- Oct 14, 2020
-
-
John Hubbard authored
This patch reduces the running time for compaction_test from about 27 sec, to 3.3 sec, which is about an 8x speedup. These numbers are for an Intel x86_64 system with 32 GB of DRAM. The compaction_test.c program was spending most of its time doing mmap(), 1 MB at a time, on about 25 GB of memory. Instead, do the mmaps 100 MB at a time. (Going past 100 MB doesn't make things go much faster, because other parts of the program are using the remaining time.) Signed-off-by:
John Hubbard <jhubbard@nvidia.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Acked-by:
Sri Jayaramappa <sjayaram@akamai.com> Cc: Shuah Khan <shuah@kernel.org> Cc: Mel Gorman <mgorman@techsingularity.net> Link: https://lkml.kernel.org/r/20201002080621.551044-2-jhubbard@nvidia.com Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Ralph Campbell authored
Some tests might not be able to be run if resources like huge pages are not available. Mark these tests as skipped instead of simply passing. Signed-off-by:
Ralph Campbell <rcampbell@nvidia.com> Signed-off-by:
Jason Gunthorpe <jgg@nvidia.com> Cc: Jerome Glisse <jglisse@redhat.com> Cc: John Hubbard <jhubbard@nvidia.com> Cc: Shuah Khan <shuah@kernel.org> Link: http://lkml.kernel.org/r/20200827190400.12608-1-rcampbell@nvidia.com Signed-off-by:
-
John Hubbard authored
Avoid accidental wrong builds, due to built-in rules working just a little bit too well--but not quite as well as required for our situation here. In other words, "make userfaultfd" (for example) is supposed to fail to build at all, because this Makefile only supports either "make" (all), or "make /full/path". However, the built-in rules, if not suppressed, will pick up CFLAGS and the initial LDLIBS (but not the target-specific LDLIBS, because those are only set for the full path target!). This causes it to get pretty far into building things despite using incorrect values such as an *occasionally* incomplete LDLIBS value. Signed-off-by:
-
John Hubbard authored
Patch series "selftests/vm: fix some minor aggravating factors in the Makefile". This fixes a couple of minor aggravating factors that I ran across while trying to do some changes in selftests/vm. These are simple things, but like most things with GNU Make, it's rarely obvious what's wrong until you understand *the entire Makefile and all of its includes*. So while there is, of course, joy in learning those details, I thought I'd fix these little things, so as to allow others to skip out on the Joy if they so choose. :) First of all, if you have an item (let's choose userfaultfd for an example) that fails to build, you might do this: $ make -j32 # ...you observe a failed item in the threaded output # OK, let's get a closer look $ make # ...but now the build quietly "succeeds". That's what Patch 0001 fixes. Second, if you instead attempt this approach for your closer look (a casual mistake, as it's not supported): $ make userfaultfd # ...userfaultfd fails to link, due to incomplete LDLIBS That's what Patch 0002 fixes. This patch (of 2): If one or more of these selftest fail to build, then after the first failure, subsequent invocations of "make" will make it appear that there are no build failures, after all. That's because the failed build products remain, with up-to-date timestamps, thus tricking Make (and you!) into believing that there's nothing else to build. Fix this by telling Make to delete targets that didn't completely succeed. Signed-off-by: -
Barry Song authored
According to Documentation/core-api/pin_user_pages.rst, FOLL_PIN is a prerequisite to FOLL_LONGTERM. Another way of saying that is, FOLL_LONGTERM is a specific case, more restrictive case of FOLL_PIN. Almost all kernel modules are using pin_user_pages() with FOLL_LONGTERM, mm/gup_benchmark.c seems to the only exception in which FOLL_PIN is not a prerequisite to FOLL_LONGTERM. Signed-off-by:
Barry Song <song.bao.hua@hisilicon.com> Signed-off-by:
-
Dan Williams authored
Break the requirement that device-dax instances are physically contiguous. With this constraint removed it allows fragmented available capacity to be fully allocated. This capability is useful to mitigate the "noisy neighbor" problem with memory-side-cache management for virtual machines, or any other scenario where a platform address boundary also designates a performance boundary. For example a direct mapped memory side cache might rotate cache colors at 1GB boundaries. With dis-contiguous allocations a device-dax instance could be configured to contain only 1 cache color. It also satisfies Joao's use case (see link) for partitioning memory for exclusive guest access. It allows for a future potential mode where the host kernel need not allocate 'struct page' capacity up-front. Reported-by:
Joao Martins <joao.m.martins@oracle.com> Signed-off-by:
Dan Williams <dan.j.williams@intel.com> Signed-off-by:
-
Dan Williams authored
The 'struct resource' in 'struct dev_pagemap' is only used for holding resource span information. The other fields, 'name', 'flags', 'desc', 'parent', 'sibling', and 'child' are all unused wasted space. This is in preparation for introducing a multi-range extension of devm_memremap_pages(). The bulk of this change is unwinding all the places internal to libnvdimm that used 'struct resource' unnecessarily, and replacing instances of 'struct dev_pagemap'.res with 'struct dev_pagemap'.range. P2PDMA had a minor usage of the resource flags field, but only to report failures with "%pR". That is replaced with an open coded print of the range. [dan.carpenter@oracle.com: mm/hmm/test: use after free in dmirror_allocate_chunk()] Link: https://lkml.kernel.org/r/20200926121402.GA7467@kadam Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
-
Dan Williams authored
The passed in dev_pagemap is only required in the pmem case as the libnvdimm core may have reserved a vmem_altmap for dev_memremap_pages() to place the memmap in pmem directly. In the hmem case there is no agent reserving an altmap so it can all be handled by a core internal default. Pass the resource range via a new @range property of 'struct dev_dax_data'. Signed-off-by:
-
- Oct 09, 2020
-
-
Daniel Latypov authored
CalledProcessError stores the output of the failed process as `bytes`, not a `str`. So when we log it on build error, the make output is all crammed into one line with "\n" instead of actually printing new lines. After this change, we get readable output with new lines, e.g. > CC lib/kunit/kunit-example-test.o > In file included from ../lib/kunit/test.c:9: > ../include/kunit/test.h:22:1: error: unknown type name ‘invalid_type_that_causes_compile’ > 22 | invalid_type_that_causes_compile errors; > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > make[3]: *** [../scripts/Makefile.build:283: lib/kunit/test.o] Error 1 Secondly, trying to concat exceptions to strings will fail with > TypeError: can only concatenate str (not "OSError") to str so fix this with an explicit cast to str. Signed-off-by:
Daniel Latypov <dlatypov@google.com> Reviewed-by:
Brendan Higgins <brendanhiggins@google.com> Tested-by:
Shuah Khan <skhan@linuxfoundation.org>
-
Christian Brauner authored
Naresh reported that selftests: pidfd: pidfd_wait hangs on linux next kernel on x86_64, i386 and arm64 Juno-r2 These devices are using NFS mounted rootfs. I have tested pidfd testcases independently and all test PASS. The Hang or exit from test run noticed when run by run_kselftest.sh pidfd_wait.c:208:wait_nonblock:Expected sys_waitid(P_PIDFD, pidfd, &info, WSTOPPED, NULL) (-1) == 0 (0) wait_nonblock: Test terminated by assertion metadata: git branch: master git repo: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git git commit: e64997027d5f171148687e58b78c8b3c869a6158 git describe: next-20200922 make_kernelversion: 5.9.0-rc6 kernel-config: http://snapshots.linaro.org/openembedded/lkft/lkft/sumo/intel-core2-32/lkft/linux-next/865/config The reason for this is a simple race in the selftests, that I overlooked and which is more likely to hit when there's a lot of processes running on the system. Basically the child process hasn't SIGSTOPed itself yet but the parent is already calling waitid() on a O_NONBLOCK pidfd. Since it doesn't find a WSTOPPED process it returns -EAGAIN correctly. The fix for this is to move the line where we're removing the O_NONBLOCK property from the fd before the waitid() WSTOPPED call so we hang until the child becomes stopped. Fixes: cd89597b ("tests: add waitid() tests for non-blocking pidfds") Reported-by:
Naresh Kamboju <naresh.kamboju@linaro.org> Tested-by:
Christian Brauner <christian.brauner@ubuntu.com>
-
- Oct 08, 2020
-
-
Kees Cook authored
As the UAPI headers start to appear in distros, we need to avoid outdated versions of struct clone_args to be able to test modern features, named "struct __clone_args". Additionally update the struct size macro names to match UAPI names. Acked-by:
Kees Cook <keescook@chromium.org>
-
Kees Cook authored
Some archs (like powerpc) only support changing the return code during syscall exit when ptrace is used. Test entry vs exit phases for which portions of the syscall number and return values need to be set at which different phases. For non-powerpc, all changes are made during ptrace syscall entry, as before. For powerpc, the syscall number is changed at ptrace syscall entry and the syscall return value is changed on ptrace syscall exit. Reported-by:
Thadeu Lima de Souza Cascardo <cascardo@canonical.com> Suggested-by:
-
Kees Cook authored
In preparation for setting syscall nr and ret values separately, refactor the helpers to take a pointer to a value, so that a NULL can indicate "do not change this respective value". This is done to keep the regset read/write happening once and in one code path. Acked-by:
-
Kees Cook authored
In preparation for performing actions during ptrace syscall exit, save the syscall number during ptrace syscall entry. Some architectures do no have the syscall number available during ptrace syscall exit. Suggested-by:
-
- Oct 07, 2020
-
-
Kees Cook authored
Currently with run_kselftest.sh there is no way to choose which test we could run. All the tests listed in kselftest-list.txt are all run every time. This patch enhanced the run_kselftest.sh to make the test collections (or tests) individually selectable. e.g.: $ ./run_kselftest.sh -c seccomp -t timers:posix_timers -t timers:nanosleep Additionally adds a way to list all known tests with "-l", usage with "-h", and perform a dry run without running tests with "-n". Co-developed-by:
Hangbin Liu <liuhangbin@gmail.com> Signed-off-by:
-
Kees Cook authored
Instead of building a script on the fly (which just repeats the same thing for each test collection), move the script out of the Makefile and into run_kselftest.sh, which reads kselftest-list.txt. Adjust the emit_tests target to report each test on a separate line so that test running tools (e.g. LAVA) can easily remove individual tests (for example, as seen in [1]). [1] https://github.com/Linaro/test-definitions/pull/208/commits/2e7b62155e4998e54ac0587704932484d4ff84c8 Signed-off-by:
-
- Oct 06, 2020
-
-
Dan Williams authored
In reaction to a proposal to introduce a memcpy_mcsafe_fast() implementation Linus points out that memcpy_mcsafe() is poorly named relative to communicating the scope of the interface. Specifically what addresses are valid to pass as source, destination, and what faults / exceptions are handled. Of particular concern is that even though x86 might be able to handle the semantics of copy_mc_to_user() with its common copy_user_generic() implementation other archs likely need / want an explicit path for this case: On Fri, May 1, 2020 at 11:28 AM Linus Torvalds <torvalds@linux-foundation.org> wrote: > > On Thu, Apr 30, 2020 at 6:21 PM Dan Williams <dan.j.williams@intel.com> wrote: > > > > However now I see that copy_user_generic() works for the wrong reason. > > It works because the exception on the source address due to poison > > looks no different than a write fault on the user address to the > > caller, it's still just a short copy. So it makes copy_to_user() work > > for the wrong reason relative to the name. > > Right. > > And it won't work that way on other architectures. On x86, we have a > generic function that can take faults on either side, and we use it > for both cases (and for the "in_user" case too), but that's an > artifact of the architecture oddity. > > In fact, it's probably wrong even on x86 - because it can hide bugs - > but writing those things is painful enough that everybody prefers > having just one function. Replace a single top-level memcpy_mcsafe() with either copy_mc_to_user(), or copy_mc_to_kernel(). Introduce an x86 copy_mc_fragile() name as the rename for the low-level x86 implementation formerly named memcpy_mcsafe(). It is used as the slow / careful backend that is supplanted by a fast copy_mc_generic() in a follow-on patch. One side-effect of this reorganization is that separating copy_mc_64.S to its own file means that perf no longer needs to track dependencies for its memcpy_64.S benchmarks. [ bp: Massage a bit. ] Signed-off-by:
Borislav Petkov <bp@suse.de> Reviewed-by:
Tony Luck <tony.luck@intel.com> Acked-by:
Michael Ellerman <mpe@ellerman.id.au> Cc: <stable@vger.kernel.org> Link: http://lore.kernel.org/r/CAHk-=wjSqtXAqfUJxFtWNwmguFASTgB0dz1dT3V-78Quiezqbg@mail.gmail.com Link: https://lkml.kernel.org/r/160195561680.2163339.11574962055305783722.stgit@dwillia2-desk3.amr.corp.intel.com
-
- Oct 05, 2020
-
-
Amit Daniel Kachhap authored
Add a testcase to check that user address with valid/invalid mte tag works in kernel mode. This test verifies that the kernel API's __arch_copy_from_user/__arch_copy_to_user works by considering if the user pointer has valid/invalid allocation tags. In MTE sync mode, file memory read/write and other similar interfaces fails if a user memory with invalid tag is accessed in kernel. In async mode no such failure occurs. Signed-off-by:
Amit Daniel Kachhap <amit.kachhap@arm.com> Tested-by:
Catalin Marinas <catalin.marinas@arm.com> Acked-by:
Will Deacon <will@kernel.org>
-
Amit Daniel Kachhap authored
Add a testcase to check that KSM should not merge pages containing same data with same/different MTE tag values. This testcase has one positive tests and passes if page merging happens according to the above rule. It also saves and restores any modified ksm sysfs entries. Signed-off-by:
-
Amit Daniel Kachhap authored
This testcase checks the different unsupported/supported options for mmap if used with PROT_MTE memory protection flag. These checks are, * Either pstate.tco enable or prctl PR_MTE_TCF_NONE option should not cause any tag mismatch faults. * Different combinations of anonymous/file memory mmap, mprotect, sync/async error mode and private/shared mappings should work. * mprotect should not be able to clear the PROT_MTE page property. Co-developed-by:
Gabor Kertesz <gabor.kertesz@arm.com> Signed-off-by:
-
Amit Daniel Kachhap authored
This test covers the mte memory behaviour of the forked process with different mapping properties and flags. It checks that all bytes of forked child memory are accessible with the same tag as that of the parent and memory accessed outside the tag range causes fault to occur. Co-developed-by:
-
Amit Daniel Kachhap authored
This testcase verifies that the tag generated with "irg" instruction contains only included tags. This is done via prtcl call. This test covers 4 scenarios, * At least one included tag. * More than one included tags. * All included. * None included. Co-developed-by:
-
Amit Daniel Kachhap authored
This test checks that the memory tag is present after mte allocation and the memory is accessible with those tags. This testcase verifies all sync, async and none mte error reporting mode. The allocated mte buffers are verified for Allocated range (no error expected while accessing buffer), Underflow range, and Overflow range. Different test scenarios covered here are, * Verify that mte memory are accessible at byte/block level. * Force underflow and overflow to occur and check the data consistency. * Check to/from between tagged and untagged memory. * Check that initial allocated memory to have 0 tag. This change also creates the necessary infrastructure to add mte test cases. MTE kselftests can use the several utility functions provided here to add wide variety of mte test scenarios. GCC compiler need flag '-march=armv8.5-a+memtag' so those flags are verified before compilation. The mte testcases can be launched with kselftest framework as, make TARGETS=arm64 ARM64_SUBTARGETS=mte kselftest or compiled as, make -C tools/testing/selftests TARGETS=arm64 ARM64_SUBTARGETS=mte CC='compiler' Co-developed-by:
-
Scott Branden authored
Add additional hooks to test_firmware to pass in support for partial file read using request_firmware_into_buf(): buf_size: size of buffer to request firmware into partial: indicates that a partial file request is being made file_offset: to indicate offset into file to request Also update firmware selftests to use the new partial read test API. Signed-off-by:
Scott Branden <scott.branden@broadcom.com> Co-developed-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
- Oct 02, 2020
-
-
Joe Perches authored
Use semicolons and braces. Signed-off-by:
Joe Perches <joe@perches.com> Signed-off-by:
-
Brendan Higgins authored
Right now .kunitconfig and the build dir are automatically created if the build dir does not exists; however, if the build dir is present and .kunitconfig is not, kunit_tool will crash. Fix this by checking for both the build dir as well as the .kunitconfig. NOTE: This depends on commit 5578d008 ("kunit: tool: fix running kunit_tool from outside kernel tree") Link: https://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest.git/commit/?id=5578d008d9e06bb531fb3e62dd17096d9fd9c853 Signed-off-by:
-
- Sep 25, 2020
-
-
Peter Oskolkov authored
Based on Google-internal RSEQ work done by Paul Turner and Andrew Hunter. This patch adds a selftest for MEMBARRIER_CMD_PRIVATE_EXPEDITED_RSEQ. The test quite often fails without the previous patch in this patchset, but consistently passes with it. Signed-off-by:
Peter Oskolkov <posk@google.com> Signed-off-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Acked-by:
Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Link: https://lkml.kernel.org/r/20200923233618.2572849-3-posk@google.com
-
Peter Oskolkov authored
This patch adds rseq_offset_deref_addv() function to tools/testing/selftests/rseq/rseq-x86.h, to be used in a selftest in the next patch in the patchset. Once an architecture adds support for this function they should define "RSEQ_ARCH_HAS_OFFSET_DEREF_ADDV". Signed-off-by:
-
- Sep 23, 2020
-
-
Brendan Higgins authored
Alltests flag evidently stopped working when run from outside of the root of the source tree, so fix that. Also add an additional broken config to the broken_on_uml config. Signed-off-by:
-
Gabriel Krisman Bertazi authored
Prevent them from polluting git status after building selftests. Signed-off-by:
Gabriel Krisman Bertazi <krisman@collabora.com> Signed-off-by:
-
Yang Weijiang authored
This is a follow-up patch to fix an issue left in commit: 98b0bf02 selftests: kvm: Use a shorter encoding to clear RAX With the change in the commit, we also need to modify "xor" instruction length from 3 to 2 in array ss_size accordingly to pass below check: for (i = 0; i < (sizeof(ss_size) / sizeof(ss_size[0])); i++) { target_rip += ss_size[i]; CLEAR_DEBUG(); debug.control = KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_SINGLESTEP; debug.arch.debugreg[7] = 0x00000400; APPLY_DEBUG(); vcpu_run(vm, VCPU_ID); TEST_ASSERT(run->exit_reason == KVM_EXIT_DEBUG && run->debug.arch.exception == DB_VECTOR && run->debug.arch.pc == target_rip && run->debug.arch.dr6 == target_dr6, "SINGLE_STEP[%d]: exit %d exception %d rip 0x%llx " "(should be 0x%llx) dr6 0x%llx (should be 0x%llx)", i, run->exit_reason, run->debug.arch.exception, run->debug.arch.pc, target_rip, run->debug.arch.dr6, target_dr6); } Reported-by:
kernel test robot <rong.a.chen@intel.com> Signed-off-by:
Yang Weijiang <weijiang.yang@intel.com> Message-Id: <20200826015524.13251-1-weijiang.yang@intel.com> Signed-off-by:
Paolo Bonzini <pbonzini@redhat.com>
-
- Sep 19, 2020
-
-
Christophe Leroy authored
The displayed size is in bytes while the text says it is in kB. Shift it by 10 to really display kBytes. Fixes: fa7b9a80 ("tools/selftest/vm: allow choosing mem size and page size in map_hugetlb") Signed-off-by:
Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by:
-
Kees Cook authored
On powerpc, the errno is not inverted, and depends on ccr.so being set. Add this to a powerpc definition of SYSCALL_RET_SET(). Co-developed-by:
-
Kees Cook authored
Instead of special-casing the specific case of shared registers, create a default SYSCALL_RET_SET() macro (mirroring SYSCALL_NUM_SET()), that writes to the SYSCALL_RET register. For architectures that can't set the return value (for whatever reason), they can define SYSCALL_RET_SET() without an associated SYSCALL_RET() macro. This also paves the way for architectures that need to do special things to set the return value (e.g. powerpc). Signed-off-by:
-
Kees Cook authored
When none of the registers have changed, don't flush them back. This can happen if the architecture uses a non-register way to change the syscall (e.g. arm64) , and a return value hasn't been written. Signed-off-by:
-
Kees Cook authored
Consolidate the REGSET logic into the new ARCH_GETREG() and ARCH_SETREG() macros, avoiding more #ifdef code in function bodies. Signed-off-by:
-
Kees Cook authored
Instead of special-casing the get/set-registers routines, move the HAVE_GETREG logic into the new ARCH_GETREG() and ARCH_SETREG() macros. Signed-off-by:
-
Kees Cook authored
With all architectures now using the common SYSCALL_NUM_SET() macro, the arch-specific #ifdef can be removed from change_syscall() itself. Signed-off-by:
-
Kees Cook authored
Instead of having the mips O32 macro special-cased, pull the logic into the SYSCALL_NUM() macro. Additionally include the ABI headers, since these appear to have been missing, leaving __NR_O32_Linux undefined. Signed-off-by:
-
Kees Cook authored
Remove the arm64 special-case in change_syscall(). Signed-off-by:
-
