* [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets
@ 2026-03-31 4:01 Li Wang
2026-03-31 4:01 ` [PATCH 2/4] selftests/mm: serialize local_config generation for parallel builds Li Wang
` (3 more replies)
0 siblings, 4 replies; 12+ messages in thread
From: Li Wang @ 2026-03-31 4:01 UTC (permalink / raw)
To: akpm, david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko,
shuah
Cc: linux-mm, linux-kselftest, linux-kernel
The 32-bit and 64-bit compilation rules invoke $(CC) directly, bypassing
the $(Q) quiet prefix and $(call msg,...) helper used by the rest of the
selftests build system. This causes these rules to always print the full
compiler command line, even when V=0 (the default).
Wrap the commands with $(Q) and $(call msg,CC,,$@) to match the
convention used by lib.mk, so that quiet and verbose builds behave
consistently across all targets.
==== Build logs ====
...
CC merge
CC rmap
CC soft-dirty
gcc -Wall -O2 -I /usr/src/25/tools/testing/selftests/../../..
-isystem /usr/src/25/tools/testing/selftests/../../../usr/include
-isystem /usr/src/25/tools/testing/selftests/../../../tools/include/uapi
-Wunreachable-code -U_FORTIFY_SOURCE -no-pie -D_GNU_SOURCE=
-I/usr/src/25/tools/testing/selftests/../../../tools/testing/selftests
-m32 -mxsave protection_keys.c vm_util.c thp_settings.c pkey_util.c
-lrt -lpthread -lm -lrt -ldl -lm
-o /usr/src/25/tools/testing/selftests/mm/protection_keys_32
gcc -Wall -O2 -I /usr/src/25/tools/testing/selftests/../../..
-isystem /usr/src/25/tools/testing/selftests/../../../usr/include
-isystem /usr/src/25/tools/testing/selftests/../../../tools/include/uapi
-Wunreachable-code -U_FORTIFY_SOURCE -no-pie -D_GNU_SOURCE=
-I/usr/src/25/tools/testing/selftests/../../../tools/testing/selftests
-m32 -mxsave pkey_sighandler_tests.c vm_util.c thp_settings.c pkey_util.c
-lrt -lpthread -lm -lrt -ldl -lm
-o /usr/src/25/tools/testing/selftests/mm/pkey_sighandler_tests_32
...
Reported-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Li Wang <liwang@redhat.com>
---
tools/testing/selftests/mm/Makefile | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/mm/Makefile b/tools/testing/selftests/mm/Makefile
index 7a5de4e9bf52..3b222cd6a048 100644
--- a/tools/testing/selftests/mm/Makefile
+++ b/tools/testing/selftests/mm/Makefile
@@ -215,7 +215,8 @@ ifeq ($(CAN_BUILD_I386),1)
$(BINARIES_32): CFLAGS += -m32 -mxsave
$(BINARIES_32): LDLIBS += -lrt -ldl -lm
$(BINARIES_32): $(OUTPUT)/%_32: %.c
- $(CC) $(CFLAGS) $(EXTRA_CFLAGS) $(notdir $^) $(LDLIBS) -o $@
+ $(call msg,CC,,$@)
+ $(Q)$(CC) $(CFLAGS) $(EXTRA_CFLAGS) $(notdir $^) $(LDLIBS) -o $@
$(foreach t,$(VMTARGETS),$(eval $(call gen-target-rule-32,$(t))))
endif
@@ -223,7 +224,8 @@ ifeq ($(CAN_BUILD_X86_64),1)
$(BINARIES_64): CFLAGS += -m64 -mxsave
$(BINARIES_64): LDLIBS += -lrt -ldl
$(BINARIES_64): $(OUTPUT)/%_64: %.c
- $(CC) $(CFLAGS) $(EXTRA_CFLAGS) $(notdir $^) $(LDLIBS) -o $@
+ $(call msg,CC,,$@)
+ $(Q)$(CC) $(CFLAGS) $(EXTRA_CFLAGS) $(notdir $^) $(LDLIBS) -o $@
$(foreach t,$(VMTARGETS),$(eval $(call gen-target-rule-64,$(t))))
endif
--
2.53.0
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH 2/4] selftests/mm: serialize local_config generation for parallel builds
2026-03-31 4:01 [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets Li Wang
@ 2026-03-31 4:01 ` Li Wang
2026-03-31 5:29 ` Andrew Morton
2026-03-31 4:01 ` [PATCH 3/4] selftests/lib.mk: set PWD from CURDIR to avoid wrong extmod path Li Wang
` (2 subsequent siblings)
3 siblings, 1 reply; 12+ messages in thread
From: Li Wang @ 2026-03-31 4:01 UTC (permalink / raw)
To: akpm, david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko,
shuah
Cc: linux-mm, linux-kselftest, linux-kernel
The mm selftests generate both local_config.mk and local_config.h from
check_config.sh. With high parallelism (-jN), this can race and lead to
incomplete target builds (e.g. only a few binaries get built after
'make -j100', while 'make -j1' builds everything).
Switch to a stamp-based dependency:
local_config.stamp: check_config.sh
... run check_config.sh ...
touch local_config.stamp
and make local_config.mk/local_config.h depend on the stamp.
This ensures check_config.sh is executed once per update decision and
removes the parallel race window. Also hook local_config.stamp into
EXTRA_CLEAN.
No functional change intended for non-parallel builds.
Reported-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Li Wang <liwang@redhat.com>
---
tools/testing/selftests/mm/Makefile | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/tools/testing/selftests/mm/Makefile b/tools/testing/selftests/mm/Makefile
index 3b222cd6a048..78496f705386 100644
--- a/tools/testing/selftests/mm/Makefile
+++ b/tools/testing/selftests/mm/Makefile
@@ -261,10 +261,14 @@ $(OUTPUT)/migration: LDLIBS += -lnuma
$(OUTPUT)/rmap: LDLIBS += -lnuma
-local_config.mk local_config.h: check_config.sh
- CC="$(CC)" CFLAGS="$(CFLAGS)" ./check_config.sh
+local_config.stamp: check_config.sh
+ $(call msg,CHK,config,$@)
+ $(Q)CC="$(CC)" CFLAGS="$(CFLAGS)" ./check_config.sh
+ $(Q)touch $@
-EXTRA_CLEAN += local_config.mk local_config.h
+local_config.mk local_config.h: local_config.stamp
+
+EXTRA_CLEAN += local_config.mk local_config.h local_config.stamp
ifeq ($(IOURING_EXTRA_LIBS),)
all: warn_missing_liburing
--
2.53.0
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH 3/4] selftests/lib.mk: set PWD from CURDIR to avoid wrong extmod path
2026-03-31 4:01 [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets Li Wang
2026-03-31 4:01 ` [PATCH 2/4] selftests/mm: serialize local_config generation for parallel builds Li Wang
@ 2026-03-31 4:01 ` Li Wang
2026-03-31 5:30 ` Andrew Morton
2026-03-31 4:01 ` [PATCH 4/4] selftests/mm: fix cow/gup_longterm link failures when liburing flags are missing Li Wang
2026-03-31 5:29 ` [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets Andrew Morton
3 siblings, 1 reply; 12+ messages in thread
From: Li Wang @ 2026-03-31 4:01 UTC (permalink / raw)
To: akpm, david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko,
shuah
Cc: linux-mm, linux-kselftest, linux-kernel
When running selftests from the kernel top-level (e.g. make kselftest-all),
sub-makes might still inherit the caller's PWD from the environment.
Some selftests use $(PWD) in recursive kbuild invocations, which can
then incorrectly resolve to the kernel top directory instead of the
current test directory.
In that case, kbuild may generate an external-module wrapper Makefile in
the wrong location, potentially clobbering the top-level Makefile and
causing recursive include failures ("Too many open files").
Export PWD := $(CURDIR) in selftests/lib.mk so $(PWD) always matches the
actual current selftest directory.
Reported-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Li Wang <liwang@redhat.com>
---
tools/testing/selftests/lib.mk | 2 ++
tools/testing/selftests/mm/local_config.stamp | 0
2 files changed, 2 insertions(+)
create mode 100644 tools/testing/selftests/mm/local_config.stamp
diff --git a/tools/testing/selftests/lib.mk b/tools/testing/selftests/lib.mk
index f02cc8a2e4ae..2840e98c225c 100644
--- a/tools/testing/selftests/lib.mk
+++ b/tools/testing/selftests/lib.mk
@@ -60,6 +60,8 @@ endif
selfdir = $(realpath $(dir $(filter %/lib.mk,$(MAKEFILE_LIST))))
top_srcdir = $(selfdir)/../../..
+export PWD := $(CURDIR)
+
# msg: emit succinct information message describing current building step
# $1 - generic step name (e.g., CC, LINK, etc);
# $2 - optional "flavor" specifier; if provided, will be emitted as [flavor];
diff --git a/tools/testing/selftests/mm/local_config.stamp b/tools/testing/selftests/mm/local_config.stamp
new file mode 100644
index 000000000000..e69de29bb2d1
--
2.53.0
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH 4/4] selftests/mm: fix cow/gup_longterm link failures when liburing flags are missing
2026-03-31 4:01 [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets Li Wang
2026-03-31 4:01 ` [PATCH 2/4] selftests/mm: serialize local_config generation for parallel builds Li Wang
2026-03-31 4:01 ` [PATCH 3/4] selftests/lib.mk: set PWD from CURDIR to avoid wrong extmod path Li Wang
@ 2026-03-31 4:01 ` Li Wang
2026-03-31 7:13 ` Li Wang
2026-03-31 5:29 ` [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets Andrew Morton
3 siblings, 1 reply; 12+ messages in thread
From: Li Wang @ 2026-03-31 4:01 UTC (permalink / raw)
To: akpm, david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko,
shuah
Cc: linux-mm, linux-kselftest, linux-kernel
mm selftests rely on IOURING_EXTRA_LIBS (typically from local_config.mk)
to link io_uring-based tests (cow and gup_longterm). On some systems
with liburing installed, IOURING_EXTRA_LIBS can still be empty, causing
link failures with unresolved io_uring symbols.
Add a fallback detection:
IOURING_EXTRA_LIBS := $(shell pkg-config --libs liburing)
when IOURING_EXTRA_LIBS was not set by local_config.mk.
Also gate io_uring-dependent test binaries (cow and gup_longterm) on
IOURING_EXTRA_LIBS being non-empty, so missing liburing support cleanly
skips those tests instead of failing the build.
Signed-off-by: Li Wang <liwang@redhat.com>
---
tools/testing/selftests/mm/Makefile | 13 ++++++++++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/tools/testing/selftests/mm/Makefile b/tools/testing/selftests/mm/Makefile
index 78496f705386..01bdd25e04e4 100644
--- a/tools/testing/selftests/mm/Makefile
+++ b/tools/testing/selftests/mm/Makefile
@@ -10,6 +10,11 @@ LOCAL_HDRS += $(selfdir)/mm/mseal_helpers.h
include local_config.mk
+PKG_CONFIG ?= pkg-config
+ifeq ($(IOURING_EXTRA_LIBS),)
+IOURING_EXTRA_LIBS := $(shell $(PKG_CONFIG) --libs liburing 2>/dev/null)
+endif
+
ifeq ($(ARCH),)
ifeq ($(CROSS_COMPILE),)
@@ -55,10 +60,12 @@ else
PAGE_FRAG_WARNING = "missing Module.symvers, please have the kernel built first"
endif
-TEST_GEN_FILES = cow
-TEST_GEN_FILES += compaction_test
-TEST_GEN_FILES += gup_longterm
+TEST_GEN_FILES = compaction_test
TEST_GEN_FILES += gup_test
+ifneq ($(IOURING_EXTRA_LIBS),)
+TEST_GEN_FILES += cow
+TEST_GEN_FILES += gup_longterm
+endif
TEST_GEN_FILES += hmm-tests
TEST_GEN_FILES += hugetlb-madvise
TEST_GEN_FILES += hugetlb-read-hwpoison
--
2.53.0
^ permalink raw reply related [flat|nested] 12+ messages in thread
* Re: [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets
2026-03-31 4:01 [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets Li Wang
` (2 preceding siblings ...)
2026-03-31 4:01 ` [PATCH 4/4] selftests/mm: fix cow/gup_longterm link failures when liburing flags are missing Li Wang
@ 2026-03-31 5:29 ` Andrew Morton
2026-03-31 9:50 ` Li Wang
3 siblings, 1 reply; 12+ messages in thread
From: Andrew Morton @ 2026-03-31 5:29 UTC (permalink / raw)
To: Li Wang
Cc: david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko, shuah,
linux-mm, linux-kselftest, linux-kernel
On Tue, 31 Mar 2026 12:01:53 +0800 Li Wang <liwang@redhat.com> wrote:
> The 32-bit and 64-bit compilation rules invoke $(CC) directly, bypassing
> the $(Q) quiet prefix and $(call msg,...) helper used by the rest of the
> selftests build system. This causes these rules to always print the full
> compiler command line, even when V=0 (the default).
>
> Wrap the commands with $(Q) and $(call msg,CC,,$@) to match the
> convention used by lib.mk, so that quiet and verbose builds behave
> consistently across all targets.
>
> ==== Build logs ====
> ...
> CC merge
> CC rmap
> CC soft-dirty
> gcc -Wall -O2 -I /usr/src/25/tools/testing/selftests/../../..
> -isystem /usr/src/25/tools/testing/selftests/../../../usr/include
> -isystem /usr/src/25/tools/testing/selftests/../../../tools/include/uapi
> -Wunreachable-code -U_FORTIFY_SOURCE -no-pie -D_GNU_SOURCE=
> -I/usr/src/25/tools/testing/selftests/../../../tools/testing/selftests
> -m32 -mxsave protection_keys.c vm_util.c thp_settings.c pkey_util.c
> -lrt -lpthread -lm -lrt -ldl -lm
> -o /usr/src/25/tools/testing/selftests/mm/protection_keys_32
> gcc -Wall -O2 -I /usr/src/25/tools/testing/selftests/../../..
> -isystem /usr/src/25/tools/testing/selftests/../../../usr/include
> -isystem /usr/src/25/tools/testing/selftests/../../../tools/include/uapi
> -Wunreachable-code -U_FORTIFY_SOURCE -no-pie -D_GNU_SOURCE=
> -I/usr/src/25/tools/testing/selftests/../../../tools/testing/selftests
> -m32 -mxsave pkey_sighandler_tests.c vm_util.c thp_settings.c pkey_util.c
> -lrt -lpthread -lm -lrt -ldl -lm
> -o /usr/src/25/tools/testing/selftests/mm/pkey_sighandler_tests_32
> ...
>
> Reported-by: Andrew Morton <akpm@linux-foundation.org>
wow, thanks, do you want my credit card number?
> tools/testing/selftests/mm/Makefile | 6 ++++--
Before:
ts:/usr/src/25/tools/testing/selftests/mm> make -j100
CC="gcc" CFLAGS="-Wall -O2 -I /usr/src/25/tools/testing/selftests/../../.. -isystem /usr/src/25/tools/testing/selftests/../../../usr/include -isystem /usr/src/25/tools/testing/selftests/../../../tools/include/uapi -Wunreachable-code -U_FORTIFY_SOURCE -no-pie -D_GNU_SOURCE= -I/usr/src/25/tools/testing/selftests/../../../tools/testing/selftests " ./check_config.sh
CC cow
CC compaction_test
CC gup_longterm
CC gup_test
CC hmm-tests
CC hugetlb-madvise
CC hugetlb-read-hwpoison
CC hugetlb-soft-offline
CC hugepage-mmap
CC hugepage-mremap
CC hugepage-shm
CC hugepage-vmemmap
CC khugepaged
CC madv_populate
CC map_fixed_noreplace
CC map_populate
CC map_hugetlb
CC memfd_secret
CC memory-failure
CC migration
CC mkdirty
CC mlock-random-test
CC mlock2-tests
CC mremap_dontunmap
CC mrelease_test
...
