diff mbox series

[10/34] generic/759,760: skip test if we can't set up a hugepage for IO

Message ID 173933094507.1758477.425979019420266054.stgit@frogsfrogsfrogs (mailing list archive)
State New
Headers show
Series [01/34] generic/476: fix fsstress process management | expand

Commit Message

Darrick J. Wong Feb. 12, 2025, 3:33 a.m. UTC 
From: Darrick J. Wong <djwong@kernel.org>

On an arm64 VM with 64k base pages and a paltry 8G of RAM, this test
will frequently fail like this:

>  QA output created by 759
>  fsx -N 10000 -l 500000 -h
> -fsx -N 10000 -o 8192 -l 500000 -h
> -fsx -N 10000 -o 128000 -l 500000 -h
> +Seed set to 1
> +madvise collapse for buf: Cannot allocate memory
> +init_hugepages_buf failed for good_buf: Cannot allocate memory

This system has a 512MB hugepage size, which means that there's a good
chance that memory is so fragmented that we won't be able to create a
huge page (in 1/16th the available DRAM).  Create a _run_hugepage_fsx
helper that will detect this situation at the start of the test and skip
it, having refactored run_fsx into a properly namespaced version that
won't exit the test on failure.

Cc: <fstests@vger.kernel.org> # v2025.02.02
Cc: joannelkoong@gmail.com
Fixes: 627289232371e3 ("generic: add tests for read/writes from hugepages-backed buffers")
Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 common/rc         |   34 ++++++++++++++++++++++++++++++----
 ltp/fsx.c         |    6 ++++--
 tests/generic/759 |    6 +++---
 tests/generic/760 |    6 +++---
 4 files changed, 40 insertions(+), 12 deletions(-)

Comments

Joanne Koong Feb. 12, 2025, 6:39 p.m. UTC | #1 
On Tue, Feb 11, 2025 at 7:33 PM Darrick J. Wong <djwong@kernel.org> wrote:
>
> From: Darrick J. Wong <djwong@kernel.org>
>
> On an arm64 VM with 64k base pages and a paltry 8G of RAM, this test
> will frequently fail like this:
>
> >  QA output created by 759
> >  fsx -N 10000 -l 500000 -h
> > -fsx -N 10000 -o 8192 -l 500000 -h
> > -fsx -N 10000 -o 128000 -l 500000 -h
> > +Seed set to 1
> > +madvise collapse for buf: Cannot allocate memory
> > +init_hugepages_buf failed for good_buf: Cannot allocate memory
>
> This system has a 512MB hugepage size, which means that there's a good
> chance that memory is so fragmented that we won't be able to create a
> huge page (in 1/16th the available DRAM).  Create a _run_hugepage_fsx
> helper that will detect this situation at the start of the test and skip
> it, having refactored run_fsx into a properly namespaced version that
> won't exit the test on failure.
>
> Cc: <fstests@vger.kernel.org> # v2025.02.02
> Cc: joannelkoong@gmail.com
> Fixes: 627289232371e3 ("generic: add tests for read/writes from hugepages-backed buffers")
> Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>

Thanks for adding this.

Reviewed-by: Joanne Koong <joannelkoong@gmail.com>

> ---
>  common/rc         |   34 ++++++++++++++++++++++++++++++----
>  ltp/fsx.c         |    6 ++++--
>  tests/generic/759 |    6 +++---
>  tests/generic/760 |    6 +++---
>  4 files changed, 40 insertions(+), 12 deletions(-)
>
>
> diff --git a/common/rc b/common/rc
> index b7736173e6e839..36e270abbc082a 100644
> --- a/common/rc
> +++ b/common/rc
> @@ -4982,20 +4982,46 @@ _require_hugepage_fsx()
>                 _notrun "fsx binary does not support MADV_COLLAPSE"
>  }
>
> -run_fsx()
> +_run_fsx()
>  {
> -       echo fsx $@
> +       echo "fsx $*"
>         local args=`echo $@ | sed -e "s/ BSIZE / $bsize /g" -e "s/ PSIZE / $psize /g"`
>         set -- $here/ltp/fsx $args $FSX_AVOID $TEST_DIR/junk
>         echo "$@" >>$seqres.full
>         rm -f $TEST_DIR/junk
>         "$@" 2>&1 | tee -a $seqres.full >$tmp.fsx
> -       if [ ${PIPESTATUS[0]} -ne 0 ]; then
> +       local res=${PIPESTATUS[0]}
> +       if [ $res -ne 0 ]; then
>                 cat $tmp.fsx
>                 rm -f $tmp.fsx
> -               exit 1
> +               return $res
>         fi
>         rm -f $tmp.fsx
> +       return 0

I think this could also be further simplified to

if [$res -ne 0]; then
    cat $tmp.fsx
fi
rm -rf $tmp.fsx
return $res


> +}
> +
> +# Run fsx with -h(ugepage buffers).  If we can't set up a hugepage then skip
> +# the test, but if any other error occurs then exit the test.
> +_run_hugepage_fsx() {
> +       _run_fsx "$@" -h &> $tmp.hugepage_fsx
> +       local res=$?
> +       if [ $res -eq 103 ]; then
> +               # According to the MADV_COLLAPSE manpage, these three errors
> +               # can happen if the kernel could not collapse a collection of
> +               # pages into a single huge page.
> +               grep -q -E ' for hugebuf: (Cannot allocate memory|Device or resource busy|Resource temporarily unavailable)' $tmp.hugepage_fsx && \
> +                       _notrun "Could not set up huge page for test"
> +       fi
> +       cat $tmp.hugepage_fsx
> +       rm -f $tmp.hugepage_fsx
> +       test $res -ne 0 && exit 1
> +       return 0
> +}
> +
> +# run fsx or exit the test
> +run_fsx()
> +{
> +       _run_fsx "$@" || exit 1
>  }
>
>  _require_statx()
> diff --git a/ltp/fsx.c b/ltp/fsx.c
> index cf9502a74c17a7..d1b0f245582b31 100644
> --- a/ltp/fsx.c
> +++ b/ltp/fsx.c
> @@ -2974,13 +2974,15 @@ init_hugepages_buf(unsigned len, int hugepage_size, int alignment, long *buf_siz
>
>         ret = posix_memalign(&buf, hugepage_size, size);
>         if (ret) {
> -               prterr("posix_memalign for buf");
> +               /* common/rc greps this error message */
> +               prterr("posix_memalign for hugebuf");
>                 return NULL;
>         }
>         memset(buf, '\0', size);
>         ret = madvise(buf, size, MADV_COLLAPSE);
>         if (ret) {
> -               prterr("madvise collapse for buf");
> +               /* common/rc greps this error message */
> +               prterr("madvise collapse for hugebuf");
>                 free(buf);
>                 return NULL;
>         }
> diff --git a/tests/generic/759 b/tests/generic/759
> index a7dec155056abc..49c02214559a55 100755
> --- a/tests/generic/759
> +++ b/tests/generic/759
> @@ -15,9 +15,9 @@ _require_test
>  _require_thp
>  _require_hugepage_fsx
>
> -run_fsx -N 10000            -l 500000 -h
> -run_fsx -N 10000  -o 8192   -l 500000 -h
> -run_fsx -N 10000  -o 128000 -l 500000 -h
> +_run_hugepage_fsx -N 10000            -l 500000
> +_run_hugepage_fsx -N 10000  -o 8192   -l 500000
> +_run_hugepage_fsx -N 10000  -o 128000 -l 500000
>
>  status=0
>  exit
> diff --git a/tests/generic/760 b/tests/generic/760
> index 4781a8d1eec4ec..f270636e56a377 100755
> --- a/tests/generic/760
> +++ b/tests/generic/760
> @@ -19,9 +19,9 @@ _require_hugepage_fsx
>  psize=`$here/src/feature -s`
>  bsize=`$here/src/min_dio_alignment $TEST_DIR $TEST_DEV`
>
> -run_fsx -N 10000            -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W -h
> -run_fsx -N 10000  -o 8192   -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W -h
> -run_fsx -N 10000  -o 128000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W -h
> +_run_hugepage_fsx -N 10000            -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
> +_run_hugepage_fsx -N 10000  -o 8192   -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
> +_run_hugepage_fsx -N 10000  -o 128000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
>
>  status=0
>  exit
>
diff mbox series

Patch

diff --git a/common/rc b/common/rc
index b7736173e6e839..36e270abbc082a 100644
--- a/common/rc
+++ b/common/rc
@@ -4982,20 +4982,46 @@  _require_hugepage_fsx()
 		_notrun "fsx binary does not support MADV_COLLAPSE"
 }
 
-run_fsx()
+_run_fsx()
 {
-	echo fsx $@
+	echo "fsx $*"
 	local args=`echo $@ | sed -e "s/ BSIZE / $bsize /g" -e "s/ PSIZE / $psize /g"`
 	set -- $here/ltp/fsx $args $FSX_AVOID $TEST_DIR/junk
 	echo "$@" >>$seqres.full
 	rm -f $TEST_DIR/junk
 	"$@" 2>&1 | tee -a $seqres.full >$tmp.fsx
-	if [ ${PIPESTATUS[0]} -ne 0 ]; then
+	local res=${PIPESTATUS[0]}
+	if [ $res -ne 0 ]; then
 		cat $tmp.fsx
 		rm -f $tmp.fsx
-		exit 1
+		return $res
 	fi
 	rm -f $tmp.fsx
+	return 0
+}
+
+# Run fsx with -h(ugepage buffers).  If we can't set up a hugepage then skip
+# the test, but if any other error occurs then exit the test.
+_run_hugepage_fsx() {
+	_run_fsx "$@" -h &> $tmp.hugepage_fsx
+	local res=$?
+	if [ $res -eq 103 ]; then
+		# According to the MADV_COLLAPSE manpage, these three errors
+		# can happen if the kernel could not collapse a collection of
+		# pages into a single huge page.
+		grep -q -E ' for hugebuf: (Cannot allocate memory|Device or resource busy|Resource temporarily unavailable)' $tmp.hugepage_fsx && \
+			_notrun "Could not set up huge page for test"
+	fi
+	cat $tmp.hugepage_fsx
+	rm -f $tmp.hugepage_fsx
+	test $res -ne 0 && exit 1
+	return 0
+}
+
+# run fsx or exit the test
+run_fsx()
+{
+	_run_fsx "$@" || exit 1
 }
 
 _require_statx()
diff --git a/ltp/fsx.c b/ltp/fsx.c
index cf9502a74c17a7..d1b0f245582b31 100644
--- a/ltp/fsx.c
+++ b/ltp/fsx.c
@@ -2974,13 +2974,15 @@  init_hugepages_buf(unsigned len, int hugepage_size, int alignment, long *buf_siz
 
 	ret = posix_memalign(&buf, hugepage_size, size);
 	if (ret) {
-		prterr("posix_memalign for buf");
+		/* common/rc greps this error message */
+		prterr("posix_memalign for hugebuf");
 		return NULL;
 	}
 	memset(buf, '\0', size);
 	ret = madvise(buf, size, MADV_COLLAPSE);
 	if (ret) {
-		prterr("madvise collapse for buf");
+		/* common/rc greps this error message */
+		prterr("madvise collapse for hugebuf");
 		free(buf);
 		return NULL;
 	}
diff --git a/tests/generic/759 b/tests/generic/759
index a7dec155056abc..49c02214559a55 100755
--- a/tests/generic/759
+++ b/tests/generic/759
@@ -15,9 +15,9 @@  _require_test
 _require_thp
 _require_hugepage_fsx
 
-run_fsx -N 10000            -l 500000 -h
-run_fsx -N 10000  -o 8192   -l 500000 -h
-run_fsx -N 10000  -o 128000 -l 500000 -h
+_run_hugepage_fsx -N 10000            -l 500000
+_run_hugepage_fsx -N 10000  -o 8192   -l 500000
+_run_hugepage_fsx -N 10000  -o 128000 -l 500000
 
 status=0
 exit
diff --git a/tests/generic/760 b/tests/generic/760
index 4781a8d1eec4ec..f270636e56a377 100755
--- a/tests/generic/760
+++ b/tests/generic/760
@@ -19,9 +19,9 @@  _require_hugepage_fsx
 psize=`$here/src/feature -s`
 bsize=`$here/src/min_dio_alignment $TEST_DIR $TEST_DEV`
 
-run_fsx -N 10000            -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W -h
-run_fsx -N 10000  -o 8192   -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W -h
-run_fsx -N 10000  -o 128000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W -h
+_run_hugepage_fsx -N 10000            -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
+_run_hugepage_fsx -N 10000  -o 8192   -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
+_run_hugepage_fsx -N 10000  -o 128000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
 
 status=0
 exit