July 2023 - Linux-kselftest-mirror

[PATCH v1 0/9] selftests/mm fixes for arm64

by Ryan Roberts

Hi All, Given my on-going work on large anon folios and contpte mappings, I decided it would be a good idea to start running mm selftests to help guard against regressions. However, it soon became clear that I couldn't get the suite to run cleanly on arm64 with a vanilla v6.5-rc1 kernel (perhaps I'm just doing it wrong??), so got stuck in a rabbit hole trying to debug and fix all the issues. Some were down to misconfigurations, but I also found a number of issues with the tests and even a couple of issues with the kernel. This series aims to fix (most of) the test issues. It applies on top of v6.5-rc1. Reproducing ----------- What follows is a write up of how I'm running the tests and the results I see with this series applied. I don't yet have a concrete understanding of all of the remaining failures. So if anyone has any comments on my setup or reasons for the test failures it would be great to hear. Source: v6.5-rc1 + this series + [1] + [2]. [1] is a patch from Florent Revest to fix mdwe mmap_FIXED tests. [2] is a fix for a regression in the kernel that I found by running `mlock-random-test` and `mlock2-tests`. Compile the kernel (on arm64 system): $ make defconfig $ ./scripts/config --enable CONFIG_SQUASHFS_LZ4 $ ./scripts/config --enable CONFIG_SQUASHFS_LZO $ ./scripts/config --enable CONFIG_SQUASHFS_XZ $ ./scripts/config --enable CONFIG_SQUASHFS_ZSTD $ ./scripts/config --enable CONFIG_XFS_FS $ ./scripts/config --enable CONFIG_SYSVIPC $ ./scripts/config --enable CONFIG_USERFAULTFD $ ./scripts/config --enable CONFIG_TEST_VMALLOC $ ./scripts/config --enable CONFIG_GUP_TEST $ ./scripts/config --enable CONFIG_TRANSPARENT_HUGEPAGE $ ./scripts/config --enable CONFIG_MEM_SOFT_DIRTY $ make olddefconfig $ make -s -j`nproc` Image (In the above case, I'm building/testing a 4K kernel). Note that it turns out that arm64 doesn't really support ZONE_DEVICE; Although it defines ARCH_HAS_PTE_DEVMAP, it can't allocate `struct page`s for arbitrary physical addresses. This means that the TEST_HMM module causes warnings to be emitted when initializing because it tries to reserve arbitrary PA range then requests struct page's for them. I haven't fully investigated this yet, but for now, I'm just deliverately excluding ZONE_DEVICE, (which TEST_HMM depends upon). This means that the `hmm-tests` selftest gets skipped at runtime. Compile the tests: $ make -j`nproc` headers_install $ make -C tools/testing/selftests TARGETS=mm install INSTALL_PATH=<path/to/install> Start a VM running the kernel we just compiled: $ taskset -c 8-15 qemu-system-aarch64 \ -object memory-backend-file,id=mem0,size=6G,mem-path=/hugetlbfs,merge=off,prealloc=on,host-nodes=0,policy=bind,align=1G \ -object memory-backend-file,id=mem1,size=6G,mem-path=/hugetlbfs,merge=off,prealloc=on,host-nodes=0,policy=bind,align=1G \ -nographic -enable-kvm -machine virt,gic-version=3 -cpu max \ -smp 8 -m 12G \ -numa node,memdev=mem0,cpus=0-3,nodeid=0 \ -numa node,memdev=mem1,cpus=4-7,nodeid=1 \ -drive if=virtio,format=raw,file=ubuntu-22.04.xfs \ -object rng-random,filename=/dev/urandom,id=rng0 \ -device virtio-scsi-pci,id=scsi0 \ -netdev user,id=net0,hostfwd=tcp::8022-:22 \ -device virtio-rng-pci,rng=rng0 \ -device virtio-net-pci,netdev=net0 \ -kernel arch/arm64/boot/Image \ -append "earlycon root=/dev/vda2 secretmem.enable hugepagesz=1G hugepages=0:2,1:2 hugepagesz=32M hugepages=0:2,1:2 default_hugepagesz=2M hugepages=0:64,1:64 hugepagesz=64K hugepages=0:2,1:2" This starts a VM with 2 numa nodes (needed by ksm and migration tests), with 6G of memory and 4 CPUs on each node. The kernel command line enables secretmem (needed for `memfd_secret` test), and preallocates a bunch of huge pages (divined by reading the comments and source for a bunch of tests that require huge pages). 128M of the default huge page size, and 4 pages of each of the other sizes appear to be sufficient. I'm allocating half on each numa node. Once booted, copy the selftests we just compiled onto it. On the VM, run the tests: $ cd path/to/selftests $ sudo ./run_kselftest.sh or alternatively: $ cd path/to/selftests/mm $ sudo ./run_vmtests.sh Test Results ------------ TOP-LEVEL SUMMARY: PASS=42 SKIP=4 FAIL=2 Only showing nested tests if they are skipped or failed. [PASS] hugepage-mmap [PASS] hugepage-shm [PASS] map_hugetlb [PASS] hugepage-mremap [PASS] hugepage-vmemmap [PASS] hugetlb-madvise [PASS] map_fixed_noreplace [PASS] gup_test -u [PASS] gup_test -a [PASS] gup_test -ct -F 0x1 0 19 0x1000 [PASS] gup_longterm [PASS] uffd-unit-tests [PASS] uffd-stress anon 20 16 [PASS] uffd-stress hugetlb 128 32 [PASS] uffd-stress hugetlb-private 128 32 [PASS] uffd-stress shmem 20 16 [PASS] uffd-stress shmem-private 20 16 [PASS] compaction_test [PASS] on-fault-limit [PASS] map_populate [PASS] mlock-random-test [PASS] mlock2-tests [PASS] mrelease_test [PASS] mremap_test [PASS] thuge-gen [PASS] virtual_address_range [SKIP] va_high_addr_switch.sh # 4K kernel does not support big enough VA space for test [SKIP] test_vmalloc.sh smoke # Test requires test_vmalloc kernel module which isn't present [PASS] mremap_dontunmap [SKIP] test_hmm.sh smoke # Test requires test_hmm kernel module - see ZONE_DEVICE issue above [PASS] madv_populate [PASS] test_softdirty [SKIP] range is not softdirty [SKIP] MADV_POPULATE_READ [SKIP] range is not softdirty [SKIP] MADV_POPULATE_WRITE [SKIP] range is softdirty # All skipped because arm64 does not support soft-dirty [PASS] memfd_secret [PASS] ksm_tests -M -p 10 [PASS] ksm_tests -U [PASS] ksm_tests -Z -p 10 -z 0 [PASS] ksm_tests -Z -p 10 -z 1 [PASS] ksm_tests -N -m 1 [PASS] ksm_tests -N -m 0 [PASS] ksm_functional_tests [SKIP] test_unmerge_uffd_wp # UFFD_FEATURE_PAGEFAULT_FLAG_WP not available on arm64 [PASS] ksm_functional_tests [SKIP] test_unmerge_uffd_wp # UFFD_FEATURE_PAGEFAULT_FLAG_WP not available on arm64 [SKIP] soft-dirty # Skipped because arm64 does not support soft-dirty [FAIL] cow [FAIL] vmsplice() + unmap in child ... with hugetlb [FAIL] vmsplice() + unmap in child with mprotect() optimization ... with hugetlb [FAIL] vmsplice() before fork(), unmap in parent after fork() ... with hugetlb [FAIL] vmsplice() + unmap in parent after fork() ... with hugetlb # Above are known issues for vmsplice + hugetlb # Reproduces on x86 [SKIP] Basic COW after fork() ... with swapped-out, PTE-mapped THP [SKIP] Basic COW after fork() with mprotect() optimization ... with swapped-out, PTE-mapped THP [SKIP] vmsplice() + unmap in child ... with swapped-out, PTE-mapped THP [SKIP] vmsplice() + unmap in child with mprotect() optimization ... with swapped-out, PTE-mapped THP [SKIP] vmsplice() before fork(), unmap in parent after fork() ... with swapped-out, PTE-mapped THP [SKIP] vmsplice() + unmap in parent after fork() ... with swapped-out, PTE-mapped THP [SKIP] R/O-mapping a page registered as iouring fixed buffer ... with swapped-out, PTE-mapped THP [SKIP] fork() with an iouring fixed buffer ... with swapped-out, PTE-mapped THP [SKIP] R/O GUP pin on R/O-mapped shared page ... with swapped-out, PTE-mapped THP [SKIP] R/O GUP-fast pin on R/O-mapped shared page ... with swapped-out, PTE-mapped THP [SKIP] R/O GUP pin on R/O-mapped previously-shared page ... with swapped-out, PTE-mapped THP [SKIP] R/O GUP-fast pin on R/O-mapped previously-shared page ... with swapped-out, PTE-mapped THP [SKIP] R/O GUP pin on R/O-mapped exclusive page ... with swapped-out, PTE-mapped THP [SKIP] R/O GUP-fast pin on R/O-mapped exclusive page ... with swapped-out, PTE-mapped THP # Above all skipped due to "MADV_PAGEOUT did not work, is swap enabled?" # swap is enabled though # Reproduces on x86 [SKIP] Basic COW after fork() when collapsing after fork() (fully shared) # MADV_COLLAPSE failed: Invalid argument [PASS] khugepaged [PASS] transhuge-stress -d 20 [PASS] split_huge_page_test [FAIL] migration [FAIL] migration.shared_anon # move_pages() reports that the requested page was not migrated # after a few iterations. [PASS] mkdirty [PASS] mdwe_test [1] https://lore.kernel.org/lkml/20230704153630.1591122-3-revest@chromium.org/ [2] https://lore.kernel.org/linux-mm/20230711175020.4091336-1-Liam.Howlett@orac… Thanks, Ryan Ryan Roberts (9): selftests: Line buffer test program's stdout selftests/mm: Give scripts execute permission selftests/mm: Skip soft-dirty tests on arm64 selftests/mm: Enable mrelease_test for arm64 selftests/mm: Fix thuge-gen test bugs selftests/mm: va_high_addr_switch should skip unsupported arm64 configs selftests/mm: Make migration test robust to failure selftests/mm: Optionally pass duration to transhuge-stress selftests/mm: Run all tests from run_vmtests.sh tools/testing/selftests/kselftest/runner.sh | 5 +- tools/testing/selftests/mm/Makefile | 79 ++++++++++--------- .../selftests/mm/charge_reserved_hugetlb.sh | 0 tools/testing/selftests/mm/check_config.sh | 0 .../selftests/mm/hugetlb_reparenting_test.sh | 0 tools/testing/selftests/mm/madv_populate.c | 18 +++-- tools/testing/selftests/mm/migration.c | 14 +++- tools/testing/selftests/mm/mrelease_test.c | 1 + tools/testing/selftests/mm/run_vmtests.sh | 23 ++++++ tools/testing/selftests/mm/settings | 2 +- tools/testing/selftests/mm/soft-dirty.c | 3 + tools/testing/selftests/mm/test_hmm.sh | 0 tools/testing/selftests/mm/test_vmalloc.sh | 0 tools/testing/selftests/mm/thuge-gen.c | 4 +- tools/testing/selftests/mm/transhuge-stress.c | 12 ++- .../selftests/mm/va_high_addr_switch.c | 3 +- .../selftests/mm/va_high_addr_switch.sh | 0 tools/testing/selftests/mm/vm_util.c | 17 ++++ tools/testing/selftests/mm/vm_util.h | 1 + .../selftests/mm/write_hugetlb_memory.sh | 0 20 files changed, 127 insertions(+), 55 deletions(-) mode change 100644 => 100755 tools/testing/selftests/mm/charge_reserved_hugetlb.sh mode change 100644 => 100755 tools/testing/selftests/mm/check_config.sh mode change 100644 => 100755 tools/testing/selftests/mm/hugetlb_reparenting_test.sh mode change 100644 => 100755 tools/testing/selftests/mm/run_vmtests.sh mode change 100644 => 100755 tools/testing/selftests/mm/test_hmm.sh mode change 100644 => 100755 tools/testing/selftests/mm/test_vmalloc.sh mode change 100644 => 100755 tools/testing/selftests/mm/va_high_addr_switch.sh mode change 100644 => 100755 tools/testing/selftests/mm/write_hugetlb_memory.sh -- 2.25.1