CC merge
CC rmap
gcc -Wall -O2 -I /usr/src/25/tools/testing/selftests/../../.. -isystem /usr/src/25/tools/testing/selftests/../../../usr/include -isystem /usr/src/25/tools/testing/selftests/../../../tools/include/uapi -Wunreachable-code -U_FORTIFY_SOURCE -no-pie -D_GNU_SOURCE= -I/usr/src/25/tools/testing/selftests/../../../tools/testing/selftests -m32 -mxsave protection_keys.c vm_util.c thp_settings.c pkey_util.c -lrt -lpthread -lm -lrt -ldl -lm -o /usr/src/25/tools/testing/selftests/mm/protection_keys_32
CC folio_split_race_test
CC soft-dirty
gcc -Wall -O2 -I /usr/src/25/tools/testing/selftests/../../.. -isystem /usr/src/25/tools/testing/selftests/../../../usr/include -isystem /usr/src/25/tools/testing/selftests/../../../tools/include/uapi -Wunreachable-code -U_FORTIFY_SOURCE -no-pie -D_GNU_SOURCE= -I/usr/src/25/tools/testing/selftests/../../../tools/testing/selftests -m32 -mxsave pkey_sighandler_tests.c vm_util.c thp_settings.c pkey_util.c -lrt -lpthread -lm -lrt -ldl -lm -o /usr/src/25/tools/testing/selftests/mm/pkey_sighandler_tests_32
gcc -Wall -O2 -I /usr/src/25/tools/testing/selftests/../../.. -isystem /usr/src/25/tools/testing/selftests/../../../usr/include -isystem /usr/src/25/tools/testing/selftests/../../../tools/include/uapi -Wunreachable-code -U_FORTIFY_SOURCE -no-pie -D_GNU_SOURCE= -I/usr/src/25/tools/testing/selftests/../../../tools/testing/selftests -m64 -mxsave protection_keys.c vm_util.c thp_settings.c pkey_util.c -lrt -lpthread -lm -lrt -ldl -o /usr/src/25/tools/testing/selftests/mm/protection_keys_64
gcc -Wall -O2 -I /usr/src/25/tools/testing/selftests/../../.. -isystem /usr/src/25/tools/testing/selftests/../../../usr/include -isystem /usr/src/25/tools/testing/selftests/../../../tools/include/uapi -Wunreachable-code -U_FORTIFY_SOURCE -no-pie -D_GNU_SOURCE= -I/usr/src/25/tools/testing/selftests/../../../tools/testing/selftests -m64 -mxsave pkey_sighandler_tests.c vm_util.c thp_settings.c pkey_util.c -lrt -lpthread -lm -lrt -ldl -o /usr/src/25/tools/testing/selftests/mm/pkey_sighandler_tests_64
CC write_to_hugetlbfs
after:
ts:/usr/src/25/tools/testing/selftests/mm> make clean
...
ts:/usr/src/25/tools/testing/selftests/mm> make -j100
CC="gcc" CFLAGS="-Wall -O2 -I /usr/src/25/tools/testing/selftests/../../.. -isystem /usr/src/25/tools/testing/selftests/../../../usr/include -isystem /usr/src/25/tools/testing/selftests/../../../tools/include/uapi -Wunreachable-code -U_FORTIFY_SOURCE -no-pie -D_GNU_SOURCE= -I/usr/src/25/tools/testing/selftests/../../../tools/testing/selftests " ./check_config.sh
CC cow
CC compaction_test
CC gup_longterm
CC gup_test
CC hmm-tests
CC hugetlb-madvise
CC hugetlb-read-hwpoison
CC hugetlb-soft-offline
CC hugepage-mmap
CC hugepage-mremap
CC hugepage-shm
CC hugepage-vmemmap
CC khugepaged
CC madv_populate
CC map_fixed_noreplace
CC map_hugetlb
CC map_populate
CC memfd_secret
CC memory-failure
CC migration
CC mkdirty
CC mlock-random-test
CC mlock2-tests
CC mrelease_test
CC mremap_dontunmap
CC mremap_test
CC mseal_test
CC on-fault-limit
CC pagemap_ioctl
CC pfnmap
CC process_madv
CC prctl_thp_disable
CC thuge-gen
CC transhuge-stress
CC uffd-stress
CC uffd-unit-tests
CC uffd-wp-mremap
CC split_huge_page_test
CC ksm_tests
CC ksm_functional_tests
CC mdwe_test
CC hugetlb_fault_after_madv
CC hugetlb_madv_vs_map
CC hugetlb_dio
CC droppable
CC guard-regions
CC merge
CC rmap
CC folio_split_race_test
CC soft-dirty
CC protection_keys_32
CC pkey_sighandler_tests_32
CC protection_keys_64
CC pkey_sighandler_tests_64
CC va_high_addr_switch
CC write_to_hugetlbfs
Tested-by: Andrew Morton <akpm@linux-foundation.org>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 2/4] selftests/mm: serialize local_config generation for parallel builds
2026-03-31 4:01 ` [PATCH 2/4] selftests/mm: serialize local_config generation for parallel builds Li Wang
@ 2026-03-31 5:29 ` Andrew Morton
2026-03-31 9:16 ` Li Wang
0 siblings, 1 reply; 12+ messages in thread
From: Andrew Morton @ 2026-03-31 5:29 UTC (permalink / raw)
To: Li Wang
Cc: david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko, shuah,
linux-mm, linux-kselftest, linux-kernel
On Tue, 31 Mar 2026 12:01:54 +0800 Li Wang <liwang@redhat.com> wrote:
> The mm selftests generate both local_config.mk and local_config.h from
> check_config.sh. With high parallelism (-jN), this can race and lead to
> incomplete target builds (e.g. only a few binaries get built after
> 'make -j100', while 'make -j1' builds everything).
>
> Switch to a stamp-based dependency:
>
> local_config.stamp: check_config.sh
> ... run check_config.sh ...
> touch local_config.stamp
>
> and make local_config.mk/local_config.h depend on the stamp.
>
> This ensures check_config.sh is executed once per update decision and
> removes the parallel race window. Also hook local_config.stamp into
> EXTRA_CLEAN.
>
> No functional change intended for non-parallel builds.
>
> Reported-by: Andrew Morton <akpm@linux-foundation.org>
Well, the wind changed direction. I saw this race a single time today
but now I'm trying to produce a reliable before-and-after report, it
isn't cooperating - everything is now building with -j100 with and
without this patch.
So not very helpful, sorry.
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 3/4] selftests/lib.mk: set PWD from CURDIR to avoid wrong extmod path
2026-03-31 4:01 ` [PATCH 3/4] selftests/lib.mk: set PWD from CURDIR to avoid wrong extmod path Li Wang
@ 2026-03-31 5:30 ` Andrew Morton
2026-03-31 7:19 ` Li Wang
0 siblings, 1 reply; 12+ messages in thread
From: Andrew Morton @ 2026-03-31 5:30 UTC (permalink / raw)
To: Li Wang
Cc: david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko, shuah,
linux-mm, linux-kselftest, linux-kernel
On Tue, 31 Mar 2026 12:01:55 +0800 Li Wang <liwang@redhat.com> wrote:
> When running selftests from the kernel top-level (e.g. make kselftest-all),
> sub-makes might still inherit the caller's PWD from the environment.
>
> Some selftests use $(PWD) in recursive kbuild invocations, which can
> then incorrectly resolve to the kernel top directory instead of the
> current test directory.
>
> In that case, kbuild may generate an external-module wrapper Makefile in
> the wrong location, potentially clobbering the top-level Makefile and
> causing recursive include failures ("Too many open files").
>
> Export PWD := $(CURDIR) in selftests/lib.mk so $(PWD) always matches the
> actual current selftest directory.
>
> Reported-by: Andrew Morton <akpm@linux-foundation.org>
> Signed-off-by: Li Wang <liwang@redhat.com>
I couldn't clearly figure out what I was doing to cause kbuild to
destroy my top-level Makefile. It happened three times and it wasn't
fun so I Stopped Doing That.
I tried a bunch of things *without* this patch and of course, Makefile
is still intact. So I cannot confirm or deny, sorry.
I've switched my script so I'll henceforth be running -j100, shall keep
an eye on things.
I have no comment on [4/4] - it addresses something I haven't seen.
I'll assume this patchset will be handled via the kbuild tree.
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 4/4] selftests/mm: fix cow/gup_longterm link failures when liburing flags are missing
2026-03-31 4:01 ` [PATCH 4/4] selftests/mm: fix cow/gup_longterm link failures when liburing flags are missing Li Wang
@ 2026-03-31 7:13 ` Li Wang
2026-04-01 13:44 ` David Hildenbrand (Arm)
0 siblings, 1 reply; 12+ messages in thread
From: Li Wang @ 2026-03-31 7:13 UTC (permalink / raw)
To: akpm, david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko,
shuah
Cc: linux-mm, linux-kselftest, linux-kernel
> --- a/tools/testing/selftests/mm/Makefile
> +++ b/tools/testing/selftests/mm/Makefile
> @@ -10,6 +10,11 @@ LOCAL_HDRS += $(selfdir)/mm/mseal_helpers.h
>
> include local_config.mk
>
> +PKG_CONFIG ?= pkg-config
> +ifeq ($(IOURING_EXTRA_LIBS),)
> +IOURING_EXTRA_LIBS := $(shell $(PKG_CONFIG) --libs liburing 2>/dev/null)
> +endif
> +
> ifeq ($(ARCH),)
>
> ifeq ($(CROSS_COMPILE),)
> @@ -55,10 +60,12 @@ else
> PAGE_FRAG_WARNING = "missing Module.symvers, please have the kernel built first"
> endif
>
> -TEST_GEN_FILES = cow
> -TEST_GEN_FILES += compaction_test
> -TEST_GEN_FILES += gup_longterm
> +TEST_GEN_FILES = compaction_test
> TEST_GEN_FILES += gup_test
> +ifneq ($(IOURING_EXTRA_LIBS),)
> +TEST_GEN_FILES += cow
> +TEST_GEN_FILES += gup_longterm
> +endif
Please ignore this one, as Sashiko points:
"The io_uring specific test cases within them are already isolated via
#ifdef LOCAL_CONFIG_HAVE_LIBURING and do not require liburing to compile the
rest of the tests."
I overlooked that two tests can be built without liburing-devel pkg.
Sashiko is correct here, we don't need define IOURING_EXTRA_LIBS.
The compiling error comes from a temp file in check_config.sh, and I
am not sure whether we should hide the error:
/tmp/tmp.kIIOIqwe3n.c:2:10: fatal error: liburing.h: No such file or directory
2 | #include <liburing.h>
| ^~~~~~~~~~~~
--
Regards,
Li Wang
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 3/4] selftests/lib.mk: set PWD from CURDIR to avoid wrong extmod path
2026-03-31 5:30 ` Andrew Morton
@ 2026-03-31 7:19 ` Li Wang
0 siblings, 0 replies; 12+ messages in thread
From: Li Wang @ 2026-03-31 7:19 UTC (permalink / raw)
To: Andrew Morton
Cc: david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko, shuah,
linux-mm, linux-kselftest, linux-kernel
On Mon, Mar 30, 2026 at 10:30:25PM -0700, Andrew Morton wrote:
> On Tue, 31 Mar 2026 12:01:55 +0800 Li Wang <liwang@redhat.com> wrote:
>
> > When running selftests from the kernel top-level (e.g. make kselftest-all),
> > sub-makes might still inherit the caller's PWD from the environment.
> >
> > Some selftests use $(PWD) in recursive kbuild invocations, which can
> > then incorrectly resolve to the kernel top directory instead of the
> > current test directory.
> >
> > In that case, kbuild may generate an external-module wrapper Makefile in
> > the wrong location, potentially clobbering the top-level Makefile and
> > causing recursive include failures ("Too many open files").
> >
> > Export PWD := $(CURDIR) in selftests/lib.mk so $(PWD) always matches the
> > actual current selftest directory.
> >
> > Reported-by: Andrew Morton <akpm@linux-foundation.org>
> > Signed-off-by: Li Wang <liwang@redhat.com>
>
> I couldn't clearly figure out what I was doing to cause kbuild to
> destroy my top-level Makefile. It happened three times and it wasn't
> fun so I Stopped Doing That.
>
> I tried a bunch of things *without* this patch and of course, Makefile
> is still intact. So I cannot confirm or deny, sorry.
>
> I've switched my script so I'll henceforth be running -j100, shall keep
> an eye on things.
No problem, actually I wasn't able to reproduce it, the patch just worked
out based on your description. We can kick out it from patchset unless
someone confirm it is needed.
--
Regards,
Li Wang
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 2/4] selftests/mm: serialize local_config generation for parallel builds
2026-03-31 5:29 ` Andrew Morton
@ 2026-03-31 9:16 ` Li Wang
0 siblings, 0 replies; 12+ messages in thread
From: Li Wang @ 2026-03-31 9:16 UTC (permalink / raw)
To: Andrew Morton
Cc: david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko, shuah,
linux-mm, linux-kselftest, linux-kernel
On Mon, Mar 30, 2026 at 10:29:48PM -0700, Andrew Morton wrote:
> On Tue, 31 Mar 2026 12:01:54 +0800 Li Wang <liwang@redhat.com> wrote:
>
> > The mm selftests generate both local_config.mk and local_config.h from
> > check_config.sh. With high parallelism (-jN), this can race and lead to
> > incomplete target builds (e.g. only a few binaries get built after
> > 'make -j100', while 'make -j1' builds everything).
> >
> > Switch to a stamp-based dependency:
> >
> > local_config.stamp: check_config.sh
> > ... run check_config.sh ...
> > touch local_config.stamp
> >
> > and make local_config.mk/local_config.h depend on the stamp.
> >
> > This ensures check_config.sh is executed once per update decision and
> > removes the parallel race window. Also hook local_config.stamp into
> > EXTRA_CLEAN.
> >
> > No functional change intended for non-parallel builds.
> >
> > Reported-by: Andrew Morton <akpm@linux-foundation.org>
>
> Well, the wind changed direction. I saw this race a single time today
> but now I'm trying to produce a reliable before-and-after report, it
> isn't cooperating - everything is now building with -j100 with and
> without this patch.
>
> So not very helpful, sorry.
No worries, we can hang this one as well.
Let me send to kbuild ML with the two valid fix (1/4, 4/4),
I will reformat them.
--
Regards,
Li Wang
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets
2026-03-31 5:29 ` [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets Andrew Morton
@ 2026-03-31 9:50 ` Li Wang
0 siblings, 0 replies; 12+ messages in thread
From: Li Wang @ 2026-03-31 9:50 UTC (permalink / raw)
To: Andrew Morton
Cc: david, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko, shuah,
linux-mm, linux-kselftest, linux-kernel, linux-kbuild
> > Reported-by: Andrew Morton <akpm@linux-foundation.org>
>
> wow, thanks, do you want my credit card number?
Only if you want me to add a Funded-by: tag to the commit message. :)
> Tested-by: Andrew Morton <akpm@linux-foundation.org>
Thank you for testing!
--
Regards,
Li Wang
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH 4/4] selftests/mm: fix cow/gup_longterm link failures when liburing flags are missing
2026-03-31 7:13 ` Li Wang
@ 2026-04-01 13:44 ` David Hildenbrand (Arm)
0 siblings, 0 replies; 12+ messages in thread
From: David Hildenbrand (Arm) @ 2026-04-01 13:44 UTC (permalink / raw)
To: Li Wang, akpm, ljs, Liam.Howlett, vbabka, rppt, surenb, mhocko,
shuah
Cc: linux-mm, linux-kselftest, linux-kernel
On 3/31/26 09:13, Li Wang wrote:
>> --- a/tools/testing/selftests/mm/Makefile
>> +++ b/tools/testing/selftests/mm/Makefile
>> @@ -10,6 +10,11 @@ LOCAL_HDRS += $(selfdir)/mm/mseal_helpers.h
>>
>> include local_config.mk
>>
>> +PKG_CONFIG ?= pkg-config
>> +ifeq ($(IOURING_EXTRA_LIBS),)
>> +IOURING_EXTRA_LIBS := $(shell $(PKG_CONFIG) --libs liburing 2>/dev/null)
>> +endif
>> +
>> ifeq ($(ARCH),)
>>
>> ifeq ($(CROSS_COMPILE),)
>> @@ -55,10 +60,12 @@ else
>> PAGE_FRAG_WARNING = "missing Module.symvers, please have the kernel built first"
>> endif
>>
>> -TEST_GEN_FILES = cow
>> -TEST_GEN_FILES += compaction_test
>> -TEST_GEN_FILES += gup_longterm
>> +TEST_GEN_FILES = compaction_test
>> TEST_GEN_FILES += gup_test
>> +ifneq ($(IOURING_EXTRA_LIBS),)
>> +TEST_GEN_FILES += cow
>> +TEST_GEN_FILES += gup_longterm
>> +endif
>
> Please ignore this one, as Sashiko points:
Right, this should be dropped.
--
Cheers,
David
^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2026-04-01 13:45 UTC | newest]
Thread overview: 12+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-03-31 4:01 [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets Li Wang
2026-03-31 4:01 ` [PATCH 2/4] selftests/mm: serialize local_config generation for parallel builds Li Wang
2026-03-31 5:29 ` Andrew Morton
2026-03-31 9:16 ` Li Wang
2026-03-31 4:01 ` [PATCH 3/4] selftests/lib.mk: set PWD from CURDIR to avoid wrong extmod path Li Wang
2026-03-31 5:30 ` Andrew Morton
2026-03-31 7:19 ` Li Wang
2026-03-31 4:01 ` [PATCH 4/4] selftests/mm: fix cow/gup_longterm link failures when liburing flags are missing Li Wang
2026-03-31 7:13 ` Li Wang
2026-04-01 13:44 ` David Hildenbrand (Arm)
2026-03-31 5:29 ` [PATCH 1/4] selftests/mm: respect build verbosity settings for 32/64-bit targets Andrew Morton
2026-03-31 9:50 ` Li Wang
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox