- Linux-kselftest-mirror - lists.linaro.org

[PATCH v14 0/5] Implement IOCTL to get and optionally clear info about PTEs

by Muhammad Usama Anjum

*Changes in v14* - Fix build error caused by #ifdef added at last minute in some configs *Changes in v13* - Rebase on top of next-20230414 - Give-up on using uffd_wp_range() and write new helpers, flush tlb only once *Changes in v12* - Update and other memory types to UFFD_FEATURE_WP_ASYNC - Rebaase on top of next-20230406 - Review updates *Changes in v11* - Rebase on top of next-20230307 - Base patches on UFFD_FEATURE_WP_UNPOPULATED - Do a lot of cosmetic changes and review updates - Remove ENGAGE_WP + !GET operation as it can be performed with UFFDIO_WRITEPROTECT *Changes in v10* - Add specific condition to return error if hugetlb is used with wp async - Move changes in tools/include/uapi/linux/fs.h to separate patch - Add documentation *Changes in v9:* - Correct fault resolution for userfaultfd wp async - Fix build warnings and errors which were happening on some configs - Simplify pagemap ioctl's code *Changes in v8:* - Update uffd async wp implementation - Improve PAGEMAP_IOCTL implementation *Changes in v7:* - Add uffd wp async - Update the IOCTL to use uffd under the hood instead of soft-dirty flags *Motivation* The real motivation for adding PAGEMAP_SCAN IOCTL is to emulate Windows GetWriteWatch() syscall [1]. The GetWriteWatch{} retrieves the addresses of the pages that are written to in a region of virtual memory. This syscall is used in Windows applications and games etc. This syscall is being emulated in pretty slow manner in userspace. Our purpose is to enhance the kernel such that we translate it efficiently in a better way. Currently some out of tree hack patches are being used to efficiently emulate it in some kernels. We intend to replace those with these patches. So the whole gaming on Linux can effectively get benefit from this. It means there would be tons of users of this code. CRIU use case [2] was mentioned by Andrei and Danylo: > Use cases for migrating sparse VMAs are binaries sanitized with ASAN, > MSAN or TSAN [3]. All of these sanitizers produce sparse mappings of > shadow memory [4]. Being able to migrate such binaries allows to highly > reduce the amount of work needed to identify and fix post-migration > crashes, which happen constantly. Andrei's defines the following uses of this code: * it is more granular and allows us to track changed pages more effectively. The current interface can clear dirty bits for the entire process only. In addition, reading info about pages is a separate operation. It means we must freeze the process to read information about all its pages, reset dirty bits, only then we can start dumping pages. The information about pages becomes more and more outdated, while we are processing pages. The new interface solves both these downsides. First, it allows us to read pte bits and clear the soft-dirty bit atomically. It means that CRIU will not need to freeze processes to pre-dump their memory. Second, it clears soft-dirty bits for a specified region of memory. It means CRIU will have actual info about pages to the moment of dumping them. * The new interface has to be much faster because basic page filtering is happening in the kernel. With the old interface, we have to read pagemap for each page. *Implementation Evolution (Short Summary)* From the definition of GetWriteWatch(), we feel like kernel's soft-dirty feature can be used under the hood with some additions like: * reset soft-dirty flag for only a specific region of memory instead of clearing the flag for the entire process * get and clear soft-dirty flag for a specific region atomically So we decided to use ioctl on pagemap file to read or/and reset soft-dirty flag. But using soft-dirty flag, sometimes we get extra pages which weren't even written. They had become soft-dirty because of VMA merging and VM_SOFTDIRTY flag. This breaks the definition of GetWriteWatch(). We were able to by-pass this short coming by ignoring VM_SOFTDIRTY until David reported that mprotect etc messes up the soft-dirty flag while ignoring VM_SOFTDIRTY [5]. This wasn't happening until [6] got introduced. We discussed if we can revert these patches. But we could not reach to any conclusion. So at this point, I made couple of tries to solve this whole VM_SOFTDIRTY issue by correcting the soft-dirty implementation: * [7] Correct the bug fixed wrongly back in 2014. It had potential to cause regression. We left it behind. * [8] Keep a list of soft-dirty part of a VMA across splits and merges. I got the reply don't increase the size of the VMA by 8 bytes. At this point, we left soft-dirty considering it is too much delicate and userfaultfd [9] seemed like the only way forward. From there onward, we have been basing soft-dirty emulation on userfaultfd wp feature where kernel resolves the faults itself when WP_ASYNC feature is used. It was straight forward to add WP_ASYNC feature in userfautlfd. Now we get only those pages dirty or written-to which are really written in reality. (PS There is another WP_UNPOPULATED userfautfd feature is required which is needed to avoid pre-faulting memory before write-protecting [9].) All the different masks were added on the request of CRIU devs to create interface more generic and better. [1] https://learn.microsoft.com/en-us/windows/win32/api/memoryapi/nf-memoryapi-… [2] https://lore.kernel.org/all/20221014134802.1361436-1-mdanylo@google.com [3] https://github.com/google/sanitizers [4] https://github.com/google/sanitizers/wiki/AddressSanitizerAlgorithm#64-bit [5] https://lore.kernel.org/all/bfcae708-db21-04b4-0bbe-712badd03071@redhat.com [6] https://lore.kernel.org/all/20220725142048.30450-1-peterx@redhat.com/ [7] https://lore.kernel.org/all/20221122115007.2787017-1-usama.anjum@collabora.… [8] https://lore.kernel.org/all/20221220162606.1595355-1-usama.anjum@collabora.… [9] https://lore.kernel.org/all/20230306213925.617814-1-peterx@redhat.com [10] https://lore.kernel.org/all/20230125144529.1630917-1-mdanylo@google.com * Original Cover letter from v8* Hello, Note: Soft-dirty pages and pages which have been written-to are synonyms. As kernel already has soft-dirty feature inside which we have given up to use, we are using written-to terminology while using UFFD async WP under the hood. This IOCTL, PAGEMAP_SCAN on pagemap file can be used to get and/or clear the info about page table entries. The following operations are supported in this ioctl: - Get the information if the pages have been written-to (PAGE_IS_WRITTEN), file mapped (PAGE_IS_FILE), present (PAGE_IS_PRESENT) or swapped (PAGE_IS_SWAPPED). - Write-protect the pages (PAGEMAP_WP_ENGAGE) to start finding which pages have been written-to. - Find pages which have been written-to and write protect the pages (atomic PAGE_IS_WRITTEN + PAGEMAP_WP_ENGAGE) It is possible to find and clear soft-dirty pages entirely in userspace. But it isn't efficient: - The mprotect and SIGSEGV handler for bookkeeping - The userfaultfd wp (synchronous) with the handler for bookkeeping Some benchmarks can be seen here[1]. This series adds features that weren't present earlier: - There is no atomic get soft-dirty/Written-to status and clear present in the kernel. - The pages which have been written-to can not be found in accurate way. (Kernel's soft-dirty PTE bit + sof_dirty VMA bit shows more soft-dirty pages than there actually are.) Historically, soft-dirty PTE bit tracking has been used in the CRIU project. The procfs interface is enough for finding the soft-dirty bit status and clearing the soft-dirty bit of all the pages of a process. We have the use case where we need to track the soft-dirty PTE bit for only specific pages on-demand. We need this tracking and clear mechanism of a region of memory while the process is running to emulate the getWriteWatch() syscall of Windows. *(Moved to using UFFD instead of soft-dirtyi feature to find pages which have been written-to from v7 patch series)*: Stop using the soft-dirty flags for finding which pages have been written to. It is too delicate and wrong as it shows more soft-dirty pages than the actual soft-dirty pages. There is no interest in correcting it [2][3] as this is how the feature was written years ago. It shouldn't be updated to changed behaviour. Peter Xu has suggested using the async version of the UFFD WP [4] as it is based inherently on the PTEs. So in this patch series, I've added a new mode to the UFFD which is asynchronous version of the write protect. When this variant of the UFFD WP is used, the page faults are resolved automatically by the kernel. The pages which have been written-to can be found by reading pagemap file (!PM_UFFD_WP). This feature can be used successfully to find which pages have been written to from the time the pages were write protected. This works just like the soft-dirty flag without showing any extra pages which aren't soft-dirty in reality. The information related to pages if the page is file mapped, present and swapped is required for the CRIU project [5][6]. The addition of the required mask, any mask, excluded mask and return masks are also required for the CRIU project [5]. The IOCTL returns the addresses of the pages which match the specific masks. The page addresses are returned in struct page_region in a compact form. The max_pages is needed to support a use case where user only wants to get a specific number of pages. So there is no need to find all the pages of interest in the range when max_pages is specified. The IOCTL returns when the maximum number of the pages are found. The max_pages is optional. If max_pages is specified, it must be equal or greater than the vec_size. This restriction is needed to handle worse case when one page_region only contains info of one page and it cannot be compacted. This is needed to emulate the Windows getWriteWatch() syscall. The patch series include the detailed selftest which can be used as an example for the uffd async wp test and PAGEMAP_IOCTL. It shows the interface usages as well. [1] https://lore.kernel.org/lkml/54d4c322-cd6e-eefd-b161-2af2b56aae24@collabora… [2] https://lore.kernel.org/all/20221220162606.1595355-1-usama.anjum@collabora.… [3] https://lore.kernel.org/all/20221122115007.2787017-1-usama.anjum@collabora.… [4] https://lore.kernel.org/all/Y6Hc2d+7eTKs7AiH@x1n [5] https://lore.kernel.org/all/YyiDg79flhWoMDZB@gmail.com/ [6] https://lore.kernel.org/all/20221014134802.1361436-1-mdanylo@google.com/ Regards, Muhammad Usama Anjum Muhammad Usama Anjum (4): fs/proc/task_mmu: Implement IOCTL to get and optionally clear info about PTEs tools headers UAPI: Update linux/fs.h with the kernel sources mm/pagemap: add documentation of PAGEMAP_SCAN IOCTL selftests: mm: add pagemap ioctl tests Peter Xu (1): userfaultfd: UFFD_FEATURE_WP_ASYNC Documentation/admin-guide/mm/pagemap.rst | 56 + Documentation/admin-guide/mm/userfaultfd.rst | 35 + fs/proc/task_mmu.c | 478 +++++++ fs/userfaultfd.c | 26 +- include/linux/userfaultfd_k.h | 21 +- include/uapi/linux/fs.h | 53 + include/uapi/linux/userfaultfd.h | 9 +- mm/hugetlb.c | 32 +- mm/memory.c | 27 +- tools/include/uapi/linux/fs.h | 53 + tools/testing/selftests/mm/.gitignore | 1 + tools/testing/selftests/mm/Makefile | 3 +- tools/testing/selftests/mm/config | 1 + tools/testing/selftests/mm/pagemap_ioctl.c | 1326 ++++++++++++++++++ tools/testing/selftests/mm/run_vmtests.sh | 4 + 15 files changed, 2102 insertions(+), 23 deletions(-) create mode 100644 tools/testing/selftests/mm/pagemap_ioctl.c mode change 100644 => 100755 tools/testing/selftests/mm/run_vmtests.sh -- 2.39.2

2 years, 5 months

2
6
0 0

[PATCH linux-next 0/3] selftests: net: udpgso_bench_rx: Fix verifty, gsosize, and packet number exceptions

by yang.yang29＠zte.com.cn

From: zhang yunkai (CGEL ZTE) <zhang.yunkai(a)zte.com.cn> 1.Fix verifty exception Executing the following command fails: bash# udpgso_bench_tx -l 4 -4 -D "$DST" bash# udpgso_bench_tx -l 4 -4 -D "$DST" -S 0 bash# udpgso_bench_rx -4 -G -S 1472 -v udpgso_bench_rx: data[1472]: len 2944, a(97) != q(113) This is because the sending buffers are not aligned by 26 bytes, and the GRO is not merged sequentially, and the receiver does not judge this situation. We think of the receiver to split the data and then validate it, just as the application actually uses this feature. 2.Fix gsosize exception Executing the following command fails: bash# udpgso_bench_tx -l 4 -4 -D "$DST" bash# udpgso_bench_tx -l 4 -4 -D "$DST" -S 0 bash# udpgso_bench_rx -4 -G -S 1472 udp rx: 15 MB/s 256 calls/s udp rx: 30 MB/s 512 calls/s udpgso_bench_rx: recv: bad gso size, got -1, expected 1472 (-1 == no gso cmsg)) IP 192.168.2.199.55238 > 192.168.2.203.8000: UDP, length 7360 IP 192.168.2.199.55238 > 192.168.2.203.8000: UDP, length 1472 IP 192.168.2.199.55238 > 192.168.2.203.8000: UDP, length 1472 IP 192.168.2.199.55238 > 192.168.2.203.8000: UDP, length 4416 IP 192.168.2.199.55238 > 192.168.2.203.8000: UDP, length 11776 IP 192.168.2.199.55238 > 192.168.2.203.8000: UDP, length 20608 recv: got one message len:1472, probably not an error. recv: got one message len:1472, probably not an error. This is due to network, NAPI, timer, etc., only one message being received. We believe that this situation should be normal. 3.Fix packet number exception bash# udpgso_bench_rx -4 -n 100 bash# udpgso_bench_tx -l 1 -4 -D "$DST" udpgso_bench_rx: wrong packet number! got 0, expected 100 This is because the packets is cleared after print. Zhang Yunkai (3): selftests: net: udpgso_bench_rx: Fix verifty exceptions selftests: net: udpgso_bench_rx: Fix gsosize exceptions selftests: net: udpgso_bench_rx: Fix packet number exceptions tools/testing/selftests/net/udpgso_bench_rx.c | 45 +++++++++++++++++++++------ 1 file changed, 35 insertions(+), 10 deletions(-) -- 2.15.2

2 years, 5 months

4
8
0 0

[PATCH] kunit: Set the current KUnit context when cleaning up

by David Gow

KUnit tests run in a kthread, with the current->kunit_test pointer set to the test's context. This allows the kunit_get_current_test() and kunit_fail_current_test() macros to work. Normally, this pointer is still valid during test shutdown (i.e., the suite->exit function, and any resource cleanup). However, if the test has exited early (e.g., due to a failed assertion), the cleanup is done in the parent KUnit thread, which does not have an active context. Fix this by setting the active KUnit context for the duration of the test shutdown procedure. When the test exits normally, this does nothing. When run from the KUits previous value (probably NULL) afterwards. This should make it easier to get access to the KUnit context, particularly from within resource cleanup functions, which may, for example, need access to data in test->priv. Signed-off-by: David Gow <davidgow(a)google.com> --- This becomes useful with the current kunit_add_action() implementation, as actions do not get the KUnit context passed in by default: https://lore.kernel.org/linux-kselftest/CABVgOSmjs0wLUa4=ErkB9tH8p6A1P6N33b… I think it's probably correct anyway, though, so we should either do this, or totally rule out using kunit_get_current_test() here at all, by resetting current->kunit_test to NULL before running cleanup even in the normal case. I've only given this the most cursory testing so far (I'm not sure how much of the executor innards I want to expose to be able to actually write a proper test for it), so more eyes and/or suggestions are welcome. Cheers, -- David --- lib/kunit/test.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/lib/kunit/test.c b/lib/kunit/test.c index e2910b261112..2d7cad249863 100644 --- a/lib/kunit/test.c +++ b/lib/kunit/test.c @@ -392,10 +392,21 @@ static void kunit_case_internal_cleanup(struct kunit *test) static void kunit_run_case_cleanup(struct kunit *test, struct kunit_suite *suite) { + /* + * If we're no-longer running from within the test kthread() because it failed + * or timed out, we still need the context to be okay when running exit and + * cleanup functions. + */ + struct kunit *old_current = current->kunit_test; + + current->kunit_test = test; if (suite->exit) suite->exit(test); kunit_case_internal_cleanup(test); + + /* Restore the thread's previous test context (probably NULL or test). */ + current->kunit_test = old_current; } struct kunit_try_catch_context { -- 2.40.0.634.g4ca3ef3211-goog

2 years, 5 months

2
3
0 0

[PATCH 0/2] Access variable length array relaxed for integer type

by Feng zhou

From: Feng Zhou <zhoufeng.zf(a)bytedance.com> Add support for integer type of accessing variable length array. Add a selftest to check it. Feng Zhou (2): bpf: support access variable length array of integer type selftests/bpf: Add test to access integer type of variable array kernel/bpf/btf.c | 8 +++++--- .../selftests/bpf/bpf_testmod/bpf_testmod.c | 20 +++++++++++++++++++ .../selftests/bpf/prog_tests/tracing_struct.c | 2 ++ .../selftests/bpf/progs/tracing_struct.c | 13 ++++++++++++ 4 files changed, 40 insertions(+), 3 deletions(-) -- 2.20.1

2 years, 5 months

3
4
0 0

[PATCH] selftests/futex: Fix pthread_create() error handling

by tavianator＠tavianator.com

From: Tavian Barnes <tavianator(a)tavianator.com> The pthread APIs return error codes, rather than setting errno. Signed-off-by: Tavian Barnes <tavianator(a)tavianator.com> --- .../selftests/futex/functional/futex_requeue.c | 10 ++++++---- .../functional/futex_requeue_pi_mismatched_ops.c | 5 +++-- .../selftests/futex/functional/futex_wait.c | 15 +++++++++------ .../functional/futex_wait_private_mapped_file.c | 2 +- .../functional/futex_wait_uninitialized_heap.c | 2 +- .../selftests/futex/functional/futex_waitv.c | 10 ++++++---- 6 files changed, 26 insertions(+), 18 deletions(-) diff --git a/tools/testing/selftests/futex/functional/futex_requeue.c b/tools/testing/selftests/futex/functional/futex_requeue.c index 51485be6eb2f..76f672ad5263 100644 --- a/tools/testing/selftests/futex/functional/futex_requeue.c +++ b/tools/testing/selftests/futex/functional/futex_requeue.c @@ -73,8 +73,9 @@ int main(int argc, char *argv[]) /* * Requeue a waiter from f1 to f2, and wake f2. */ - if (pthread_create(&waiter[0], NULL, waiterfn, NULL)) - error("pthread_create failed\n", errno); + res = pthread_create(&waiter[0], NULL, waiterfn, NULL); + if (res) + error("pthread_create failed\n", res); usleep(WAKE_WAIT_US); @@ -105,8 +106,9 @@ int main(int argc, char *argv[]) * At futex_wake, wake INT_MAX (should be exactly 7). */ for (i = 0; i < 10; i++) { - if (pthread_create(&waiter[i], NULL, waiterfn, NULL)) - error("pthread_create failed\n", errno); + res = pthread_create(&waiter[i], NULL, waiterfn, NULL); + if (res) + error("pthread_create failed\n", res); } usleep(WAKE_WAIT_US); diff --git a/tools/testing/selftests/futex/functional/futex_requeue_pi_mismatched_ops.c b/tools/testing/selftests/futex/functional/futex_requeue_pi_mismatched_ops.c index d0a4d332ea44..e28de62bc03a 100644 --- a/tools/testing/selftests/futex/functional/futex_requeue_pi_mismatched_ops.c +++ b/tools/testing/selftests/futex/functional/futex_requeue_pi_mismatched_ops.c @@ -79,8 +79,9 @@ int main(int argc, char *argv[]) ksft_print_msg("%s: Detect mismatched requeue_pi operations\n", basename(argv[0])); - if (pthread_create(&child, NULL, blocking_child, NULL)) { - error("pthread_create\n", errno); + ret = pthread_create(&child, NULL, blocking_child, NULL); + if (ret) { + error("pthread_create\n", ret); ret = RET_ERROR; goto out; } diff --git a/tools/testing/selftests/futex/functional/futex_wait.c b/tools/testing/selftests/futex/functional/futex_wait.c index 685140d9b93d..2e06539e5bb7 100644 --- a/tools/testing/selftests/futex/functional/futex_wait.c +++ b/tools/testing/selftests/futex/functional/futex_wait.c @@ -78,8 +78,9 @@ int main(int argc, char *argv[]) /* Testing a private futex */ info("Calling private futex_wait on futex: %p\n", futex); - if (pthread_create(&waiter, NULL, waiterfn, (void *) &flags)) - error("pthread_create failed\n", errno); + res = pthread_create(&waiter, NULL, waiterfn, (void *) &flags); + if (res) + error("pthread_create failed\n", res); usleep(WAKE_WAIT_US); @@ -106,8 +107,9 @@ int main(int argc, char *argv[]) futex = shared_data; info("Calling shared (page anon) futex_wait on futex: %p\n", futex); - if (pthread_create(&waiter, NULL, waiterfn, NULL)) - error("pthread_create failed\n", errno); + res = pthread_create(&waiter, NULL, waiterfn, NULL); + if (res) + error("pthread_create failed\n", res); usleep(WAKE_WAIT_US); @@ -145,8 +147,9 @@ int main(int argc, char *argv[]) futex = shm; info("Calling shared (file backed) futex_wait on futex: %p\n", futex); - if (pthread_create(&waiter, NULL, waiterfn, NULL)) - error("pthread_create failed\n", errno); + res = pthread_create(&waiter, NULL, waiterfn, NULL); + if (res) + error("pthread_create failed\n", res); usleep(WAKE_WAIT_US); diff --git a/tools/testing/selftests/futex/functional/futex_wait_private_mapped_file.c b/tools/testing/selftests/futex/functional/futex_wait_private_mapped_file.c index fb4148f23fa3..7b4a28e8dc1a 100644 --- a/tools/testing/selftests/futex/functional/futex_wait_private_mapped_file.c +++ b/tools/testing/selftests/futex/functional/futex_wait_private_mapped_file.c @@ -100,7 +100,7 @@ int main(int argc, char **argv) basename(argv[0])); ret = pthread_create(&thr, NULL, thr_futex_wait, NULL); - if (ret < 0) { + if (ret) { fprintf(stderr, "pthread_create error\n"); ret = RET_ERROR; goto out; diff --git a/tools/testing/selftests/futex/functional/futex_wait_uninitialized_heap.c b/tools/testing/selftests/futex/functional/futex_wait_uninitialized_heap.c index ed9cd07e31c1..bf7529bd7f0d 100644 --- a/tools/testing/selftests/futex/functional/futex_wait_uninitialized_heap.c +++ b/tools/testing/selftests/futex/functional/futex_wait_uninitialized_heap.c @@ -103,7 +103,7 @@ int main(int argc, char **argv) ret = pthread_create(&thr, NULL, wait_thread, NULL); if (ret) { - error("pthread_create\n", errno); + error("pthread_create\n", ret); ret = RET_ERROR; goto out; } diff --git a/tools/testing/selftests/futex/functional/futex_waitv.c b/tools/testing/selftests/futex/functional/futex_waitv.c index a94337f677e1..d2400b3c7d0a 100644 --- a/tools/testing/selftests/futex/functional/futex_waitv.c +++ b/tools/testing/selftests/futex/functional/futex_waitv.c @@ -94,8 +94,9 @@ int main(int argc, char *argv[]) } /* Private waitv */ - if (pthread_create(&waiter, NULL, waiterfn, NULL)) - error("pthread_create failed\n", errno); + res = pthread_create(&waiter, NULL, waiterfn, NULL); + if (res) + error("pthread_create failed\n", res); usleep(WAKE_WAIT_US); @@ -127,8 +128,9 @@ int main(int argc, char *argv[]) waitv[i].__reserved = 0; } - if (pthread_create(&waiter, NULL, waiterfn, NULL)) - error("pthread_create failed\n", errno); + res = pthread_create(&waiter, NULL, waiterfn, NULL); + if (res) + error("pthread_create failed\n", res); usleep(WAKE_WAIT_US); -- 2.40.0

2 years, 5 months

2
1
0 0

[PATCH v13 0/5] Implement IOCTL to get and optionally clear info about PTEs

by Muhammad Usama Anjum

*Changes in v13* - Rebase on top of next-20230414 - Give-up on using uffd_wp_range() and write new helpers, flush tlb only once *Changes in v12* - Update and other memory types to UFFD_FEATURE_WP_ASYNC - Rebaase on top of next-20230406 - Review updates *Changes in v11* - Rebase on top of next-20230307 - Base patches on UFFD_FEATURE_WP_UNPOPULATED - Do a lot of cosmetic changes and review updates - Remove ENGAGE_WP + !GET operation as it can be performed with UFFDIO_WRITEPROTECT *Changes in v10* - Add specific condition to return error if hugetlb is used with wp async - Move changes in tools/include/uapi/linux/fs.h to separate patch - Add documentation *Changes in v9:* - Correct fault resolution for userfaultfd wp async - Fix build warnings and errors which were happening on some configs - Simplify pagemap ioctl's code *Changes in v8:* - Update uffd async wp implementation - Improve PAGEMAP_IOCTL implementation *Changes in v7:* - Add uffd wp async - Update the IOCTL to use uffd under the hood instead of soft-dirty flags *Motivation* The real motivation for adding PAGEMAP_SCAN IOCTL is to emulate Windows GetWriteWatch() syscall [1]. The GetWriteWatch{} retrieves the addresses of the pages that are written to in a region of virtual memory. This syscall is used in Windows applications and games etc. This syscall is being emulated in pretty slow manner in userspace. Our purpose is to enhance the kernel such that we translate it efficiently in a better way. Currently some out of tree hack patches are being used to efficiently emulate it in some kernels. We intend to replace those with these patches. So the whole gaming on Linux can effectively get benefit from this. It means there would be tons of users of this code. CRIU use case [2] was mentioned by Andrei and Danylo: > Use cases for migrating sparse VMAs are binaries sanitized with ASAN, > MSAN or TSAN [3]. All of these sanitizers produce sparse mappings of > shadow memory [4]. Being able to migrate such binaries allows to highly > reduce the amount of work needed to identify and fix post-migration > crashes, which happen constantly. Andrei's defines the following uses of this code: * it is more granular and allows us to track changed pages more effectively. The current interface can clear dirty bits for the entire process only. In addition, reading info about pages is a separate operation. It means we must freeze the process to read information about all its pages, reset dirty bits, only then we can start dumping pages. The information about pages becomes more and more outdated, while we are processing pages. The new interface solves both these downsides. First, it allows us to read pte bits and clear the soft-dirty bit atomically. It means that CRIU will not need to freeze processes to pre-dump their memory. Second, it clears soft-dirty bits for a specified region of memory. It means CRIU will have actual info about pages to the moment of dumping them. * The new interface has to be much faster because basic page filtering is happening in the kernel. With the old interface, we have to read pagemap for each page. *Implementation Evolution (Short Summary)* From the definition of GetWriteWatch(), we feel like kernel's soft-dirty feature can be used under the hood with some additions like: * reset soft-dirty flag for only a specific region of memory instead of clearing the flag for the entire process * get and clear soft-dirty flag for a specific region atomically So we decided to use ioctl on pagemap file to read or/and reset soft-dirty flag. But using soft-dirty flag, sometimes we get extra pages which weren't even written. They had become soft-dirty because of VMA merging and VM_SOFTDIRTY flag. This breaks the definition of GetWriteWatch(). We were able to by-pass this short coming by ignoring VM_SOFTDIRTY until David reported that mprotect etc messes up the soft-dirty flag while ignoring VM_SOFTDIRTY [5]. This wasn't happening until [6] got introduced. We discussed if we can revert these patches. But we could not reach to any conclusion. So at this point, I made couple of tries to solve this whole VM_SOFTDIRTY issue by correcting the soft-dirty implementation: * [7] Correct the bug fixed wrongly back in 2014. It had potential to cause regression. We left it behind. * [8] Keep a list of soft-dirty part of a VMA across splits and merges. I got the reply don't increase the size of the VMA by 8 bytes. At this point, we left soft-dirty considering it is too much delicate and userfaultfd [9] seemed like the only way forward. From there onward, we have been basing soft-dirty emulation on userfaultfd wp feature where kernel resolves the faults itself when WP_ASYNC feature is used. It was straight forward to add WP_ASYNC feature in userfautlfd. Now we get only those pages dirty or written-to which are really written in reality. (PS There is another WP_UNPOPULATED userfautfd feature is required which is needed to avoid pre-faulting memory before write-protecting [9].) All the different masks were added on the request of CRIU devs to create interface more generic and better. [1] https://learn.microsoft.com/en-us/windows/win32/api/memoryapi/nf-memoryapi-… [2] https://lore.kernel.org/all/20221014134802.1361436-1-mdanylo@google.com [3] https://github.com/google/sanitizers [4] https://github.com/google/sanitizers/wiki/AddressSanitizerAlgorithm#64-bit [5] https://lore.kernel.org/all/bfcae708-db21-04b4-0bbe-712badd03071@redhat.com [6] https://lore.kernel.org/all/20220725142048.30450-1-peterx@redhat.com/ [7] https://lore.kernel.org/all/20221122115007.2787017-1-usama.anjum@collabora.… [8] https://lore.kernel.org/all/20221220162606.1595355-1-usama.anjum@collabora.… [9] https://lore.kernel.org/all/20230306213925.617814-1-peterx@redhat.com [10] https://lore.kernel.org/all/20230125144529.1630917-1-mdanylo@google.com * Original Cover letter from v8* Hello, Note: Soft-dirty pages and pages which have been written-to are synonyms. As kernel already has soft-dirty feature inside which we have given up to use, we are using written-to terminology while using UFFD async WP under the hood. This IOCTL, PAGEMAP_SCAN on pagemap file can be used to get and/or clear the info about page table entries. The following operations are supported in this ioctl: - Get the information if the pages have been written-to (PAGE_IS_WRITTEN), file mapped (PAGE_IS_FILE), present (PAGE_IS_PRESENT) or swapped (PAGE_IS_SWAPPED). - Write-protect the pages (PAGEMAP_WP_ENGAGE) to start finding which pages have been written-to. - Find pages which have been written-to and write protect the pages (atomic PAGE_IS_WRITTEN + PAGEMAP_WP_ENGAGE) It is possible to find and clear soft-dirty pages entirely in userspace. But it isn't efficient: - The mprotect and SIGSEGV handler for bookkeeping - The userfaultfd wp (synchronous) with the handler for bookkeeping Some benchmarks can be seen here[1]. This series adds features that weren't present earlier: - There is no atomic get soft-dirty/Written-to status and clear present in the kernel. - The pages which have been written-to can not be found in accurate way. (Kernel's soft-dirty PTE bit + sof_dirty VMA bit shows more soft-dirty pages than there actually are.) Historically, soft-dirty PTE bit tracking has been used in the CRIU project. The procfs interface is enough for finding the soft-dirty bit status and clearing the soft-dirty bit of all the pages of a process. We have the use case where we need to track the soft-dirty PTE bit for only specific pages on-demand. We need this tracking and clear mechanism of a region of memory while the process is running to emulate the getWriteWatch() syscall of Windows. *(Moved to using UFFD instead of soft-dirtyi feature to find pages which have been written-to from v7 patch series)*: Stop using the soft-dirty flags for finding which pages have been written to. It is too delicate and wrong as it shows more soft-dirty pages than the actual soft-dirty pages. There is no interest in correcting it [2][3] as this is how the feature was written years ago. It shouldn't be updated to changed behaviour. Peter Xu has suggested using the async version of the UFFD WP [4] as it is based inherently on the PTEs. So in this patch series, I've added a new mode to the UFFD which is asynchronous version of the write protect. When this variant of the UFFD WP is used, the page faults are resolved automatically by the kernel. The pages which have been written-to can be found by reading pagemap file (!PM_UFFD_WP). This feature can be used successfully to find which pages have been written to from the time the pages were write protected. This works just like the soft-dirty flag without showing any extra pages which aren't soft-dirty in reality. The information related to pages if the page is file mapped, present and swapped is required for the CRIU project [5][6]. The addition of the required mask, any mask, excluded mask and return masks are also required for the CRIU project [5]. The IOCTL returns the addresses of the pages which match the specific masks. The page addresses are returned in struct page_region in a compact form. The max_pages is needed to support a use case where user only wants to get a specific number of pages. So there is no need to find all the pages of interest in the range when max_pages is specified. The IOCTL returns when the maximum number of the pages are found. The max_pages is optional. If max_pages is specified, it must be equal or greater than the vec_size. This restriction is needed to handle worse case when one page_region only contains info of one page and it cannot be compacted. This is needed to emulate the Windows getWriteWatch() syscall. The patch series include the detailed selftest which can be used as an example for the uffd async wp test and PAGEMAP_IOCTL. It shows the interface usages as well. [1] https://lore.kernel.org/lkml/54d4c322-cd6e-eefd-b161-2af2b56aae24@collabora… [2] https://lore.kernel.org/all/20221220162606.1595355-1-usama.anjum@collabora.… [3] https://lore.kernel.org/all/20221122115007.2787017-1-usama.anjum@collabora.… [4] https://lore.kernel.org/all/Y6Hc2d+7eTKs7AiH@x1n [5] https://lore.kernel.org/all/YyiDg79flhWoMDZB@gmail.com/ [6] https://lore.kernel.org/all/20221014134802.1361436-1-mdanylo@google.com/ Regards, Muhammad Usama Anjum Muhammad Usama Anjum (4): fs/proc/task_mmu: Implement IOCTL to get and optionally clear info about PTEs tools headers UAPI: Update linux/fs.h with the kernel sources mm/pagemap: add documentation of PAGEMAP_SCAN IOCTL selftests: mm: add pagemap ioctl tests Peter Xu (1): userfaultfd: UFFD_FEATURE_WP_ASYNC Documentation/admin-guide/mm/pagemap.rst | 56 + Documentation/admin-guide/mm/userfaultfd.rst | 35 + fs/proc/task_mmu.c | 478 +++++++ fs/userfaultfd.c | 26 +- include/linux/userfaultfd_k.h | 21 +- include/uapi/linux/fs.h | 53 + include/uapi/linux/userfaultfd.h | 9 +- mm/hugetlb.c | 32 +- mm/memory.c | 27 +- tools/include/uapi/linux/fs.h | 53 + tools/testing/selftests/mm/.gitignore | 1 + tools/testing/selftests/mm/Makefile | 3 +- tools/testing/selftests/mm/config | 1 + tools/testing/selftests/mm/pagemap_ioctl.c | 1326 ++++++++++++++++++ tools/testing/selftests/mm/run_vmtests.sh | 4 + 15 files changed, 2102 insertions(+), 23 deletions(-) create mode 100644 tools/testing/selftests/mm/pagemap_ioctl.c mode change 100644 => 100755 tools/testing/selftests/mm/run_vmtests.sh -- 2.39.2

2 years, 5 months

3
8
0 0

[PATCH 0/6] tools/nolibc: add stackprotector support for more architectures

by Thomas Weißschuh

Add stackprotector support for all remaining architectures, except s390. On s390 the stackprotectors are not supported in "global" mode; only "sysreg" mode which is not suppored in nolibc. The series also contains a small optimization to strace output during execution of nolibc-test. This series is based on the 20230415-nolibc-updates-4a branch of the nolibc tree. Signed-off-by: Thomas Weißschuh <linux(a)weissschuh.net> --- Thomas Weißschuh (6): selftests/nolibc: reduce syscalls during space padding tools/nolibc: riscv: add stackprotector support tools/nolibc: aarch64: add stackprotector support tools/nolibc: arm: add stackprotector support tools/nolibc: loongarch: add stackprotector support tools/nolibc: mips: add stackprotector support tools/include/nolibc/arch-aarch64.h | 7 ++++++- tools/include/nolibc/arch-arm.h | 7 ++++++- tools/include/nolibc/arch-loongarch.h | 7 ++++++- tools/include/nolibc/arch-mips.h | 8 +++++++- tools/include/nolibc/arch-riscv.h | 7 ++++++- tools/testing/selftests/nolibc/Makefile | 5 +++++ tools/testing/selftests/nolibc/nolibc-test.c | 15 +++++++++++---- 7 files changed, 47 insertions(+), 9 deletions(-) --- base-commit: e35214ea9db7477a02e67a8b412e8046534bb97c change-id: 20230408-nolibc-stackprotector-archs-42244674616e Best regards, -- Thomas Weißschuh <linux(a)weissschuh.net>

2 years, 5 months

2
7
0 0

[PATCH] kselftest/arm64: Add a smoke test for ptracing hardware break/watch points

by Mark Brown

There was a report that the hardware breakpoints and watch points weren't reporting the debug architecture version as expected, they were reporting a version of 0 which is not defined in the architecture. This happens when running in a KVM guest if the host has a debug architecture version not supported by KVM, it in turn confuses GDB which rejects any debug architecture version it does not know about. Add a test that covers that situation and while we're at it reports the debug architecture version and number of slots available to aid with figuring out problems that may arise. Signed-off-by: Mark Brown <broonie(a)kernel.org> --- tools/testing/selftests/arm64/abi/ptrace.c | 32 +++++++++++++++++++++++++++++- 1 file changed, 31 insertions(+), 1 deletion(-) diff --git a/tools/testing/selftests/arm64/abi/ptrace.c b/tools/testing/selftests/arm64/abi/ptrace.c index be952511af22..abe4d58d731d 100644 --- a/tools/testing/selftests/arm64/abi/ptrace.c +++ b/tools/testing/selftests/arm64/abi/ptrace.c @@ -20,7 +20,7 @@ #include "../../kselftest.h" -#define EXPECTED_TESTS 7 +#define EXPECTED_TESTS 11 #define MAX_TPIDRS 2 @@ -132,6 +132,34 @@ static void test_tpidr(pid_t child) } } +static void test_hw_debug(pid_t child, int type, const char *type_name) +{ + struct user_hwdebug_state state; + struct iovec iov; + int slots, arch, ret; + + iov.iov_len = sizeof(state); + iov.iov_base = &state; + + /* Should be able to read the values */ + ret = ptrace(PTRACE_GETREGSET, child, type, &iov); + ksft_test_result(ret == 0, "read_%s\n", type_name); + + if (ret == 0) { + /* Low 8 bits is the number of slots, next 4 bits the arch */ + slots = state.dbg_info & 0xff; + arch = (state.dbg_info >> 8) & 0xf; + + ksft_print_msg("%s version %d with %d slots\n", type_name, + arch, slots); + + /* Zero is not currently architecturally valid */ + ksft_test_result(arch, "%s_arch_set\n", type_name); + } else { + ksft_test_result_skip("%s_arch_set\n"); + } +} + static int do_child(void) { if (ptrace(PTRACE_TRACEME, -1, NULL, NULL)) @@ -207,6 +235,8 @@ static int do_parent(pid_t child) ksft_print_msg("Parent is %d, child is %d\n", getpid(), child); test_tpidr(child); + test_hw_debug(child, NT_ARM_HW_WATCH, "NT_ARM_HW_WATCH"); + test_hw_debug(child, NT_ARM_HW_BREAK, "NT_ARM_HW_BREAK"); ret = EXIT_SUCCESS; --- base-commit: e8d018dd0257f744ca50a729e3d042cf2ec9da65 change-id: 20230414-arm64-test-hw-breakpoint-83fe02f607fc Best regards, -- Mark Brown <broonie(a)kernel.org>

2 years, 5 months

1
0
0 0

Re: [PATCH v8 3/3] selftests/mm: add new selftests for KSM

by David Hildenbrand

> +/* Verify that KSM can be enabled / queried with prctl. */ > +static void test_prctl(void) > +{ > + int ret; > + > + ksft_print_msg("[RUN] %s\n", __func__); > + > + ret = prctl(PR_SET_MEMORY_MERGE, 1, 0, 0, 0); > + if (ret < 0 && errno == EINVAL) { > + ksft_test_result_skip("PR_SET_MEMORY_MERGE not supported\n"); > + return; > + } else if (ret) { > + ksft_test_result_fail("PR_SET_MEMORY_MERGE=1 failed\n"); > + } Just realized we're missing a "return;" in case of the failure here. -- Thanks, David / dhildenb

2 years, 5 months

2
3
0 0

Re: [PATCH v8 1/3] mm: add new api to enable ksm per process

by David Hildenbrand

> +static void __ksm_add_vma(struct vm_area_struct *vma) > +{ > + unsigned long vm_flags = vma->vm_flags; > + > + if (vm_flags & VM_MERGEABLE) > + return; > + > + if (vma_ksm_compatible(vma)) { > + vm_flags |= VM_MERGEABLE; > + vm_flags_reset(vma, vm_flags); > + } > +} We can do the following simplification on top: diff --git a/mm/ksm.c b/mm/ksm.c index 905c8edce5cf..26e7f585d65d 100644 --- a/mm/ksm.c +++ b/mm/ksm.c @@ -2511,15 +2511,11 @@ static int ksm_scan_thread(void *nothing) static void __ksm_add_vma(struct vm_area_struct *vma) { - unsigned long vm_flags = vma->vm_flags; - - if (vm_flags & VM_MERGEABLE) + if (vma->vm_flags & VM_MERGEABLE) return; - if (vma_ksm_compatible(vma)) { - vm_flags |= VM_MERGEABLE; - vm_flags_reset(vma, vm_flags); - } + if (vma_ksm_compatible(vma)) + vm_flags_set(vma, VM_MERGEABLE); } /** -- 2.39.2 I have some patches based on your patch set (handling/testing unmerging whens setting PR_SET_MEMORY_MERGE=0. Will send out once you changes are queued. -- Thanks, David / dhildenb

2 years, 5 months

2
1
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-kselftest-mirror