1 year, 5 months

6
39
0 0

[PATCH] selftests/bpf: Fix errors in log_fixup.c

by shijie001＠208suo.com

The following checkpatch errors are removed: ERROR: "foo* bar" should be "foo *bar" Signed-off-by: Jie Shi <shijie001(a)208suo.com> --- tools/testing/selftests/bpf/prog_tests/log_fixup.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/tools/testing/selftests/bpf/prog_tests/log_fixup.c b/tools/testing/selftests/bpf/prog_tests/log_fixup.c index dba71d98a227..69959748d466 100644 --- a/tools/testing/selftests/bpf/prog_tests/log_fixup.c +++ b/tools/testing/selftests/bpf/prog_tests/log_fixup.c @@ -14,7 +14,7 @@ enum trunc_type { static void bad_core_relo(size_t log_buf_size, enum trunc_type trunc_type) { char log_buf[8 * 1024]; - struct test_log_fixup* skel; + struct test_log_fixup *skel; int err; skel = test_log_fixup__open(); @@ -72,7 +72,7 @@ static void bad_core_relo(size_t log_buf_size, enum trunc_type trunc_type) static void bad_core_relo_subprog(void) { char log_buf[8 * 1024]; - struct test_log_fixup* skel; + struct test_log_fixup *skel; int err; skel = test_log_fixup__open(); @@ -104,7 +104,7 @@ static void bad_core_relo_subprog(void) static void missing_map(void) { char log_buf[8 * 1024]; - struct test_log_fixup* skel; + struct test_log_fixup *skel; int err; skel = test_log_fixup__open(); @@ -138,7 +138,7 @@ static void missing_map(void) static void missing_kfunc(void) { char log_buf[8 * 1024]; - struct test_log_fixup* skel; + struct test_log_fixup *skel; int err; skel = test_log_fixup__open();

1 year, 5 months

1
0
0 0

[PATCH] selftests/bpf: Fix errors in btf_write.c

by shijie001＠208suo.com

The following checkpatch errors are removed: ERROR: Bad function definition - void test_btf_add() should probably be void test_btf_add(void) ERROR: Bad function definition - void test_btf_add_btf() should probably be void test_btf_add_btf(void) ERROR: Bad function definition - void test_btf_write() should probably be void test_btf_write(void) Signed-off-by: Jie Shi <shijie001(a)208suo.com> --- tools/testing/selftests/bpf/prog_tests/btf_write.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tools/testing/selftests/bpf/prog_tests/btf_write.c b/tools/testing/selftests/bpf/prog_tests/btf_write.c index 6e36de1302fc..209a230ffeec 100644 --- a/tools/testing/selftests/bpf/prog_tests/btf_write.c +++ b/tools/testing/selftests/bpf/prog_tests/btf_write.c @@ -352,7 +352,7 @@ static void gen_btf(struct btf *btf) "\t'v1' val=18446744073709551615", "raw_dump"); } -static void test_btf_add() +static void test_btf_add(void) { struct btf *btf; @@ -401,7 +401,7 @@ static void test_btf_add() btf__free(btf); } -static void test_btf_add_btf() +static void test_btf_add_btf(void) { struct btf *btf1 = NULL, *btf2 = NULL; int id; @@ -497,7 +497,7 @@ static void test_btf_add_btf() btf__free(btf2); } -void test_btf_write() +void test_btf_write(void) { if (test__start_subtest("btf_add")) test_btf_add();

1 year, 5 months

1
0
0 0

[PATCH v3] selftests/net: replace manual array size calc with ARRAYSIZE macro.

by Mahmoud Maatuq

fixes coccinelle WARNING: Use ARRAY_SIZE changelog since v2: - integrat a commit that contains actual replacement for ARRAY_SIZE. - use ARRAY_SIZE for net/psock_lib.h changelog since v1: - remove unnecessary extra new line changelog since v0: - update net/Makefile to include kselftest.h - remove redefinition of ARRAYSIZE. Signed-off-by: Mahmoud Maatuq <mahmoudmatook.mm(a)gmail.com> --- tools/testing/selftests/net/Makefile | 2 ++ tools/testing/selftests/net/csum.c | 6 ++++-- tools/testing/selftests/net/hwtstamp_config.c | 6 ++++-- tools/testing/selftests/net/psock_lib.h | 4 +++- 4 files changed, 13 insertions(+), 5 deletions(-) diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile index 7f3ab2a93ed6..a06cc25489f9 100644 --- a/tools/testing/selftests/net/Makefile +++ b/tools/testing/selftests/net/Makefile @@ -3,6 +3,8 @@ CFLAGS = -Wall -Wl,--no-as-needed -O2 -g CFLAGS += -I../../../../usr/include/ $(KHDR_INCLUDES) +# Additional include paths needed by kselftest.h +CFLAGS += -I../ TEST_PROGS := run_netsocktests run_afpackettests test_bpf.sh netdevice.sh \ rtnetlink.sh xfrm_policy.sh test_blackhole_dev.sh diff --git a/tools/testing/selftests/net/csum.c b/tools/testing/selftests/net/csum.c index 82a1c1839da6..90eb06fefa59 100644 --- a/tools/testing/selftests/net/csum.c +++ b/tools/testing/selftests/net/csum.c @@ -91,6 +91,8 @@ #include <sys/types.h> #include <unistd.h> +#include "kselftest.h" + static bool cfg_bad_csum; static int cfg_family = PF_INET6; static int cfg_num_pkt = 4; @@ -450,7 +452,7 @@ static void send_packet(int fd, const char *buf, int len) iov[2].iov_len = len; msg.msg_iov = iov; - msg.msg_iovlen = sizeof(iov) / sizeof(iov[0]); + msg.msg_iovlen = ARRAY_SIZE(iov); msg.msg_name = &addr; msg.msg_namelen = sizeof(addr); @@ -505,7 +507,7 @@ static void __recv_prepare_packet_filter(int fd, int off_nexthdr, int off_dport) struct sock_fprog prog = {}; prog.filter = filter; - prog.len = sizeof(filter) / sizeof(struct sock_filter); + prog.len = ARRAY_SIZE(filter); if (setsockopt(fd, SOL_SOCKET, SO_ATTACH_FILTER, &prog, sizeof(prog))) error(1, errno, "setsockopt filter"); } diff --git a/tools/testing/selftests/net/hwtstamp_config.c b/tools/testing/selftests/net/hwtstamp_config.c index e1fdee841021..170728c96c46 100644 --- a/tools/testing/selftests/net/hwtstamp_config.c +++ b/tools/testing/selftests/net/hwtstamp_config.c @@ -16,6 +16,8 @@ #include <linux/net_tstamp.h> #include <linux/sockios.h> +#include "kselftest.h" + static int lookup_value(const char **names, int size, const char *name) { @@ -50,7 +52,7 @@ static const char *tx_types[] = { TX_TYPE(ONESTEP_SYNC) #undef TX_TYPE }; -#define N_TX_TYPES ((int)(sizeof(tx_types) / sizeof(tx_types[0]))) +#define N_TX_TYPES ((int)(ARRAY_SIZE(tx_types))) static const char *rx_filters[] = { #define RX_FILTER(name) [HWTSTAMP_FILTER_ ## name] = #name @@ -71,7 +73,7 @@ static const char *rx_filters[] = { RX_FILTER(PTP_V2_DELAY_REQ), #undef RX_FILTER }; -#define N_RX_FILTERS ((int)(sizeof(rx_filters) / sizeof(rx_filters[0]))) +#define N_RX_FILTERS ((int)(ARRAY_SIZE(rx_filters))) static void usage(void) { diff --git a/tools/testing/selftests/net/psock_lib.h b/tools/testing/selftests/net/psock_lib.h index faa884385c45..6e4fef560873 100644 --- a/tools/testing/selftests/net/psock_lib.h +++ b/tools/testing/selftests/net/psock_lib.h @@ -14,6 +14,8 @@ #include <arpa/inet.h> #include <unistd.h> +#include "kselftest.h" + #define DATA_LEN 100 #define DATA_CHAR 'a' #define DATA_CHAR_1 'b' @@ -63,7 +65,7 @@ static __maybe_unused void pair_udp_setfilter(int fd) struct sock_fprog bpf_prog; bpf_prog.filter = bpf_filter; - bpf_prog.len = sizeof(bpf_filter) / sizeof(struct sock_filter); + bpf_prog.len = ARRAY_SIZE(bpf_filter); if (setsockopt(fd, SOL_SOCKET, SO_ATTACH_FILTER, &bpf_prog, sizeof(bpf_prog))) { -- 2.34.1

1 year, 5 months

1
0
0 0

[PATCH v2] selftests/net: replace manual array size calc with ARRAYSIZE macro.

by Mahmoud Maatuq

fixes coccinelle WARNING: Use ARRAY_SIZE changelog since v1: - remove unnecessary extra new line changelog since v0: - update net/Makefile to include kselftest.h - remove redefinition of ARRAYSIZE. Signed-off-by: Mahmoud Maatuq <mahmoudmatook.mm(a)gmail.com> --- tools/testing/selftests/net/Makefile | 2 ++ tools/testing/selftests/net/csum.c | 6 ++---- tools/testing/selftests/net/hwtstamp_config.c | 4 +--- 3 files changed, 5 insertions(+), 7 deletions(-) diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile index 7f3ab2a93ed6..a06cc25489f9 100644 --- a/tools/testing/selftests/net/Makefile +++ b/tools/testing/selftests/net/Makefile @@ -3,6 +3,8 @@ CFLAGS = -Wall -Wl,--no-as-needed -O2 -g CFLAGS += -I../../../../usr/include/ $(KHDR_INCLUDES) +# Additional include paths needed by kselftest.h +CFLAGS += -I../ TEST_PROGS := run_netsocktests run_afpackettests test_bpf.sh netdevice.sh \ rtnetlink.sh xfrm_policy.sh test_blackhole_dev.sh diff --git a/tools/testing/selftests/net/csum.c b/tools/testing/selftests/net/csum.c index 702f34a9d506..90eb06fefa59 100644 --- a/tools/testing/selftests/net/csum.c +++ b/tools/testing/selftests/net/csum.c @@ -91,6 +91,8 @@ #include <sys/types.h> #include <unistd.h> +#include "kselftest.h" + static bool cfg_bad_csum; static int cfg_family = PF_INET6; static int cfg_num_pkt = 4; @@ -123,10 +125,6 @@ static struct sockaddr_in6 cfg_saddr6 = {.sin6_family = AF_INET6}; #define MAX_HEADER_LEN (sizeof(struct ipv6hdr) + ENC_HEADER_LEN + sizeof(struct tcphdr)) #define MAX_PAYLOAD_LEN 1024 -#ifndef ARRAY_SIZE -#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) -#endif - /* Trivial demo encap. Stand-in for transport layer protocols like ESP or PSP */ struct udp_encap_hdr { uint8_t nexthdr; diff --git a/tools/testing/selftests/net/hwtstamp_config.c b/tools/testing/selftests/net/hwtstamp_config.c index 263cc1c34165..60970d98eb9b 100644 --- a/tools/testing/selftests/net/hwtstamp_config.c +++ b/tools/testing/selftests/net/hwtstamp_config.c @@ -16,9 +16,7 @@ #include <linux/net_tstamp.h> #include <linux/sockios.h> -#ifndef ARRAY_SIZE -#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) -#endif +#include "kselftest.h" static int lookup_value(const char **names, int size, const char *name) -- 2.34.1

1 year, 5 months

1
0
0 0

[PATCH v4 00/18] tools/nolibc: shrink arch support

by Zhangjin Wu

Hi, Willy, Thomas Thanks very much for your careful review and great suggestions, now, we get v4 revision of the arch shrink series [1], it mainly include a new fixup for -O0 under gcc < 11.1.0, the stackprotector support for _start_c(), new testcases for startup code and two new test targets. All of the tests passed or skipped (tinyconfig + few options + qemu-system) for both -Os and -O0: arch/board | result ------------|------------ arm/versatilepb | 165 test(s): 158 passed, 7 skipped, 0 failed => status: warning. arm/vexpress-a9 | 165 test(s): 158 passed, 7 skipped, 0 failed => status: warning. arm/virt | 165 test(s): 158 passed, 7 skipped, 0 failed => status: warning. aarch64/virt | 165 test(s): 158 passed, 7 skipped, 0 failed => status: warning. i386/pc | 165 test(s): 158 passed, 7 skipped, 0 failed => status: warning. x86_64/pc | 165 test(s): 158 passed, 7 skipped, 0 failed => status: warning. mipsel/malta | 165 test(s): 158 passed, 7 skipped, 0 failed => status: warning. loongarch64/virt | 165 test(s): 158 passed, 7 skipped, 0 failed => status: warning. riscv64/virt | 165 test(s): 158 passed, 7 skipped, 0 failed => status: warning. s390x/s390-ccw-virtio | 165 test(s): 158 passed, 7 skipped, 0 failed => status: warning. And more, for both -Os and -O0: $ for r in run-user run-nolibc-test run-libc-test; do make clean > /dev/null; make $r | grep status; done 165 test(s): 157 passed, 8 skipped, 0 failed => status: warning 165 test(s): 157 passed, 8 skipped, 0 failed => status: warning 165 test(s): 153 passed, 12 skipped, 0 failed => status: warning // for make run-user, the euid0 and 32bit limit related tests are // skipped $ make clean && make run-user $ grep -i skip run.out 17 chroot_root [SKIPPED] 39 link_dir [SKIPPED] 62 limit_intptr_min_32 [SKIPPED] 63 limit_intptr_max_32 [SKIPPED] 64 limit_uintptr_max_32 [SKIPPED] 65 limit_ptrdiff_min_32 [SKIPPED] 66 limit_ptrdiff_max_32 [SKIPPED] 67 limit_size_max_32 [SKIPPED] // for run-libc-test, the _auxv variables, euid0, 32bits limit and // stackprotector related tests are skipped $ make clean && make run-libc-test $ grep -i skip run.out 9 environ_auxv [SKIPPED] 10 environ_total [SKIPPED] 12 auxv_addr [SKIPPED] 17 chroot_root [SKIPPED] 39 link_dir [SKIPPED] 62 limit_intptr_min_32 [SKIPPED] 63 limit_intptr_max_32 [SKIPPED] 64 limit_uintptr_max_32 [SKIPPED] 65 limit_ptrdiff_min_32 [SKIPPED] 66 limit_ptrdiff_max_32 [SKIPPED] 67 limit_size_max_32 [SKIPPED] 0 -fstackprotector not supported [SKIPPED] $ make clean >/dev/null; make run-libc-test CC=/labs/linux-lab/src/examples/musl-install/bin/musl-gcc | grep status 165 test(s): 151 passed, 12 skipped, 2 failed => status: failure // The failures are expected for musl has disabled both sbrk and brk // but not the sbrk(0); the _auxv variables, euid0, 32bits limit and // stackprotector related tests are skipped for musl too $ grep FAIL -ur run.out 9 sbrk = 1 ENOMEM [FAIL] 10 brk = -1 ENOMEM [FAIL] $ grep "SKIP" -ur run.out 9 environ_auxv [SKIPPED] 10 environ_total [SKIPPED] 12 auxv_addr [SKIPPED] 17 chroot_root [SKIPPED] 39 link_dir [SKIPPED] 62 limit_intptr_min_32 [SKIPPED] 63 limit_intptr_max_32 [SKIPPED] 64 limit_uintptr_max_32 [SKIPPED] 65 limit_ptrdiff_min_32 [SKIPPED] 66 limit_ptrdiff_max_32 [SKIPPED] 67 limit_size_max_32 [SKIPPED] 0 -fstackprotector not supported [SKIPPED] For stackprotector, gcc 13.1.0 is used to test on x86_64 standalonely: $ make run-user CROSS_COMPILE=x86_64-linux- | grep status 165 test(s): 157 passed, 8 skipped, 0 failed => status: warning $ grep stack -ur run.out 0 -fstackprotector [OK] $ make run-nolibc-test CROSS_COMPILE=x86_64-linux- | grep status 165 test(s): 157 passed, 8 skipped, 0 failed => status: warning $ grep stack -ur run.out 0 -fstackprotector [OK] Changes from v3 --> v4: * tools/nolibc: arch-*.h: add missing space after ',' tools/nolibc: fix up startup failures for -O0 under gcc < 11.1.0 Both of the above changes are for _start, it is able to merge them if necessary. The first one is old for format errors reported by scripts/checkpatch.pl The second one is for -O0 failure under gcc < 11.1.0, applied the optimize("-Os", "omit-frame-pointer") suggestion from Thomas. * tools/nolibc: remove the old sys_stat support As suggested by Willy, Document carefully about the statx supported Linux version info. * tools/nolibc: add new crt.h with _start_c The code is polished carefully for smaller size and better readability. * tools/nolibc: stackprotector.h: add empty __stack_chk_init for !_NOLIBC_STACKPROTECTOR tools/nolibc: crt.h: initialize stack protector As suggested by Thomas, init stackprotector in _start_c() too. * tools/nolibc: arm: shrink _start with _start_c tools/nolibc: aarch64: shrink _start with _start_c tools/nolibc: i386: shrink _start with _start_c tools/nolibc: x86_64: shrink _start with _start_c tools/nolibc: mips: shrink _start with _start_c tools/nolibc: loongarch: shrink _start with _start_c tools/nolibc: riscv: shrink _start with _start_c tools/nolibc: s390: shrink _start with _start_c Removed the stackprotector initialization from _start too, we already have it in _start_c(). * selftests/nolibc: add EXPECT_PTRGE, EXPECT_PTRGT, EXPECT_PTRLE, EXPECT_PTRLT selftests/nolibc: add testcases for startup code Add a new startup test group to cover the testing of argc, argv/argv0, envp/environ and _auxv. Some testcases are enhanced, some are newly added from after the discussion during v3 review. * selftests/nolibc: allow run nolibc-test locally selftests/nolibc: allow test -include /path/to/nolibc.h Two new test targets are added to cover more scenes. Hope you like this revisoin ;-) Next patchset is powerpc & powerpc64 support, after that we will send the v2 of tinyconfig support, at last the left rv32 patches (mainly 64bit time). Best regards, Zhangjin --- [1]: https://lore.kernel.org/lkml/20230715100134.GD24086@1wt.eu/ Zhangjin Wu (18): tools/nolibc: arch-*.h: add missing space after ',' tools/nolibc: fix up startup failures for -O0 under gcc < 11.1.0 tools/nolibc: remove the old sys_stat support tools/nolibc: add new crt.h with _start_c tools/nolibc: stackprotector.h: add empty __stack_chk_init for !_NOLIBC_STACKPROTECTOR tools/nolibc: crt.h: initialize stack protector tools/nolibc: arm: shrink _start with _start_c tools/nolibc: aarch64: shrink _start with _start_c tools/nolibc: i386: shrink _start with _start_c tools/nolibc: x86_64: shrink _start with _start_c tools/nolibc: mips: shrink _start with _start_c tools/nolibc: loongarch: shrink _start with _start_c tools/nolibc: riscv: shrink _start with _start_c tools/nolibc: s390: shrink _start with _start_c selftests/nolibc: add EXPECT_PTRGE, EXPECT_PTRGT, EXPECT_PTRLE, EXPECT_PTRLT selftests/nolibc: add testcases for startup code selftests/nolibc: allow run nolibc-test locally selftests/nolibc: allow test -include /path/to/nolibc.h tools/include/nolibc/Makefile | 1 + tools/include/nolibc/arch-aarch64.h | 57 +--------- tools/include/nolibc/arch-arm.h | 83 ++------------- tools/include/nolibc/arch-i386.h | 62 ++--------- tools/include/nolibc/arch-loongarch.h | 46 +------- tools/include/nolibc/arch-mips.h | 76 ++----------- tools/include/nolibc/arch-riscv.h | 69 ++---------- tools/include/nolibc/arch-s390.h | 63 ++--------- tools/include/nolibc/arch-x86_64.h | 58 ++-------- tools/include/nolibc/crt.h | 61 +++++++++++ tools/include/nolibc/stackprotector.h | 2 + tools/include/nolibc/sys.h | 63 ++--------- tools/include/nolibc/types.h | 4 +- tools/testing/selftests/nolibc/Makefile | 12 +++ tools/testing/selftests/nolibc/nolibc-test.c | 106 ++++++++++++++++++- 15 files changed, 246 insertions(+), 517 deletions(-) create mode 100644 tools/include/nolibc/crt.h -- 2.25.1

1 year, 5 months

2
21
0 0

[PATCH v3 00/11] tools/nolibc: shrink arch support

by Zhangjin Wu

Hi, Willy, Thomas Here is the revision of the v2 arch support shrink patchset [1], it mainly applies suggestions from you. It is based on the 20230710-nolibc-ser2-tom-syscall-configv4-report branch of nolibc repo. Tested for all of the supported archs: arch/board | result ------------|------------ arm/versatilepb | 151 test(s): 144 passed, 7 skipped, 0 failed => status: warning. arm/vexpress-a9 | 151 test(s): 144 passed, 7 skipped, 0 failed => status: warning. arm/virt | 151 test(s): 144 passed, 7 skipped, 0 failed => status: warning. aarch64/virt | 151 test(s): 144 passed, 7 skipped, 0 failed => status: warning. i386/pc | 151 test(s): 144 passed, 7 skipped, 0 failed => status: warning. x86_64/pc | 151 test(s): 144 passed, 7 skipped, 0 failed => status: warning. mipsel/malta | 151 test(s): 144 passed, 7 skipped, 0 failed => status: warning. loongarch64/virt | 151 test(s): 144 passed, 7 skipped, 0 failed => status: warning. riscv64/virt | 151 test(s): 144 passed, 7 skipped, 0 failed => status: warning. riscv32/virt | 151 test(s): 122 passed, 7 skipped, 22 failed => status: failure. s390x/s390-ccw-virtio | 151 test(s): 144 passed, 7 skipped, 0 failed => status: warning. Changes from v2 --> v3: * tools/nolibc: remove the old sys_stat support Revert the reorg operation, basically the same as v1 * tools/nolibc: add new crt.h with _start_c Revert the reorg operation Add crt.h in Makefile Add _nolibc_main alias for main to silence the compile warning about * tools/nolibc: arm: shrink _start with _start_c tools/nolibc: aarch64: shrink _start with _start_c tools/nolibc: i386: shrink _start with _start_c tools/nolibc: x86_64: shrink _start with _start_c tools/nolibc: mips: shrink _start with _start_c tools/nolibc: loongarch: shrink _start with _start_c tools/nolibc: riscv: shrink _start with _start_c tools/nolibc: s390: shrink _start with _start_c Revert the reorg operation, but still use post-whitespaces instead of post-tab. Include "crt.h" in arch-<ARCH>.h * tools/nolibc: arch-*.h: add missing space after ',' Fix up the errors reported by scripts/checkpatch.pl Best regards, Zhangjin --- [1]: https://lore.kernel.org/lkml/cover.1688828139.git.falcon@tinylab.org/ Zhangjin Wu (11): tools/nolibc: remove the old sys_stat support tools/nolibc: add new crt.h with _start_c tools/nolibc: arm: shrink _start with _start_c tools/nolibc: aarch64: shrink _start with _start_c tools/nolibc: i386: shrink _start with _start_c tools/nolibc: x86_64: shrink _start with _start_c tools/nolibc: mips: shrink _start with _start_c tools/nolibc: loongarch: shrink _start with _start_c tools/nolibc: riscv: shrink _start with _start_c tools/nolibc: s390: shrink _start with _start_c tools/nolibc: arch-*.h: add missing space after ',' tools/include/nolibc/Makefile | 1 + tools/include/nolibc/arch-aarch64.h | 56 ++---------------- tools/include/nolibc/arch-arm.h | 82 +++------------------------ tools/include/nolibc/arch-i386.h | 61 +++----------------- tools/include/nolibc/arch-loongarch.h | 45 ++------------- tools/include/nolibc/arch-mips.h | 76 ++++--------------------- tools/include/nolibc/arch-riscv.h | 68 +++------------------- tools/include/nolibc/arch-s390.h | 63 ++------------------ tools/include/nolibc/arch-x86_64.h | 57 +++---------------- tools/include/nolibc/crt.h | 59 +++++++++++++++++++ tools/include/nolibc/sys.h | 63 ++++---------------- tools/include/nolibc/types.h | 4 +- 12 files changed, 133 insertions(+), 502 deletions(-) create mode 100644 tools/include/nolibc/crt.h -- 2.25.1

1 year, 5 months

3
26
0 0

kselftest/next kselftest-seccomp: 2 runs, 1 regressions (v6.5-rc1-1-gfeac01aaf677)

by kernelci.org bot

kselftest/next kselftest-seccomp: 2 runs, 1 regressions (v6.5-rc1-1-gfeac01aaf677) Regressions Summary ------------------- platform | arch | lab | compiler | defconfig | regressions ----------------+-------+---------------+----------+------------------------------+------------ mt8173-elm-hana | arm64 | lab-collabora | gcc-10 | defconfig+kse...4-chromebook | 1 Details: https://kernelci.org/test/job/kselftest/branch/next/kernel/v6.5-rc1-1-gfeac… Test: kselftest-seccomp Tree: kselftest Branch: next Describe: v6.5-rc1-1-gfeac01aaf677 URL: https://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest.git SHA: feac01aaf677827a0b77115554e504ee444c9b7c Test Regressions ---------------- platform | arch | lab | compiler | defconfig | regressions ----------------+-------+---------------+----------+------------------------------+------------ mt8173-elm-hana | arm64 | lab-collabora | gcc-10 | defconfig+kse...4-chromebook | 1 Details: https://kernelci.org/test/plan/id/64b243511b2af9a9d18ace31 Results: 0 PASS, 1 FAIL, 0 SKIP Full config: defconfig+kselftest+arm64-chromebook Compiler: gcc-10 (aarch64-linux-gnu-gcc (Debian 10.2.1-6) 10.2.1 20210110) Plain log: https://storage.kernelci.org//kselftest/next/v6.5-rc1-1-gfeac01aaf677/arm64… HTML log: https://storage.kernelci.org//kselftest/next/v6.5-rc1-1-gfeac01aaf677/arm64… Rootfs: http://storage.kernelci.org/images/rootfs/debian/bullseye-kselftest/2023062… * kselftest-seccomp.login: https://kernelci.org/test/case/id/64b243511b2af9a9d18ace32 failing since 270 days (last pass: linux-kselftest-next-6.0-rc2-11-g144eeb2fc761, first fail: v6.1-rc1)

1 year, 5 months

1
0
0 0

kselftest/next build: 6 builds: 0 failed, 6 passed, 4 warnings (v6.5-rc1-1-gfeac01aaf677)

by kernelci.org bot

kselftest/next build: 6 builds: 0 failed, 6 passed, 4 warnings (v6.5-rc1-1-gfeac01aaf677) Full Build Summary: https://kernelci.org/build/kselftest/branch/next/kernel/v6.5-rc1-1-gfeac01a… Tree: kselftest Branch: next Git Describe: v6.5-rc1-1-gfeac01aaf677 Git Commit: feac01aaf677827a0b77115554e504ee444c9b7c Git URL: https://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest.git Built: 4 unique architectures Warnings Detected: arm64: arm: i386: x86_64: x86_64_defconfig+kselftest (gcc-10): 3 warnings x86_64_defconfig+kselftest (clang-16): 1 warning Warnings summary: 1 vmlinux.o: warning: objtool: set_ftrace_ops_ro+0x23: relocation to !ENDBR: relocate_range+0xbb 1 vmlinux.o: warning: objtool: iovec_from_user+0x88: call to copy_iovec_from_user.part.0() with UACCESS enabled 1 vmlinux.o: warning: objtool: __import_iovec+0x147: call to copy_iovec_from_user.part.0() with UACCESS enabled 1 vmlinux.o: warning: objtool: .altinstr_replacement+0x1d1e: redundant UACCESS disable ================================================================================ Detailed per-defconfig build reports: -------------------------------------------------------------------------------- defconfig+kselftest (arm64, gcc-10) — PASS, 0 errors, 0 warnings, 0 section mismatches -------------------------------------------------------------------------------- defconfig+kselftest+arm64-chromebook (arm64, gcc-10) — PASS, 0 errors, 0 warnings, 0 section mismatches -------------------------------------------------------------------------------- i386_defconfig+kselftest (i386, gcc-10) — PASS, 0 errors, 0 warnings, 0 section mismatches -------------------------------------------------------------------------------- multi_v7_defconfig+kselftest (arm, gcc-10) — PASS, 0 errors, 0 warnings, 0 section mismatches -------------------------------------------------------------------------------- x86_64_defconfig+kselftest (x86_64, gcc-10) — PASS, 0 errors, 3 warnings, 0 section mismatches Warnings: vmlinux.o: warning: objtool: .altinstr_replacement+0x1d1e: redundant UACCESS disable vmlinux.o: warning: objtool: iovec_from_user+0x88: call to copy_iovec_from_user.part.0() with UACCESS enabled vmlinux.o: warning: objtool: __import_iovec+0x147: call to copy_iovec_from_user.part.0() with UACCESS enabled -------------------------------------------------------------------------------- x86_64_defconfig+kselftest (x86_64, clang-16) — PASS, 0 errors, 1 warning, 0 section mismatches Warnings: vmlinux.o: warning: objtool: set_ftrace_ops_ro+0x23: relocation to !ENDBR: relocate_range+0xbb --- For more info write to <info(a)kernelci.org>

1 year, 5 months

1
0
0 0

[PATCH v1] selftests/net: replace manual array size calc with ARRAYSIZE macro.

by Mahmoud Maatuq

fixes coccinelle WARNING: Use ARRAY_SIZE changelog since v0: - update net/Makefile to include kselftest.h - remove redefinition of ARRAYSIZE. Signed-off-by: Mahmoud Maatuq <mahmoudmatook.mm(a)gmail.com> --- tools/testing/selftests/net/Makefile | 2 ++ tools/testing/selftests/net/csum.c | 6 ++---- tools/testing/selftests/net/hwtstamp_config.c | 5 ++--- 3 files changed, 6 insertions(+), 7 deletions(-) diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile index 7f3ab2a93ed6..a06cc25489f9 100644 --- a/tools/testing/selftests/net/Makefile +++ b/tools/testing/selftests/net/Makefile @@ -3,6 +3,8 @@ CFLAGS = -Wall -Wl,--no-as-needed -O2 -g CFLAGS += -I../../../../usr/include/ $(KHDR_INCLUDES) +# Additional include paths needed by kselftest.h +CFLAGS += -I../ TEST_PROGS := run_netsocktests run_afpackettests test_bpf.sh netdevice.sh \ rtnetlink.sh xfrm_policy.sh test_blackhole_dev.sh diff --git a/tools/testing/selftests/net/csum.c b/tools/testing/selftests/net/csum.c index 702f34a9d506..90eb06fefa59 100644 --- a/tools/testing/selftests/net/csum.c +++ b/tools/testing/selftests/net/csum.c @@ -91,6 +91,8 @@ #include <sys/types.h> #include <unistd.h> +#include "kselftest.h" + static bool cfg_bad_csum; static int cfg_family = PF_INET6; static int cfg_num_pkt = 4; @@ -123,10 +125,6 @@ static struct sockaddr_in6 cfg_saddr6 = {.sin6_family = AF_INET6}; #define MAX_HEADER_LEN (sizeof(struct ipv6hdr) + ENC_HEADER_LEN + sizeof(struct tcphdr)) #define MAX_PAYLOAD_LEN 1024 -#ifndef ARRAY_SIZE -#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) -#endif - /* Trivial demo encap. Stand-in for transport layer protocols like ESP or PSP */ struct udp_encap_hdr { uint8_t nexthdr; diff --git a/tools/testing/selftests/net/hwtstamp_config.c b/tools/testing/selftests/net/hwtstamp_config.c index 263cc1c34165..873f4937f37f 100644 --- a/tools/testing/selftests/net/hwtstamp_config.c +++ b/tools/testing/selftests/net/hwtstamp_config.c @@ -16,9 +16,8 @@ #include <linux/net_tstamp.h> #include <linux/sockios.h> -#ifndef ARRAY_SIZE -#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) -#endif +#include "kselftest.h" + static int lookup_value(const char **names, int size, const char *name) -- 2.34.1

1 year, 5 months

2
1
0 0

2024

2023

2022

2021

2020

2019

2018

2017

Linux-kselftest-mirror July 2023