_pid(s32 pid) * bpf_dynptr_slice() - Obtain a read-only pointer to the dynptr data. * @ptr: The dynptr whose data slice to retrieve * @offset: Offset into the dynptr - * @buffer: User-provided buffer to copy contents into - * @buffer__szk: Size (in bytes) of the buffer. This is the length of the - * requested slice. This must be a constant. + * @buffer__opt: User-provided buffer to copy contents into. May be NULL + * @buffer__szk: Size (in bytes) of the buffer if present. This is the + * length of the requested slice. This must be a constant. * * For non-skb and non-xdp type dynptrs, there is no difference between * bpf_dynptr_slice and bpf_dynptr_data. * + * If buffer__opt is NULL, the call will fail if buffer_opt was needed. + * * If the intention is to write to the data slice, please use * bpf_dynptr_slice_rdwr. * @@ -2190,7 +2192,7 @@ __bpf_kfunc struct task_struct *bpf_task_from_pid(s32 pid) * direct pointer) */ __bpf_kfunc void *bpf_dynptr_slice(const struct bpf_dynptr_kern *ptr, u32 offset, - void *buffer, u32 buffer__szk) + void *buffer__opt, u32 buffer__szk) { enum bpf_dynptr_type type; u32 len = buffer__szk; @@ -2210,15 +2212,17 @@ __bpf_kfunc void *bpf_dynptr_slice(const struct bpf_dynptr_kern *ptr, u32 offset case BPF_DYNPTR_TYPE_RINGBUF: return ptr->data + ptr->offset + offset; case BPF_DYNPTR_TYPE_SKB: - return skb_header_pointer(ptr->data, ptr->offset + offset, len, buffer); + return skb_header_pointer(ptr->data, ptr->offset + offset, len, buffer__opt); case BPF_DYNPTR_TYPE_XDP: { void *xdp_ptr = bpf_xdp_pointer(ptr->data, ptr->offset + offset, len); if (xdp_ptr) return xdp_ptr; - bpf_xdp_copy_buf(ptr->data, ptr->offset + offset, buffer, len, false); - return buffer; + if (!buffer__opt) + return NULL; + bpf_xdp_copy_buf(ptr->data, ptr->offset + offset, buffer__opt, len, false); + return buffer__opt; } default: WARN_ONCE(true, "unknown dynptr type %d\n", type); @@ -2230,13 +2234,15 @@ __bpf_kfunc void *bpf_dynptr_slice(const struct bpf_dynptr_kern *ptr, u32 offset * bpf_dynptr_slice_rdwr() - Obtain a writable pointer to the dynptr data. * @ptr: The dynptr whose data slice to retrieve * @offset: Offset into the dynptr - * @buffer: User-provided buffer to copy contents into - * @buffer__szk: Size (in bytes) of the buffer. This is the length of the - * requested slice. This must be a constant. + * @buffer__opt: User-provided buffer to copy contents into. May be NULL + * @buffer__szk: Size (in bytes) of the buffer if present. This is the + * length of the requested slice. This must be a constant. * * For non-skb and non-xdp type dynptrs, there is no difference between * bpf_dynptr_slice and bpf_dynptr_data. * + * If buffer__opt is NULL, the call will fail if buffer_opt was needed. + * * The returned pointer is writable and may point to either directly the dynptr * data at the requested offset or to the buffer if unable to obtain a direct * data pointer to (example: the requested slice is to the paged area of an skb @@ -2267,7 +2273,7 @@ __bpf_kfunc void *bpf_dynptr_slice(const struct bpf_dynptr_kern *ptr, u32 offset * direct pointer) */ __bpf_kfunc void *bpf_dynptr_slice_rdwr(const struct bpf_dynptr_kern *ptr, u32 offset, - void *buffer, u32 buffer__szk) + void *buffer__opt, u32 buffer__szk) { if (!ptr->data || bpf_dynptr_is_rdonly(ptr)) return NULL; @@ -2294,7 +2300,7 @@ __bpf_kfunc void *bpf_dynptr_slice_rdwr(const struct bpf_dynptr_kern *ptr, u32 o * will be copied out into the buffer and the user will need to call * bpf_dynptr_write() to commit changes. */ - return bpf_dynptr_slice(ptr, offset, buffer, buffer__szk); + return bpf_dynptr_slice(ptr, offset, buffer__opt, buffer__szk); } __bpf_kfunc void *bpf_cast_to_kern_ctx(void *obj) diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index fbcf5a4e2fcd..708ae7bca1fe 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -9398,6 +9398,11 @@ static bool is_kfunc_arg_const_mem_size(const struct btf *btf, return __kfunc_param_match_suffix(btf, arg, "__szk"); } +static bool is_kfunc_arg_optional(const struct btf *btf, const struct btf_param *arg) +{ + return __kfunc_param_match_suffix(btf, arg, "__opt"); +} + static bool is_kfunc_arg_constant(const struct btf *btf, const struct btf_param *arg) { return __kfunc_param_match_suffix(btf, arg, "__k"); @@ -10464,13 +10469,17 @@ static int check_kfunc_args(struct bpf_verifier_env *env, struct bpf_kfunc_call_ break; case KF_ARG_PTR_TO_MEM_SIZE: { + struct bpf_reg_state *buff_reg = &regs[regno]; + const struct btf_param *buff_arg = &args[i]; struct bpf_reg_state *size_reg = &regs[regno + 1]; const struct btf_param *size_arg = &args[i + 1]; - ret = check_kfunc_mem_size_reg(env, size_reg, regno + 1); - if (ret < 0) { - verbose(env, "arg#%d arg#%d memory, len pair leads to invalid memory access\n", i, i + 1); - return ret; + if (!register_is_null(buff_reg) || !is_kfunc_arg_optional(meta->btf, buff_arg)) { + ret = check_kfunc_mem_size_reg(env, size_reg, regno + 1); + if (ret < 0) { + verbose(env, "arg#%d arg#%d memory, len pair leads to invalid memory access\n", i, i + 1); + return ret; + } } if (is_kfunc_arg_const_mem_size(meta->btf, size_arg, size_reg)) { base-commit: 6e98b09da931a00bf4e0477d0fa52748bf28fcce -- 2.40.1.495.gc816e09b53d-goog

9 months, 1 week

5
17
0 0

[PATCH net-next 0/4] nexthop: Refactor and fix nexthop selection for multipath routes

by Benjamin Poirier

In order to select a nexthop for multipath routes, fib_select_multipath() is used with legacy nexthops and nexthop_select_path_hthr() is used with nexthop objects. Those two functions perform a validity test on the neighbor related to each nexthop but their logic is structured differently. This causes a divergence in behavior and nexthop_select_path_hthr() may return a nexthop that failed the neighbor validity test even if there was one that passed. Refactor nexthop_select_path_hthr() to make it more similar to fib_select_multipath() and fix the problem mentioned above. Benjamin Poirier (4): nexthop: Factor out hash threshold fdb nexthop selection nexthop: Factor out neighbor validity check nexthop: Do not return invalid nexthop object during multipath selection selftests: net: Add test cases for nexthop groups with invalid neighbors net/ipv4/nexthop.c | 64 +++++++--- tools/testing/selftests/net/fib_nexthops.sh | 129 ++++++++++++++++++++ 2 files changed, 174 insertions(+), 19 deletions(-) -- 2.40.1

9 months, 1 week

3
10
0 0

nolibc: KTAP output and test reports

by Thomas Weißschuh

Hi Willy, Zhangjin, after your recent discussions about the test output and report I wondered if it would make sense to switch nolibc-test to KTAP output format [0]. With this it would be possible to have a wrapper script run each architecture test as its own test subcomponent. A (K)TAP parser/runner could then directly recognize and report failing testcases, making it easier to validate. Also maybe we can hook it up into the regular kselftests setup and have the bots run it as part of that. The kernel even includes a header-only library to implement the format [1]. It also should be fairly easy to emit the format without a library. Thomas [0] Documentation/dev-tools/ktap.rst [1] Documentation/dev-tools/kselftest.rst (Test harness)

9 months, 2 weeks

3
3
0 0

[PATCH v1 00/22] selftests/nolibc: add minimal kernel config support

by Zhangjin Wu

Willy, Thomas We just sent the 'selftests/nolibc: allow run with minimal kernel config' series [1], Here is the 'minimal' kernel config support, with both of them, it is possible to run nolibc-test for all architectures with oneline command and in less than ~30 minutes - 1 hour (not fullly measured yet): // run with tiny config + qemu-system // Note: rv32 and loongarch require to download the bios at first $ time make run-tiny-all QUIET_RUN=1 // run with default config + qemu-system $ time make run-default-all QUIET_RUN=1 // run with qemu-user $ time make run-user-all QUIET_RUN=1 Besides the 'tinyconfig' suggestion from Thomas, this patch also merge the generic part of my local powerpc porting (the extconfig to add additional console support). This is applied after the test report patchset [2] and the rv32 compile patchset [3], because all of them touched the same Makefile. Even without the 'selftests/nolibc: allow run with minimal kernel config' series [1], all of the tests can pass except the /proc/self/net related ones (We haven't enable CONFIG_NET in this patchset), the chmod_net one will be removed by Thomas from this patchset [4] for the wrong chmodable attribute issue of /proc/self/net, the link_cross one can be simply fixed up by using another /proc/self interface (like /proc/self/cmdline), which will be covered in our revision of the [1] series. Beside the core 'minimal' config support, some generic patch are added together to avoid patch conflicts. * selftests/nolibc: add test for -include /path/to/nolibc.h Add a test switch to allow run nolibc-test with nolibc.h * selftests/nolibc: print result to the screen too Let the run targets print results by default, allow disable by QUIET_RUN=1 * selftests/nolibc: allow use x86_64 toolchain for i386 Allow use x86_64 toolchains for i386 * selftests/nolibc: add menuconfig target for manual config a new 'menuconfig' target added for development and debugging * selftests/nolibc: add tinyconfig target a new 'tinyconfig' compare to 'defconfig', smaller and faster, but not enough for boot and print, require following 'extconfig' target * selftests/nolibc: allow customize extra kernel config options a new 'extconfig' allows to add extra config options for 'defconfig' and 'tinyconfig' * selftests/nolibc: add common extra config options selftests/nolibc: add power reset control support selftests/nolibc: add procfs, shmem and tmpfs Add common extra configs, the 3rd one (procfs, shmem and tmpfs) can be completely reverted after [1] series, but as discuss with Thomas, procfs may be still a hard requirement. * selftests/nolibc: add extra configs for i386 selftests/nolibc: add extra configs for x86_64 selftests/nolibc: add extra configs for arm64 selftests/nolibc: add extra configs for arm selftests/nolibc: add extra configs for mips selftests/nolibc: add extra configs for riscv32 selftests/nolibc: add extra configs for riscv64 selftests/nolibc: add extra configs for s390x selftests/nolibc: add extra configs for loongarch Add architecture specific extra configs to let kernel boot and nolibc-test print. The rv32 added here is only for test, it should not be merged before the missing 64bit syscalls are added (still wait for the merging of the __sysret and -ENOSYS patches). * selftests/nolibc: config default CROSS_COMPILE selftests/nolibc: add run-tiny and run-default both run-tiny and run-default are added to do config and run together, this easier test a log. * selftests/nolibc: allow run tests on all targets selftests/nolibc: detect bios existing to avoid hang Further allow do run-user, run-tiny and run-default for all architectures at once, the -all suffix is added to do so. Since some generic patches are still in review, before sending the left rv32 patches, I'm will send more generic patches later, the coming one is arch-xxx.h cleanup, and then, the 32bit powerpc porting support. For the compile speedup, the next step may be add architecture specific 'O' support, which may allow us rerun across architectures without mrproper, for a single architecture development, this 'minimal' config should be enough ;-) Thanks. Best regards, Zhangjin --- [1]: https://lore.kernel.org/lkml/cover.1687344643.git.falcon@tinylab.org/ [2]: https://lore.kernel.org/lkml/cover.1687156559.git.falcon@tinylab.org/ [3]: https://lore.kernel.org/linux-riscv/cover.1687176996.git.falcon@tinylab.org/ [4]: https://lore.kernel.org/lkml/20230624-proc-net-setattr-v1-0-73176812adee@we… Zhangjin Wu (22): selftests/nolibc: add test for -include /path/to/nolibc.h selftests/nolibc: print result to the screen too selftests/nolibc: allow use x86_64 toolchain for i386 selftests/nolibc: add menuconfig target for manual config selftests/nolibc: add tinyconfig target selftests/nolibc: allow customize extra kernel config options selftests/nolibc: add common extra config options selftests/nolibc: add power reset control support selftests/nolibc: add procfs, shmem and tmpfs selftests/nolibc: add extra configs for i386 selftests/nolibc: add extra configs for x86_64 selftests/nolibc: add extra configs for arm64 selftests/nolibc: add extra configs for arm selftests/nolibc: add extra configs for mips selftests/nolibc: add extra configs for riscv32 selftests/nolibc: add extra configs for riscv64 selftests/nolibc: add extra configs for s390x selftests/nolibc: add extra configs for loongarch selftests/nolibc: config default CROSS_COMPILE selftests/nolibc: add run-tiny and run-default selftests/nolibc: allow run tests on all targets selftests/nolibc: detect bios existing to avoid hang tools/testing/selftests/nolibc/Makefile | 125 ++++++++++++++++++++++-- 1 file changed, 119 insertions(+), 6 deletions(-) -- 2.25.1

9 months, 2 weeks

3
30
0 0

[PATCH 0/2] kselftest/alsa: Decrease pcm-test duration to avoid timeouts

by Nícolas F. R. A. Prado

This series decreases the pcm-test duration in order to avoid timeouts by first moving the audio stream duration to a variable and subsequently decreasing it. Nícolas F. R. A. Prado (2): kselftest/alsa: pcm-test: Move stream duration and margin to variables kselftest/alsa: pcm-test: Decrease stream duration from 4 to 2 seconds tools/testing/selftests/alsa/pcm-test.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) -- 2.41.0

9 months, 2 weeks

5
15
0 0

[PATCH] Documentation: mm/memfd: vm.memfd_noexec

by jeffxu＠chromium.org

From: Jeff Xu <jeffxu(a)google.com> Add documentation for sysctl vm.memfd_noexec Link:https://lore.kernel.org/linux-mm/CABi2SkXUX_QqTQ10Yx9bBUGpN1wByOi_=gZU… Reported-by: Dominique Martinet <asmadeus(a)codewreck.org> Signed-off-by: Jeff Xu <jeffxu(a)google.com> --- Documentation/admin-guide/sysctl/vm.rst | 29 +++++++++++++++++++++++++ 1 file changed, 29 insertions(+) diff --git a/Documentation/admin-guide/sysctl/vm.rst b/Documentation/admin-guide/sysctl/vm.rst index 45ba1f4dc004..71923c3d7044 100644 --- a/Documentation/admin-guide/sysctl/vm.rst +++ b/Documentation/admin-guide/sysctl/vm.rst @@ -424,6 +424,35 @@ e.g., up to one or two maps per allocation. The default value is 65530. +memfd_noexec: +============= +This pid namespaced sysctl controls memfd_create(). + +The new MFD_NOEXEC_SEAL and MFD_EXEC flags of memfd_create() allows +application to set executable bit at creation time. + +When MFD_NOEXEC_SEAL is set, memfd is created without executable bit +(mode:0666), and sealed with F_SEAL_EXEC, so it can't be chmod to +be executable (mode: 0777) after creation. + +when MFD_EXEC flag is set, memfd is created with executable bit +(mode:0777), this is the same as the old behavior of memfd_create. + +The new pid namespaced sysctl vm.memfd_noexec has 3 values: +0: memfd_create() without MFD_EXEC nor MFD_NOEXEC_SEAL acts like + MFD_EXEC was set. +1: memfd_create() without MFD_EXEC nor MFD_NOEXEC_SEAL acts like + MFD_NOEXEC_SEAL was set. +2: memfd_create() without MFD_NOEXEC_SEAL will be rejected. + +The default value is 0. + +Once set, it can't be downgraded at runtime, i.e. 2=>1, 1=>0 +are denied. + +This is pid namespaced sysctl, child processes inherit the parent +process's pid at the time of fork. Changes to the parent process +after fork are not automatically propagated to the child process. memory_failure_early_kill: ========================== -- 2.41.0.255.g8b1d071c50-goog

9 months, 2 weeks

2
1
0 0

[RESEND PATCH] selftests/mincore: fix skip condition for check_huge_pages test

by Ricardo Cañuelo

The check_huge_pages test was failing instead of skipping on qemu-armv7 because the skip condition wasn't handled properly. Add an additional check to fix it. Signed-off-by: Ricardo Cañuelo <ricardo.canuelo(a)collabora.com> Reported-by: Naresh Kamboju <naresh.kamboju(a)linaro.org> Reported-by: Linux Kernel Functional Testing <lkft(a)linaro.org> Reviewed-by: Muhammad Usama Anjum <usama.anjum(a)collabora.com> Tested-by: Anders Roxell <anders.roxell(a)linaro.org> Closes: https://lore.kernel.org/all/CA+G9fYuoB8Ug8PcTU-YGmemL7_eeEksXFihvxWF6OikD7s… --- tools/testing/selftests/mincore/mincore_selftest.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/mincore/mincore_selftest.c b/tools/testing/selftests/mincore/mincore_selftest.c index 4c88238fc8f0..6fb3eea5b6ee 100644 --- a/tools/testing/selftests/mincore/mincore_selftest.c +++ b/tools/testing/selftests/mincore/mincore_selftest.c @@ -150,8 +149,8 @@ TEST(check_huge_pages) MAP_PRIVATE | MAP_ANONYMOUS | MAP_HUGETLB, -1, 0); if (addr == MAP_FAILED) { - if (errno == ENOMEM) - SKIP(return, "No huge pages available."); + if (errno == ENOMEM || errno == EINVAL) + SKIP(return, "No huge pages available or CONFIG_HUGETLB_PAGE disabled."); else TH_LOG("mmap error: %s", strerror(errno)); } -- 2.25.1

9 months, 2 weeks

2
1
0 0

[PATCH 0/2] proc: proc_setattr for /proc/$PID/net

by Thomas Weißschuh

/proc/$PID/net currently allows the setting of file attributes, in contrast to other /proc/$PID/ files and directories. This would break the nolibc testsuite so the first patch in the series removes the offending testcase. The "fix" for nolibc-test is intentionally kept trivial as the series will most likely go through the filesystem tree and if conflicts arise, it is obvious on how to resolve them. Technically this can lead to breakage of nolibc-test if an old nolibc-test is used with a newer kernel containing the fix. Note: Except for /proc itself this is the only "struct inode_operations" in fs/proc/ that is missing an implementation of setattr(). Signed-off-by: Thomas Weißschuh <linux(a)weissschuh.net> --- Thomas Weißschuh (2): selftests/nolibc: drop test chmod_net proc: use generic setattr() for /proc/$PID/net fs/proc/proc_net.c | 1 + tools/testing/selftests/nolibc/nolibc-test.c | 1 - 2 files changed, 1 insertion(+), 1 deletion(-) --- base-commit: a92b7d26c743b9dc06d520f863d624e94978a1d9 change-id: 20230624-proc-net-setattr-8f0a6b8eb2f5 Best regards, -- Thomas Weißschuh <linux(a)weissschuh.net>

9 months, 2 weeks

5
15
0 0

[PATCH v4 0/9] cgroup/cpuset: Support remote partitions

by Waiman Long

v4: - [v3] https://lore.kernel.org/lkml/20230627005529.1564984-1-longman@redhat.com/ - Fix compilation problem reported by kernel test robot. v3: - [v2] https://lore.kernel.org/lkml/20230531163405.2200292-1-longman@redhat.com/ - Change the new control file from root-only "cpuset.cpus.reserve" to non-root "cpuset.cpus.exclusive" which lists the set of exclusive CPUs distributed down the hierarchy. - Add a patch to restrict boot-time isolated CPUs to isolated partitions only. - Update the test_cpuset_prs.sh test script and documentation accordingly. This patch series introduces a new cpuset control file "cpuset.cpus.exclusive" which must be a subset of "cpuset.cpus" and the parent's "cpuset.cpus.exclusive". This control file lists the exclusive CPUs to be distributed down the hierarchy. Any one of the exclusive CPUs can only be distributed to at most one child cpuset. Unlike "cpuset.cpus", invalid input to "cpuset.cpus.exclusive" will be rejected with an error. This new control file has no effect on the behavior of the cpuset until it turns into a partition root. At that point, its effective CPUs will be set to its exclusive CPUs unless some of them are offline. This patch series also introduces a new category of cpuset partition called remote partitions. The existing partition category where the partition roots have to be clustered around the root cgroup in a hierarchical way is now referred to as local partitions. A remote partition can be formed far from the root cgroup with no partition root parent. While local partitions can be created without touching "cpuset.cpus.exclusive" as it can be set automatically if a cpuset becomes a local partition root. Properly set "cpuset.cpus.exclusive" values down the hierarchy are required to create a remote partition. Both scheduling and isolated partitions can be formed in a remote partition. A local partition can be created under a remote partition. A remote partition, however, cannot be formed under a local partition for now. Modern container orchestration tools like Kubernetes use the cgroup hierarchy to manage different containers. And it is relying on other middleware like systemd to help managing it. If a container needs to use isolated CPUs, it is hard to get those with the local partitions as it will require the administrative parent cgroup to be a partition root too which tool like systemd may not be ready to manage. With this patch series, we allow the creation of remote partition far from the root. The container management tool can manage the "cpuset.cpus.exclusive" file without impacting the other cpuset files that are managed by other middlewares. Of course, invalid "cpuset.cpus.exclusive" values will be rejected and changes to "cpuset.cpus" can affect the value of "cpuset.cpus.exclusive" due to the requirement that it has to be a subset of the former control file. Waiman Long (9): cgroup/cpuset: Inherit parent's load balance state in v2 cgroup/cpuset: Extract out CS_CPU_EXCLUSIVE & CS_SCHED_LOAD_BALANCE handling cgroup/cpuset: Improve temporary cpumasks handling cgroup/cpuset: Allow suppression of sched domain rebuild in update_cpumasks_hier() cgroup/cpuset: Add cpuset.cpus.exclusive for v2 cgroup/cpuset: Introduce remote partition cgroup/cpuset: Check partition conflict with housekeeping setup cgroup/cpuset: Documentation update for partition cgroup/cpuset: Extend test_cpuset_prs.sh to test remote partition Documentation/admin-guide/cgroup-v2.rst | 100 +- kernel/cgroup/cpuset.c | 1347 ++++++++++++----- .../selftests/cgroup/test_cpuset_prs.sh | 398 +++-- 3 files changed, 1291 insertions(+), 554 deletions(-) -- 2.31.1

9 months, 3 weeks

2
21
0 0

2024

2023

2022

2021

2020

2019

2018

2017

Linux-kselftest-mirror June 2023