selftests/mm: use pattern matching in .gitignore

PATCH Superseded [email protected] 3 participants 6 messages May 11 – Aug 1

Subsystems (4)

Memory management Kselftest MM misc MM GUP

lore

Share & tools

lore

[email protected]

author

2026-05-11 17:34 UTC #0 lore raw

v3 cover selftests/mm: use pattern matching in .gitignore

From: Pratyush Mallick <[email protected]>

The current selftests/mm/.gitignore hardcodes each generated test binary by name, which requires manual updates every time a new test is added.

This series switches to a pattern-matching approach (similar to KVM selftests), ignoring everything by default and allowing specific source extensions. To accommodate this without tracking generated headers, local_config.h is renamed to local_config.h_gen.

Changelog since v2:
- Split the changes into two separate patches (header rename and .gitignore update)
- Added Suggested-by for David Hildenbrand for the header rename approach.
- Dropped David's and Yosry's Reviewed-by tags.
- Kept Lorenzo Stoakes's Reviewed-by and added Mike Rapoport's Acked-by.

Changelog since v1 (RFC):
- Renamed local_config.h to local_config.h_gen to avoid conflict with !*.h.
- Updated Makefile, check_config.sh, and affected .c files for the rename.
- Removed *.mod.c as it was unnecessary.

Pratyush Mallick (2):
selftests/mm: use pattern matching in .gitignore selftests/mm: rename local_config.h to local_config.h_gen

 tools/testing/selftests/mm/.gitignore      | 71 +++-------------------
 tools/testing/selftests/mm/Makefile        |  6 +-
 tools/testing/selftests/mm/check_config.sh |  2 +-
 tools/testing/selftests/mm/cow.c           |  2 +-
 tools/testing/selftests/mm/gup_longterm.c  |  2 +-
 5 files changed, 15 insertions(+), 68 deletions(-)

Signature · 1 line

2.54.0.563.g4f69b47b94-goog

[email protected]

author

2026-05-11 17:35 UTC #1 lore raw

PATCH v3 1/2 selftests/mm: use pattern matching in .gitignore

From: Pratyush Mallick <[email protected]>

The current .gitignore hardcodes each generated test binary by name, requiring updates every time a new test is added.

Switch to the patten-matching approach similar to KVM:selftests. Ignore everything by default and then allow source extensions (.c, .h, .sh) and tracked non-source files.

Note that un-ignoring *.h files causes the generated local_config.h build artifact to incorrectly become untracked. A subsequent patch will rename this header to resolve the issue.

Reviewed-by Lorenzo Stoakes <[email protected]>

Acked-by Mike Rapoport (Microsoft) <[email protected]>

Signed-off-by Pratyush Mallick <[email protected]>

---
 tools/testing/selftests/mm/.gitignore | 71 ++++-----------------------
 1 file changed, 9 insertions(+), 62 deletions(-)

tools/testing/selftests/mm/.gitignore +9 ~~-62~~

index 83ad9454dd9d..fcd892ed21e3 100644
@@ -1,63 +1,10 @@
 # SPDX-License-Identifier: GPL-2.0-only
-cow
-hugepage-mmap
-hugepage-mremap
-hugepage-shm
-hugepage-vmemmap
-hugetlb-madvise
-hugetlb-read-hwpoison
-hugetlb-soft-offline
-khugepaged
-map_hugetlb
-map_populate
-thuge-gen
-compaction_test
-memory-failure
-migration
-mlock2-tests
-mrelease_test
-mremap_dontunmap
-mremap_test
-on-fault-limit
-transhuge-stress
-pagemap_ioctl
-pfnmap
-process_madv
-*.tmp*
-protection_keys
-protection_keys_32
-protection_keys_64
-madv_populate
-uffd-stress
-uffd-unit-tests
-uffd-wp-mremap
-mlock-intersect-test
-mlock-random-test
-virtual_address_range
-gup_test
-va_128TBswitch
-map_fixed_noreplace
-write_to_hugetlbfs
-hmm-tests
-memfd_secret
-soft-dirty
-split_huge_page_test
-ksm_tests
-local_config.h
-local_config.mk
-ksm_functional_tests
-mdwe_test
-gup_longterm
-mkdirty
-va_high_addr_switch
-hugetlb_fault_after_madv
-hugetlb_madv_vs_map
-mseal_test
-droppable
-hugetlb_dio
-pkey_sighandler_tests_32
-pkey_sighandler_tests_64
-guard-regions
-merge
-prctl_thp_disable
-rmap
+*
+!/**/
+!*.c
+!*.h
+!*.sh
+!.gitignore
+!Makefile
+!config
+!settings

Signature · 1 line

2.54.0.563.g4f69b47b94-goog

[email protected]

author

2026-05-11 17:35 UTC #2 lore raw

PATCH v3 2/2 selftests/mm: rename local_config.h to local_config.h_gen

In reply to [email protected]

From: Pratyush Mallick <[email protected]>

Because local_config.h is a generated build artifact, un-ignoring all .h files causes it to incorrectly show up as an untracked file in git status.

Rename it to local_config.h_gen so it no longer matches the !*.h inclusion rule.

Reviewed-by Lorenzo Stoakes <[email protected]>

Acked-by Mike Rapoport (Microsoft) <[email protected]>

Signed-off-by Pratyush Mallick <[email protected]>

---
 tools/testing/selftests/mm/Makefile        | 6 +++---
 tools/testing/selftests/mm/check_config.sh | 2 +-
 tools/testing/selftests/mm/cow.c           | 2 +-
 tools/testing/selftests/mm/gup_longterm.c  | 2 +-
 4 files changed, 6 insertions(+), 6 deletions(-)

tools/testing/selftests/mm/Makefile +3 -3

index 7a5de4e9bf52..32f4b016b74f 100644
@@ -5,7 +5,7 @@
 # script so kunit knows to run it, and add it to the list below.
 # If you do not YOUR TESTS WILL NOT RUN IN THE CI.
 
-LOCAL_HDRS += $(selfdir)/mm/local_config.h $(top_srcdir)/mm/gup_test.h
+LOCAL_HDRS += $(selfdir)/mm/local_config.h_gen $(top_srcdir)/mm/gup_test.h
 LOCAL_HDRS += $(selfdir)/mm/mseal_helpers.h
 
 include local_config.mk

@@ -259,10 +259,10 @@ $(OUTPUT)/migration: LDLIBS += -lnuma
 
 $(OUTPUT)/rmap: LDLIBS += -lnuma
 
-local_config.mk local_config.h: check_config.sh
+local_config.mk local_config.h_gen: check_config.sh
 	CC="$(CC)" CFLAGS="$(CFLAGS)" ./check_config.sh
 
-EXTRA_CLEAN += local_config.mk local_config.h
+EXTRA_CLEAN += local_config.mk local_config.h_gen
 
 ifeq ($(IOURING_EXTRA_LIBS),)
 all: warn_missing_liburing

tools/testing/selftests/mm/check_config.sh +1 -1

index b84c82bbf875..b46bf2d2e7f8 100755
@@ -4,7 +4,7 @@
 # Probe for libraries and create header files to record the results. Both C
 # header files and Makefile include fragments are created.
 
-OUTPUT_H_FILE=local_config.h
+OUTPUT_H_FILE=local_config.h_gen
 OUTPUT_MKFILE=local_config.mk
 
 tmpname=$(mktemp)

tools/testing/selftests/mm/cow.c +1 -1

index d9c69c04b67d..87710c5a2d3c 100644
@@ -21,7 +21,7 @@
 #include <sys/wait.h>
 #include <linux/memfd.h>
 
-#include "local_config.h"
+#include "local_config.h_gen"
 #ifdef LOCAL_CONFIG_HAVE_LIBURING
 #include <liburing.h>
 #endif /* LOCAL_CONFIG_HAVE_LIBURING */

tools/testing/selftests/mm/gup_longterm.c +1 -1

index f61150d28eb2..00ea0dd2f655 100644
@@ -21,7 +21,7 @@
 #include <linux/magic.h>
 #include <linux/memfd.h>
 
-#include "local_config.h"
+#include "local_config.h_gen"
 #ifdef LOCAL_CONFIG_HAVE_LIBURING
 #include <liburing.h>
 #endif /* LOCAL_CONFIG_HAVE_LIBURING */

Signature · 1 line

2.54.0.563.g4f69b47b94-goog

SeongJae Park

2026-05-12 01:01 UTC #3 lore raw

In reply to [email protected]

Quoted 17 lines · On Mon, 11 May 2026 17:34:09 +0000 [email protected] wrote:

From: Pratyush Mallick <[email protected]>
The current selftests/mm/.gitignore hardcodes each generated test binary by name, which requires manual updates every time a new test is added.
This series switches to a pattern-matching approach (similar to KVM selftests), ignoring everything by default and allowing specific source extensions. To accommodate this without tracking generated headers, local_config.h is renamed to local_config.h_gen.
Changelog since v2:
- Split the changes into two separate patches (header rename and .gitignore update)
- Added Suggested-by for David Hildenbrand for the header rename approach.
- Dropped David's and Yosry's Reviewed-by tags.
- Kept Lorenzo Stoakes's Reviewed-by and added Mike Rapoport's Acked-by.

Quoted 4 lines

Changelog since v1 (RFC):
- Renamed local_config.h to local_config.h_gen to avoid conflict with !*.h.
- Updated Makefile, check_config.sh, and affected .c files for the rename.
- Removed *.mod.c as it was unnecessary.

Adding previous revision links together [1] would be helpful for people who willing to see details of previous discussions.

Quoted 4 lines

Pratyush Mallick (2):
selftests/mm: use pattern matching in .gitignore selftests/mm: rename local_config.h to local_config.h_gen

So, after applying the first patch, we may show local_config.h unexpectedly shown on 'git status' output? Maybe too trivial thing, but I'm wondering if we already considered doing the renaming first.

[1] https://docs.kernel.org/process/submitting-patches.html#commentary

Thanks,
SJ

[...]

Sarthak Sharma

2026-07-31 09:16 UTC #4 lore raw

Hi Pratyush and SeongJae!

Quoted 28 lines · SeongJae Park

Quoted 17 lines · On Mon, 11 May 2026 17:34:09 +0000 [email protected] wrote:
From: Pratyush Mallick <[email protected]>
The current selftests/mm/.gitignore hardcodes each generated test binary by name, which requires manual updates every time a new test is added.
This series switches to a pattern-matching approach (similar to KVM selftests), ignoring everything by default and allowing specific source extensions. To accommodate this without tracking generated headers, local_config.h is renamed to local_config.h_gen.
Changelog since v2:
- Split the changes into two separate patches (header rename and .gitignore update)
- Added Suggested-by for David Hildenbrand for the header rename approach.
- Dropped David's and Yosry's Reviewed-by tags.
- Kept Lorenzo Stoakes's Reviewed-by and added Mike Rapoport's Acked-by.
Quoted 4 lines
Changelog since v1 (RFC):
- Renamed local_config.h to local_config.h_gen to avoid conflict with !*.h.
- Updated Makefile, check_config.sh, and affected .c files for the rename.
- Removed *.mod.c as it was unnecessary.
Adding previous revision links together [1] would be helpful for people who willing to see details of previous discussions.
Quoted 1 line

Quoted 7 lines

Quoted 3 lines
Pratyush Mallick (2):
selftests/mm: use pattern matching in .gitignore selftests/mm: rename local_config.h to local_config.h_gen
So, after applying the first patch, we may show local_config.h unexpectedly shown on 'git status' output? Maybe too trivial thing, but I'm wondering if we already considered doing the renaming first.

Just a gentle ping on this patchset, since it has been some time since the last discussion :).

Would it make sense to reorder the patches as SeongJae suggested and respin? Pratyush, if you don't have time to send a v4, I would be happy to prepare and send it. Please let me know if you'd like me to.

Quoted 9 lines

[1] https://docs.kernel.org/process/submitting-patches.html#commentary
Thanks,
SJ
[...]

Thanks,
Sarthak

Pratyush Mallick

author

2026-08-01 22:34 UTC #5 lore raw

Hi Sarthak,

Thanks for the reminder. I completely forgot about this.

Will send a V4 with the patches reordered.

Regards,
Pratyush

Quoted 60 lines · Sarthak Sharma

Hi Pratyush and SeongJae!
Quoted 28 lines · SeongJae Park
Quoted 17 lines · On Mon, 11 May 2026 17:34:09 +0000 [email protected] wrote:
From: Pratyush Mallick <[email protected]>
The current selftests/mm/.gitignore hardcodes each generated test binary by name, which requires manual updates every time a new test is added.
This series switches to a pattern-matching approach (similar to KVM selftests), ignoring everything by default and allowing specific source extensions. To accommodate this without tracking generated headers, local_config.h is renamed to local_config.h_gen.
Changelog since v2:
- Split the changes into two separate patches (header rename and .gitignore update)
- Added Suggested-by for David Hildenbrand for the header rename approach.
- Dropped David's and Yosry's Reviewed-by tags.
- Kept Lorenzo Stoakes's Reviewed-by and added Mike Rapoport's Acked-by.
Quoted 4 lines
Changelog since v1 (RFC):
- Renamed local_config.h to local_config.h_gen to avoid conflict with !*.h.
- Updated Makefile, check_config.sh, and affected .c files for the rename.
- Removed *.mod.c as it was unnecessary.
Adding previous revision links together [1] would be helpful for people who willing to see details of previous discussions.
Quoted 1 line
Quoted 7 lines
Quoted 3 lines
Pratyush Mallick (2):
selftests/mm: use pattern matching in .gitignore selftests/mm: rename local_config.h to local_config.h_gen
So, after applying the first patch, we may show local_config.h unexpectedly shown on 'git status' output? Maybe too trivial thing, but I'm wondering if we already considered doing the renaming first.
Just a gentle ping on this patchset, since it has been some time since the last discussion :).
Would it make sense to reorder the patches as SeongJae suggested and respin? Pratyush, if you don't have time to send a v4, I would be happy to prepare and send it. Please let me know if you'd like me to.
Quoted 9 lines
[1] https://docs.kernel.org/process/submitting-patches.html#commentary
Thanks,
SJ
[...]
Thanks,
Sarthak