* [PATCH 0/3] selftests/mm: add memory failure selftests
@ 2026-01-07 9:37 Miaohe Lin
2026-01-07 9:37 ` [PATCH 1/3] selftests/mm: add memory failure anonymous page test Miaohe Lin
` (3 more replies)
0 siblings, 4 replies; 5+ messages in thread
From: Miaohe Lin @ 2026-01-07 9:37 UTC (permalink / raw)
To: akpm, shuah
Cc: david, lorenzo.stoakes, Liam.Howlett, vbabka, rppt, surenb,
mhocko, nao.horiguchi, linmiaohe, linux-mm, linux-kernel,
linux-kselftest
Introduce selftests to validate the functionality of memory failure.
These tests help ensure that memory failure handling for anonymous
pages, pagecaches pages works correctly, including proper SIGBUS
delivery to user processes, page isolation, and recovery paths.
Currently madvise syscall is used to inject memory failures. And only
anonymous pages and pagecaches are tested. More test scenarios, e.g.
hugetlb, shmem, thp, will be added. Also more memory failure injecting
methods will be supported, e.g. APEI Error INJection, if required.
Thanks!
Miaohe Lin (3):
selftests/mm: add memory failure anonymous page test
selftests/mm: add memory failure clean pagecache test
selftests/mm: add memory failure dirty pagecache test
MAINTAINERS | 1 +
tools/testing/selftests/mm/.gitignore | 1 +
tools/testing/selftests/mm/Makefile | 1 +
tools/testing/selftests/mm/memory-failure.c | 335 ++++++++++++++++++++
tools/testing/selftests/mm/run_vmtests.sh | 21 ++
tools/testing/selftests/mm/vm_util.c | 41 +++
tools/testing/selftests/mm/vm_util.h | 3 +
7 files changed, 403 insertions(+)
create mode 100644 tools/testing/selftests/mm/memory-failure.c
--
2.33.0
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 1/3] selftests/mm: add memory failure anonymous page test
2026-01-07 9:37 [PATCH 0/3] selftests/mm: add memory failure selftests Miaohe Lin
@ 2026-01-07 9:37 ` Miaohe Lin
2026-01-07 9:37 ` [PATCH 2/3] selftests/mm: add memory failure clean pagecache test Miaohe Lin
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Miaohe Lin @ 2026-01-07 9:37 UTC (permalink / raw)
To: akpm, shuah
Cc: david, lorenzo.stoakes, Liam.Howlett, vbabka, rppt, surenb,
mhocko, nao.horiguchi, linmiaohe, linux-mm, linux-kernel,
linux-kselftest
This patch adds a new kselftest to validate memory failure handling for
anonymous pages. The test performs the following operations:
1. Allocates anonymous pages using mmap().
2. Injects memory failure via madvise syscall.
3. Verifies expected error handling behavior.
4. Unpoison memory.
This test helps ensure that memory failure handling for anonymous pages
works correctly, including proper SIGBUS delivery to user processes, page
isolation and recovery paths.
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
MAINTAINERS | 1 +
tools/testing/selftests/mm/.gitignore | 1 +
tools/testing/selftests/mm/Makefile | 1 +
tools/testing/selftests/mm/memory-failure.c | 239 ++++++++++++++++++++
tools/testing/selftests/mm/run_vmtests.sh | 21 ++
tools/testing/selftests/mm/vm_util.c | 41 ++++
tools/testing/selftests/mm/vm_util.h | 3 +
7 files changed, 307 insertions(+)
create mode 100644 tools/testing/selftests/mm/memory-failure.c
diff --git a/MAINTAINERS b/MAINTAINERS
index 765ad2daa218..3401b408a07e 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -11683,6 +11683,7 @@ F: include/linux/memory-failure.h
F: include/trace/events/memory-failure.h
F: mm/hwpoison-inject.c
F: mm/memory-failure.c
+F: tools/testing/selftests/mm/memory-failure.c
HYCON HY46XX TOUCHSCREEN SUPPORT
M: Giulio Benetti <giulio.benetti@benettiengineering.com>
diff --git a/tools/testing/selftests/mm/.gitignore b/tools/testing/selftests/mm/.gitignore
index c2a8586e51a1..83ad9454dd9d 100644
--- a/tools/testing/selftests/mm/.gitignore
+++ b/tools/testing/selftests/mm/.gitignore
@@ -12,6 +12,7 @@ map_hugetlb
map_populate
thuge-gen
compaction_test
+memory-failure
migration
mlock2-tests
mrelease_test
diff --git a/tools/testing/selftests/mm/Makefile b/tools/testing/selftests/mm/Makefile
index eaf9312097f7..9495f7594e28 100644
--- a/tools/testing/selftests/mm/Makefile
+++ b/tools/testing/selftests/mm/Makefile
@@ -75,6 +75,7 @@ TEST_GEN_FILES += map_populate
ifneq (,$(filter $(ARCH),arm64 riscv riscv64 x86 x86_64))
TEST_GEN_FILES += memfd_secret
endif
+TEST_GEN_FILES += memory-failure
TEST_GEN_FILES += migration
TEST_GEN_FILES += mkdirty
TEST_GEN_FILES += mlock-random-test
diff --git a/tools/testing/selftests/mm/memory-failure.c b/tools/testing/selftests/mm/memory-failure.c
new file mode 100644
index 000000000000..37806a58f4b4
--- /dev/null
+++ b/tools/testing/selftests/mm/memory-failure.c
@@ -0,0 +1,239 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Memory-failure functional tests.
+ *
+ * Author(s): Miaohe Lin <linmiaohe@huawei.com>
+ */
+
+#include "../kselftest_harness.h"
+
+#include <sys/mman.h>
+#include <linux/mman.h>
+#include <linux/string.h>
+#include <signal.h>
+#include <setjmp.h>
+#include <unistd.h>
+#include <fcntl.h>
+
+#include "vm_util.h"
+
+enum inject_type {
+ MADV_HARD,
+ MADV_SOFT,
+};
+
+enum result_type {
+ MADV_HARD_ANON,
+ MADV_SOFT_ANON,
+};
+
+static jmp_buf signal_jmp_buf;
+static siginfo_t siginfo;
+const char *pagemap_proc = "/proc/self/pagemap";
+const char *kpageflags_proc = "/proc/kpageflags";
+
+FIXTURE(memory_failure)
+{
+ unsigned long page_size;
+ unsigned long corrupted_size;
+ unsigned long pfn;
+ int pagemap_fd;
+ int kpageflags_fd;
+ bool triggered;
+};
+
+FIXTURE_VARIANT(memory_failure)
+{
+ enum inject_type type;
+ int (*inject)(FIXTURE_DATA(memory_failure) * self, void *vaddr);
+};
+
+static int madv_hard_inject(FIXTURE_DATA(memory_failure) * self, void *vaddr)
+{
+ return madvise(vaddr, self->page_size, MADV_HWPOISON);
+}
+
+FIXTURE_VARIANT_ADD(memory_failure, madv_hard)
+{
+ .type = MADV_HARD,
+ .inject = madv_hard_inject,
+};
+
+static int madv_soft_inject(FIXTURE_DATA(memory_failure) * self, void *vaddr)
+{
+ return madvise(vaddr, self->page_size, MADV_SOFT_OFFLINE);
+}
+
+FIXTURE_VARIANT_ADD(memory_failure, madv_soft)
+{
+ .type = MADV_SOFT,
+ .inject = madv_soft_inject,
+};
+
+static void sigbus_action(int signo, siginfo_t *si, void *args)
+{
+ memcpy(&siginfo, si, sizeof(siginfo_t));
+ siglongjmp(signal_jmp_buf, 1);
+}
+
+static int setup_sighandler(void)
+{
+ struct sigaction sa = {
+ .sa_sigaction = sigbus_action,
+ .sa_flags = SA_SIGINFO,
+ };
+
+ return sigaction(SIGBUS, &sa, NULL);
+}
+
+FIXTURE_SETUP(memory_failure)
+{
+ memset(self, 0, sizeof(*self));
+
+ self->page_size = (unsigned long)sysconf(_SC_PAGESIZE);
+
+ memset(&siginfo, 0, sizeof(siginfo));
+ if (setup_sighandler())
+ SKIP(return, "setup sighandler failed.\n");
+
+ self->pagemap_fd = open(pagemap_proc, O_RDONLY);
+ if (self->pagemap_fd == -1)
+ SKIP(return, "open %s failed.\n", pagemap_proc);
+
+ self->kpageflags_fd = open(kpageflags_proc, O_RDONLY);
+ if (self->kpageflags_fd == -1)
+ SKIP(return, "open %s failed.\n", kpageflags_proc);
+}
+
+static void teardown_sighandler(void)
+{
+ struct sigaction sa = {
+ .sa_handler = SIG_DFL,
+ .sa_flags = SA_SIGINFO,
+ };
+
+ sigaction(SIGBUS, &sa, NULL);
+}
+
+FIXTURE_TEARDOWN(memory_failure)
+{
+ close(self->kpageflags_fd);
+ close(self->pagemap_fd);
+ teardown_sighandler();
+}
+
+static void prepare(struct __test_metadata *_metadata, FIXTURE_DATA(memory_failure) * self,
+ void *vaddr)
+{
+ self->pfn = pagemap_get_pfn(self->pagemap_fd, vaddr);
+ ASSERT_NE(self->pfn, -1UL);
+
+ ASSERT_EQ(get_hardware_corrupted_size(&self->corrupted_size), 0);
+}
+
+static bool check_memory(void *vaddr, unsigned long size)
+{
+ char buf[64];
+
+ memset(buf, 0xce, sizeof(buf));
+ while (size >= sizeof(buf)) {
+ if (memcmp(vaddr, buf, sizeof(buf)))
+ return false;
+ size -= sizeof(buf);
+ vaddr += sizeof(buf);
+ }
+
+ return true;
+}
+
+static void check(struct __test_metadata *_metadata, FIXTURE_DATA(memory_failure) * self,
+ void *vaddr, enum result_type type, int setjmp)
+{
+ unsigned long size;
+ uint64_t pfn_flags;
+
+ switch (type) {
+ case MADV_SOFT_ANON:
+ /* It is not expected to receive a SIGBUS signal. */
+ ASSERT_EQ(setjmp, 0);
+
+ /* The page content should remain unchanged. */
+ ASSERT_TRUE(check_memory(vaddr, self->page_size));
+
+ /* The backing pfn of addr should have changed. */
+ ASSERT_NE(pagemap_get_pfn(self->pagemap_fd, vaddr), self->pfn);
+ break;
+ case MADV_HARD_ANON:
+ /* The SIGBUS signal should have been received. */
+ ASSERT_EQ(setjmp, 1);
+
+ /* Check if siginfo contains correct SIGBUS context. */
+ ASSERT_EQ(siginfo.si_signo, SIGBUS);
+ ASSERT_EQ(siginfo.si_code, BUS_MCEERR_AR);
+ ASSERT_EQ(1UL << siginfo.si_addr_lsb, self->page_size);
+ ASSERT_EQ(siginfo.si_addr, vaddr);
+
+ /* XXX Check backing pte is hwpoison entry when supported. */
+ ASSERT_TRUE(pagemap_is_swapped(self->pagemap_fd, vaddr));
+ break;
+ default:
+ SKIP(return, "unexpected inject type %d.\n", type);
+ }
+
+ /* Check if the value of HardwareCorrupted has increased. */
+ ASSERT_EQ(get_hardware_corrupted_size(&size), 0);
+ ASSERT_EQ(size, self->corrupted_size + self->page_size / 1024);
+
+ /* Check if HWPoison flag is set. */
+ ASSERT_EQ(pageflags_get(self->pfn, self->kpageflags_fd, &pfn_flags), 0);
+ ASSERT_EQ(pfn_flags & KPF_HWPOISON, KPF_HWPOISON);
+}
+
+static void cleanup(struct __test_metadata *_metadata, FIXTURE_DATA(memory_failure) * self,
+ void *vaddr)
+{
+ unsigned long size;
+ uint64_t pfn_flags;
+
+ ASSERT_EQ(unpoison_memory(self->pfn), 0);
+
+ /* Check if HWPoison flag is cleared. */
+ ASSERT_EQ(pageflags_get(self->pfn, self->kpageflags_fd, &pfn_flags), 0);
+ ASSERT_NE(pfn_flags & KPF_HWPOISON, KPF_HWPOISON);
+
+ /* Check if the value of HardwareCorrupted has decreased. */
+ ASSERT_EQ(get_hardware_corrupted_size(&size), 0);
+ ASSERT_EQ(size, self->corrupted_size);
+}
+
+TEST_F(memory_failure, anon)
+{
+ char *addr;
+ int ret;
+
+ addr = mmap(0, self->page_size, PROT_READ | PROT_WRITE,
+ MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
+ if (addr == MAP_FAILED)
+ SKIP(return, "mmap failed, not enough memory.\n");
+ memset(addr, 0xce, self->page_size);
+
+ prepare(_metadata, self, addr);
+
+ ret = sigsetjmp(signal_jmp_buf, 1);
+ if (!self->triggered) {
+ self->triggered = true;
+ ASSERT_EQ(variant->inject(self, addr), 0);
+ FORCE_READ(*addr);
+ }
+
+ if (variant->type == MADV_HARD)
+ check(_metadata, self, addr, MADV_HARD_ANON, ret);
+ else
+ check(_metadata, self, addr, MADV_SOFT_ANON, ret);
+
+ cleanup(_metadata, self, addr);
+
+ ASSERT_EQ(munmap(addr, self->page_size), 0);
+}
+
+TEST_HARNESS_MAIN
diff --git a/tools/testing/selftests/mm/run_vmtests.sh b/tools/testing/selftests/mm/run_vmtests.sh
index 2dadbfc6e535..043ac00a98b8 100755
--- a/tools/testing/selftests/mm/run_vmtests.sh
+++ b/tools/testing/selftests/mm/run_vmtests.sh
@@ -87,6 +87,8 @@ separated by spaces:
test VMA merge cases behave as expected
- rmap
test rmap behaves as expected
+- memory-failure
+ test memory-failure behaves as expected
example: ./run_vmtests.sh -t "hmm mmap ksm"
EOF
@@ -535,6 +537,25 @@ CATEGORY="page_frag" run_test ./test_page_frag.sh nonaligned
CATEGORY="rmap" run_test ./rmap
+# Try to load hwpoison_inject if not present.
+HWPOISON_DIR=/sys/kernel/debug/hwpoison/
+if [ ! -d "$HWPOISON_DIR" ]; then
+ if ! modprobe -q -R hwpoison_inject; then
+ echo "Module hwpoison_inject not found, skipping..."
+ else
+ modprobe hwpoison_inject > /dev/null 2>&1
+ LOADED_MOD=1
+ fi
+fi
+
+if [ -d "$HWPOISON_DIR" ]; then
+ CATEGORY="memory-failure" run_test ./memory-failure
+fi
+
+if [ -n "${LOADED_MOD}" ]; then
+ modprobe -r hwpoison_inject > /dev/null 2>&1
+fi
+
if [ "${HAVE_HUGEPAGES}" = 1 ]; then
echo "$orig_nr_hugepgs" > /proc/sys/vm/nr_hugepages
fi
diff --git a/tools/testing/selftests/mm/vm_util.c b/tools/testing/selftests/mm/vm_util.c
index d954bf91afd5..a6d4ff7dfdc0 100644
--- a/tools/testing/selftests/mm/vm_util.c
+++ b/tools/testing/selftests/mm/vm_util.c
@@ -723,3 +723,44 @@ int ksm_stop(void)
close(ksm_fd);
return ret == 1 ? 0 : -errno;
}
+
+int get_hardware_corrupted_size(unsigned long *val)
+{
+ unsigned long size;
+ char *line = NULL;
+ size_t linelen = 0;
+ FILE *f = fopen("/proc/meminfo", "r");
+ int ret = -1;
+
+ if (!f)
+ return ret;
+
+ while (getline(&line, &linelen, f) > 0) {
+ if (sscanf(line, "HardwareCorrupted: %12lu kB", &size) == 1) {
+ *val = size;
+ ret = 0;
+ break;
+ }
+ }
+
+ free(line);
+ fclose(f);
+ return ret;
+}
+
+int unpoison_memory(unsigned long pfn)
+{
+ int unpoison_fd, len;
+ char buf[32];
+ ssize_t ret;
+
+ unpoison_fd = open("/sys/kernel/debug/hwpoison/unpoison-pfn", O_WRONLY);
+ if (unpoison_fd < 0)
+ return -errno;
+
+ len = sprintf(buf, "0x%lx\n", pfn);
+ ret = write(unpoison_fd, buf, len);
+ close(unpoison_fd);
+
+ return ret > 0 ? 0 : -errno;
+}
diff --git a/tools/testing/selftests/mm/vm_util.h b/tools/testing/selftests/mm/vm_util.h
index 6ad32b1830f1..b0c4ea994082 100644
--- a/tools/testing/selftests/mm/vm_util.h
+++ b/tools/testing/selftests/mm/vm_util.h
@@ -20,6 +20,7 @@
#define KPF_COMPOUND_HEAD BIT_ULL(15)
#define KPF_COMPOUND_TAIL BIT_ULL(16)
+#define KPF_HWPOISON BIT_ULL(19)
#define KPF_THP BIT_ULL(22)
/*
* Ignore the checkpatch warning, we must read from x but don't want to do
@@ -147,6 +148,8 @@ long ksm_get_full_scans(void);
int ksm_use_zero_pages(void);
int ksm_start(void);
int ksm_stop(void);
+int get_hardware_corrupted_size(unsigned long *val);
+int unpoison_memory(unsigned long pfn);
/*
* On ppc64 this will only work with radix 2M hugepage size
--
2.33.0
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 2/3] selftests/mm: add memory failure clean pagecache test
2026-01-07 9:37 [PATCH 0/3] selftests/mm: add memory failure selftests Miaohe Lin
2026-01-07 9:37 ` [PATCH 1/3] selftests/mm: add memory failure anonymous page test Miaohe Lin
@ 2026-01-07 9:37 ` Miaohe Lin
2026-01-07 9:37 ` [PATCH 3/3] selftests/mm: add memory failure dirty " Miaohe Lin
2026-01-09 13:45 ` [PATCH 0/3] selftests/mm: add memory failure selftests David Hildenbrand (Red Hat)
3 siblings, 0 replies; 5+ messages in thread
From: Miaohe Lin @ 2026-01-07 9:37 UTC (permalink / raw)
To: akpm, shuah
Cc: david, lorenzo.stoakes, Liam.Howlett, vbabka, rppt, surenb,
mhocko, nao.horiguchi, linmiaohe, linux-mm, linux-kernel,
linux-kselftest
This patch adds a new testcase to validate memory failure handling for
clean pagecache. This test performs similar operations as anonymous
pages except allocating memory using mmap() with a file fd.
This test helps ensure that memory failure handling for clean pagecache
works correctly, including unchanged page content, page isolation, and
recovery paths.
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
tools/testing/selftests/mm/memory-failure.c | 46 +++++++++++++++++++++
1 file changed, 46 insertions(+)
diff --git a/tools/testing/selftests/mm/memory-failure.c b/tools/testing/selftests/mm/memory-failure.c
index 37806a58f4b4..41bc669cef60 100644
--- a/tools/testing/selftests/mm/memory-failure.c
+++ b/tools/testing/selftests/mm/memory-failure.c
@@ -10,6 +10,7 @@
#include <sys/mman.h>
#include <linux/mman.h>
#include <linux/string.h>
+#include <unistd.h>
#include <signal.h>
#include <setjmp.h>
#include <unistd.h>
@@ -24,7 +25,9 @@ enum inject_type {
enum result_type {
MADV_HARD_ANON,
+ MADV_HARD_CLEAN_PAGECACHE,
MADV_SOFT_ANON,
+ MADV_SOFT_CLEAN_PAGECACHE,
};
static jmp_buf signal_jmp_buf;
@@ -154,6 +157,8 @@ static void check(struct __test_metadata *_metadata, FIXTURE_DATA(memory_failure
switch (type) {
case MADV_SOFT_ANON:
+ case MADV_HARD_CLEAN_PAGECACHE:
+ case MADV_SOFT_CLEAN_PAGECACHE:
/* It is not expected to receive a SIGBUS signal. */
ASSERT_EQ(setjmp, 0);
@@ -236,4 +241,45 @@ TEST_F(memory_failure, anon)
ASSERT_EQ(munmap(addr, self->page_size), 0);
}
+TEST_F(memory_failure, clean_pagecache)
+{
+ const char *fname = "./clean-page-cache-test-file";
+ int fd;
+ char *addr;
+ int ret;
+
+ fd = open(fname, O_RDWR | O_CREAT, 0664);
+ if (fd < 0)
+ SKIP(return, "failed to open test file.\n");
+ unlink(fname);
+ ftruncate(fd, self->page_size);
+
+ addr = mmap(0, self->page_size, PROT_READ | PROT_WRITE,
+ MAP_SHARED, fd, 0);
+ if (addr == MAP_FAILED)
+ SKIP(return, "mmap failed, not enough memory.\n");
+ memset(addr, 0xce, self->page_size);
+ fsync(fd);
+
+ prepare(_metadata, self, addr);
+
+ ret = sigsetjmp(signal_jmp_buf, 1);
+ if (!self->triggered) {
+ self->triggered = true;
+ ASSERT_EQ(variant->inject(self, addr), 0);
+ FORCE_READ(*addr);
+ }
+
+ if (variant->type == MADV_HARD)
+ check(_metadata, self, addr, MADV_HARD_CLEAN_PAGECACHE, ret);
+ else
+ check(_metadata, self, addr, MADV_SOFT_CLEAN_PAGECACHE, ret);
+
+ cleanup(_metadata, self, addr);
+
+ ASSERT_EQ(munmap(addr, self->page_size), 0);
+
+ ASSERT_EQ(close(fd), 0);
+}
+
TEST_HARNESS_MAIN
--
2.33.0
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 3/3] selftests/mm: add memory failure dirty pagecache test
2026-01-07 9:37 [PATCH 0/3] selftests/mm: add memory failure selftests Miaohe Lin
2026-01-07 9:37 ` [PATCH 1/3] selftests/mm: add memory failure anonymous page test Miaohe Lin
2026-01-07 9:37 ` [PATCH 2/3] selftests/mm: add memory failure clean pagecache test Miaohe Lin
@ 2026-01-07 9:37 ` Miaohe Lin
2026-01-09 13:45 ` [PATCH 0/3] selftests/mm: add memory failure selftests David Hildenbrand (Red Hat)
3 siblings, 0 replies; 5+ messages in thread
From: Miaohe Lin @ 2026-01-07 9:37 UTC (permalink / raw)
To: akpm, shuah
Cc: david, lorenzo.stoakes, Liam.Howlett, vbabka, rppt, surenb,
mhocko, nao.horiguchi, linmiaohe, linux-mm, linux-kernel,
linux-kselftest
This patch adds a new testcase to validate memory failure handling for
dirty pagecache. This performs similar operations as clean pagecaches
except fsync() is not used to keep pages dirty.
This test helps ensure that memory failure handling for dirty pagecache
works correctly, including proper SIGBUS delivery, page isolation, and
recovery paths.
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
tools/testing/selftests/mm/memory-failure.c | 58 +++++++++++++++++++--
1 file changed, 54 insertions(+), 4 deletions(-)
diff --git a/tools/testing/selftests/mm/memory-failure.c b/tools/testing/selftests/mm/memory-failure.c
index 41bc669cef60..275ab1c6de0c 100644
--- a/tools/testing/selftests/mm/memory-failure.c
+++ b/tools/testing/selftests/mm/memory-failure.c
@@ -26,8 +26,10 @@ enum inject_type {
enum result_type {
MADV_HARD_ANON,
MADV_HARD_CLEAN_PAGECACHE,
+ MADV_HARD_DIRTY_PAGECACHE,
MADV_SOFT_ANON,
MADV_SOFT_CLEAN_PAGECACHE,
+ MADV_SOFT_DIRTY_PAGECACHE,
};
static jmp_buf signal_jmp_buf;
@@ -159,6 +161,7 @@ static void check(struct __test_metadata *_metadata, FIXTURE_DATA(memory_failure
case MADV_SOFT_ANON:
case MADV_HARD_CLEAN_PAGECACHE:
case MADV_SOFT_CLEAN_PAGECACHE:
+ case MADV_SOFT_DIRTY_PAGECACHE:
/* It is not expected to receive a SIGBUS signal. */
ASSERT_EQ(setjmp, 0);
@@ -169,6 +172,7 @@ static void check(struct __test_metadata *_metadata, FIXTURE_DATA(memory_failure
ASSERT_NE(pagemap_get_pfn(self->pagemap_fd, vaddr), self->pfn);
break;
case MADV_HARD_ANON:
+ case MADV_HARD_DIRTY_PAGECACHE:
/* The SIGBUS signal should have been received. */
ASSERT_EQ(setjmp, 1);
@@ -241,18 +245,27 @@ TEST_F(memory_failure, anon)
ASSERT_EQ(munmap(addr, self->page_size), 0);
}
+static int prepare_file(const char *fname, unsigned long size)
+{
+ int fd;
+
+ fd = open(fname, O_RDWR | O_CREAT, 0664);
+ if (fd >= 0) {
+ unlink(fname);
+ ftruncate(fd, size);
+ }
+ return fd;
+}
+
TEST_F(memory_failure, clean_pagecache)
{
- const char *fname = "./clean-page-cache-test-file";
int fd;
char *addr;
int ret;
- fd = open(fname, O_RDWR | O_CREAT, 0664);
+ fd = prepare_file("./clean-page-cache-test-file", self->page_size);
if (fd < 0)
SKIP(return, "failed to open test file.\n");
- unlink(fname);
- ftruncate(fd, self->page_size);
addr = mmap(0, self->page_size, PROT_READ | PROT_WRITE,
MAP_SHARED, fd, 0);
@@ -282,4 +295,41 @@ TEST_F(memory_failure, clean_pagecache)
ASSERT_EQ(close(fd), 0);
}
+TEST_F(memory_failure, dirty_pagecache)
+{
+ int fd;
+ char *addr;
+ int ret;
+
+ fd = prepare_file("./dirty-page-cache-test-file", self->page_size);
+ if (fd < 0)
+ SKIP(return, "failed to open test file.\n");
+
+ addr = mmap(0, self->page_size, PROT_READ | PROT_WRITE,
+ MAP_SHARED, fd, 0);
+ if (addr == MAP_FAILED)
+ SKIP(return, "mmap failed, not enough memory.\n");
+ memset(addr, 0xce, self->page_size);
+
+ prepare(_metadata, self, addr);
+
+ ret = sigsetjmp(signal_jmp_buf, 1);
+ if (!self->triggered) {
+ self->triggered = true;
+ ASSERT_EQ(variant->inject(self, addr), 0);
+ FORCE_READ(*addr);
+ }
+
+ if (variant->type == MADV_HARD)
+ check(_metadata, self, addr, MADV_HARD_DIRTY_PAGECACHE, ret);
+ else
+ check(_metadata, self, addr, MADV_SOFT_DIRTY_PAGECACHE, ret);
+
+ cleanup(_metadata, self, addr);
+
+ ASSERT_EQ(munmap(addr, self->page_size), 0);
+
+ ASSERT_EQ(close(fd), 0);
+}
+
TEST_HARNESS_MAIN
--
2.33.0
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 0/3] selftests/mm: add memory failure selftests
2026-01-07 9:37 [PATCH 0/3] selftests/mm: add memory failure selftests Miaohe Lin
` (2 preceding siblings ...)
2026-01-07 9:37 ` [PATCH 3/3] selftests/mm: add memory failure dirty " Miaohe Lin
@ 2026-01-09 13:45 ` David Hildenbrand (Red Hat)
3 siblings, 0 replies; 5+ messages in thread
From: David Hildenbrand (Red Hat) @ 2026-01-09 13:45 UTC (permalink / raw)
To: Miaohe Lin, akpm, shuah
Cc: lorenzo.stoakes, Liam.Howlett, vbabka, rppt, surenb, mhocko,
nao.horiguchi, linux-mm, linux-kernel, linux-kselftest
On 1/7/26 10:37, Miaohe Lin wrote:
> Introduce selftests to validate the functionality of memory failure.
> These tests help ensure that memory failure handling for anonymous
> pages, pagecaches pages works correctly, including proper SIGBUS
> delivery to user processes, page isolation, and recovery paths.
>
> Currently madvise syscall is used to inject memory failures. And only
> anonymous pages and pagecaches are tested. More test scenarios, e.g.
> hugetlb, shmem, thp, will be added. Also more memory failure injecting
> methods will be supported, e.g. APEI Error INJection, if required.
0day reports that these tests fail:
# # ------------------------
# # running ./memory-failure
# # ------------------------
# # TAP version 13
# # 1..6
# # # Starting 6 tests from 2 test cases.
# # # RUN memory_failure.madv_hard.anon ...
# # # OK memory_failure.madv_hard.anon
# # ok 1 memory_failure.madv_hard.anon
# # # RUN memory_failure.madv_hard.clean_pagecache ...
# # # memory-failure.c:166:clean_pagecache:Expected setjmp (1) == 0 (0)
# # # clean_pagecache: Test terminated by assertion
# # # FAIL memory_failure.madv_hard.clean_pagecache
# # not ok 2 memory_failure.madv_hard.clean_pagecache
# # # RUN memory_failure.madv_hard.dirty_pagecache ...
# # # memory-failure.c:207:dirty_pagecache:Expected unpoison_memory(self->pfn) (-16) == 0 (0)
# # # dirty_pagecache: Test terminated by assertion
# # # FAIL memory_failure.madv_hard.dirty_pagecache
# # not ok 3 memory_failure.madv_hard.dirty_pagecache
# # # RUN memory_failure.madv_soft.anon ...
# # # OK memory_failure.madv_soft.anon
# # ok 4 memory_failure.madv_soft.anon
# # # RUN memory_failure.madv_soft.clean_pagecache ...
# # # memory-failure.c:282:clean_pagecache:Expected variant->inject(self, addr) (-1) == 0 (0)
# # # clean_pagecache: Test terminated by assertion
# # # FAIL memory_failure.madv_soft.clean_pagecache
# # not ok 5 memory_failure.madv_soft.clean_pagecache
# # # RUN memory_failure.madv_soft.dirty_pagecache ...
# # # memory-failure.c:319:dirty_pagecache:Expected variant->inject(self, addr) (-1) == 0 (0)
# # # dirty_pagecache: Test terminated by assertion
# # # FAIL memory_failure.madv_soft.dirty_pagecache
# # not ok 6 memory_failure.madv_soft.dirty_pagecache
# # # FAILED: 2 / 6 tests passed.
# # # Totals: pass:2 fail:4 xfail:0 xpass:0 skip:0 error:0
# # [FAIL]
# not ok 71 memory-failure # exit=1
Can the test maybe not deal with running in certain environments (config options etc)?
--
Cheers
David
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2026-01-09 13:45 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-01-07 9:37 [PATCH 0/3] selftests/mm: add memory failure selftests Miaohe Lin
2026-01-07 9:37 ` [PATCH 1/3] selftests/mm: add memory failure anonymous page test Miaohe Lin
2026-01-07 9:37 ` [PATCH 2/3] selftests/mm: add memory failure clean pagecache test Miaohe Lin
2026-01-07 9:37 ` [PATCH 3/3] selftests/mm: add memory failure dirty " Miaohe Lin
2026-01-09 13:45 ` [PATCH 0/3] selftests/mm: add memory failure selftests David Hildenbrand (Red Hat)
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox