public inbox for dtrace@lists.linux.dev
 help / color / mirror / Atom feed
* [PATCH v2] test: Account for large stack-limit kernel parameter
@ 2025-10-09 21:33 eugene.loh
  2025-10-13 19:15 ` Kris Van Hees
  0 siblings, 1 reply; 2+ messages in thread
From: eugene.loh @ 2025-10-09 21:33 UTC (permalink / raw)
  To: dtrace, dtrace-devel

From: Eugene Loh <eugene.loh@oracle.com>

In commit d0fdeec50 ("test: Fix some ustackdepth tests"), the test was
tweaked to check that the full stack would be recovered if the kernel
stack-limit parameter were increased from a default 127 to greater than
about 190 frames.  But this does not work if the system already has
reset the kernel parameter to a high value.

Change the test to deliberately try a "too small" and a "big enough"
value -- again, ultimately restoring the original value.

Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
---
 .../variables/bvar/tst.ustackdepth_big.sh     | 106 +++++++++++-------
 1 file changed, 68 insertions(+), 38 deletions(-)

diff --git a/test/unittest/variables/bvar/tst.ustackdepth_big.sh b/test/unittest/variables/bvar/tst.ustackdepth_big.sh
index 4ce53d929..aeef9982c 100755
--- a/test/unittest/variables/bvar/tst.ustackdepth_big.sh
+++ b/test/unittest/variables/bvar/tst.ustackdepth_big.sh
@@ -23,52 +23,82 @@ cd $DIRNAME
 
 orig_maxstack=`sysctl -n kernel.perf_event_max_stack`
 echo kernel.perf_event_max_stack was $orig_maxstack
-trap "sysctl kernel.perf_event_max_stack=$orig_maxstack" QUIT EXIT
-sysctl kernel.perf_event_max_stack=200
+trap "sleep 2; sysctl kernel.perf_event_max_stack=$orig_maxstack" QUIT EXIT
 
-$dtrace $dt_flags -c $TRIGGER -qn '
-profile-1
-/pid == $target/
-{
-    printf("DEPTH %d\n", ustackdepth);
-    printf("TRACE BEGIN\n");
-    ustack(200);
-    printf("TRACE END\n");
-    exit(0);
-}
-ERROR
-{
-    exit(1);
-}
-' > D.out
-if [ $? -ne 0 ]; then
-    echo DTrace failure
-    exit 1
-fi
+# set bounds on the full stack depth (which is ambiguous)
+lo=188
+hi=192
 
-sleep 2
-sysctl kernel.perf_event_max_stack=$orig_maxstack
+function do_dtrace() {
+    # set the kernel parameter
+    sysctl kernel.perf_event_max_stack=$stack_limit
 
-$POSTPROC D.out > awk.out
-if [ $? -ne 0 ]; then
-    echo post processing failure
-    exit 1
-fi
+    # run dtrace
+    $dtrace $dt_flags -c $TRIGGER -qn '
+    profile-1
+    /pid == $target/
+    {
+        printf("DEPTH %d\n", ustackdepth);
+        printf("TRACE BEGIN\n");
+        ustack(200);
+        printf("TRACE END\n");
+        exit(0);
+    }
+    ERROR
+    {
+        exit(1);
+    }
+    ' > D.out.$stack_limit
+    if [ $? -ne 0 ]; then
+        echo ERROR: DTrace failure with $stack_limit
+        exit 1
+    fi
 
-if echo "Stack depth OK" | diff -q - awk.out; then
-    mydepth=`gawk '/DEPTH/ { print $2 }' D.out`
-    if [ $mydepth -gt $orig_maxstack ]; then
-        echo success depth $mydepth exceeded original limit $orig_maxstack
-        exit 0
-    else
-        echo ERROR: $mydepth does not exceed original limit $orig_maxstack
+    # check stackdepth consistency
+    $POSTPROC D.out.$stack_limit > awk.out.$stack_limit
+    if [ $? -ne 0 ]; then
+        echo ERROR: post processing failure
+        exit 1
+    fi
+    if ! grep -q "Stack depth OK" awk.out.$stack_limit; then
+        echo ERROR: stack depth does not match stack
         cat D.out
         exit 1
     fi
-else
-    echo "ERROR: stack depth does not match stack"
-    cat D.out
+
+    # get actual stack depth
+    mydepth=`gawk '/DEPTH/ { print $2 }' D.out.$stack_limit`
+    echo with limit $stack_limit got stack depth $mydepth
+
+    # provide breathing room between dtrace and resetting kernel parameter
+    sleep 2
+}
+
+# try a stack limit that is too small
+stack_limit=$(($lo / 2))
+do_dtrace
+echo "   " stack limit $stack_limit is too small for the entire stack
+if [ $mydepth -ne $stack_limit ]; then
+    echo ERROR: $mydepth does not match $stack_limit
     exit 1
 fi
+echo "   " success:  actual depth matches limit
+
+# try a stack limit that is large enough
+stack_limit=$(($hi + 20))
+do_dtrace
+echo "   " stack limit $stack_limit is large enough
+if [ $mydepth -lt $lo ]; then
+    echo ERROR: $mydepth is lower than low bound $lo
+    exit 1
+fi
+if [ $mydepth -gt $hi ]; then
+    echo ERROR: $mydepth is greater than high bound $hi
+    exit 1
+fi
+echo "   " success:  actual depth captures full stack
+
+# restore the kernel parameter
+sysctl kernel.perf_event_max_stack=$orig_maxstack
 
 exit 0
-- 
2.47.3


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH v2] test: Account for large stack-limit kernel parameter
  2025-10-09 21:33 [PATCH v2] test: Account for large stack-limit kernel parameter eugene.loh
@ 2025-10-13 19:15 ` Kris Van Hees
  0 siblings, 0 replies; 2+ messages in thread
From: Kris Van Hees @ 2025-10-13 19:15 UTC (permalink / raw)
  To: eugene.loh; +Cc: dtrace, dtrace-devel

On Thu, Oct 09, 2025 at 05:33:42PM -0400, eugene.loh@oracle.com wrote:
> From: Eugene Loh <eugene.loh@oracle.com>
> 
> In commit d0fdeec50 ("test: Fix some ustackdepth tests"), the test was
> tweaked to check that the full stack would be recovered if the kernel
> stack-limit parameter were increased from a default 127 to greater than
> about 190 frames.  But this does not work if the system already has
> reset the kernel parameter to a high value.
> 
> Change the test to deliberately try a "too small" and a "big enough"
> value -- again, ultimately restoring the original value.
> 
> Signed-off-by: Eugene Loh <eugene.loh@oracle.com>

Reviewed-by: Kris Van Hees <kris.van.hees@oracle.com>

> ---
>  .../variables/bvar/tst.ustackdepth_big.sh     | 106 +++++++++++-------
>  1 file changed, 68 insertions(+), 38 deletions(-)
> 
> diff --git a/test/unittest/variables/bvar/tst.ustackdepth_big.sh b/test/unittest/variables/bvar/tst.ustackdepth_big.sh
> index 4ce53d929..aeef9982c 100755
> --- a/test/unittest/variables/bvar/tst.ustackdepth_big.sh
> +++ b/test/unittest/variables/bvar/tst.ustackdepth_big.sh
> @@ -23,52 +23,82 @@ cd $DIRNAME
>  
>  orig_maxstack=`sysctl -n kernel.perf_event_max_stack`
>  echo kernel.perf_event_max_stack was $orig_maxstack
> -trap "sysctl kernel.perf_event_max_stack=$orig_maxstack" QUIT EXIT
> -sysctl kernel.perf_event_max_stack=200
> +trap "sleep 2; sysctl kernel.perf_event_max_stack=$orig_maxstack" QUIT EXIT
>  
> -$dtrace $dt_flags -c $TRIGGER -qn '
> -profile-1
> -/pid == $target/
> -{
> -    printf("DEPTH %d\n", ustackdepth);
> -    printf("TRACE BEGIN\n");
> -    ustack(200);
> -    printf("TRACE END\n");
> -    exit(0);
> -}
> -ERROR
> -{
> -    exit(1);
> -}
> -' > D.out
> -if [ $? -ne 0 ]; then
> -    echo DTrace failure
> -    exit 1
> -fi
> +# set bounds on the full stack depth (which is ambiguous)
> +lo=188
> +hi=192
>  
> -sleep 2
> -sysctl kernel.perf_event_max_stack=$orig_maxstack
> +function do_dtrace() {
> +    # set the kernel parameter
> +    sysctl kernel.perf_event_max_stack=$stack_limit
>  
> -$POSTPROC D.out > awk.out
> -if [ $? -ne 0 ]; then
> -    echo post processing failure
> -    exit 1
> -fi
> +    # run dtrace
> +    $dtrace $dt_flags -c $TRIGGER -qn '
> +    profile-1
> +    /pid == $target/
> +    {
> +        printf("DEPTH %d\n", ustackdepth);
> +        printf("TRACE BEGIN\n");
> +        ustack(200);
> +        printf("TRACE END\n");
> +        exit(0);
> +    }
> +    ERROR
> +    {
> +        exit(1);
> +    }
> +    ' > D.out.$stack_limit
> +    if [ $? -ne 0 ]; then
> +        echo ERROR: DTrace failure with $stack_limit
> +        exit 1
> +    fi
>  
> -if echo "Stack depth OK" | diff -q - awk.out; then
> -    mydepth=`gawk '/DEPTH/ { print $2 }' D.out`
> -    if [ $mydepth -gt $orig_maxstack ]; then
> -        echo success depth $mydepth exceeded original limit $orig_maxstack
> -        exit 0
> -    else
> -        echo ERROR: $mydepth does not exceed original limit $orig_maxstack
> +    # check stackdepth consistency
> +    $POSTPROC D.out.$stack_limit > awk.out.$stack_limit
> +    if [ $? -ne 0 ]; then
> +        echo ERROR: post processing failure
> +        exit 1
> +    fi
> +    if ! grep -q "Stack depth OK" awk.out.$stack_limit; then
> +        echo ERROR: stack depth does not match stack
>          cat D.out
>          exit 1
>      fi
> -else
> -    echo "ERROR: stack depth does not match stack"
> -    cat D.out
> +
> +    # get actual stack depth
> +    mydepth=`gawk '/DEPTH/ { print $2 }' D.out.$stack_limit`
> +    echo with limit $stack_limit got stack depth $mydepth
> +
> +    # provide breathing room between dtrace and resetting kernel parameter
> +    sleep 2
> +}
> +
> +# try a stack limit that is too small
> +stack_limit=$(($lo / 2))
> +do_dtrace
> +echo "   " stack limit $stack_limit is too small for the entire stack
> +if [ $mydepth -ne $stack_limit ]; then
> +    echo ERROR: $mydepth does not match $stack_limit
>      exit 1
>  fi
> +echo "   " success:  actual depth matches limit
> +
> +# try a stack limit that is large enough
> +stack_limit=$(($hi + 20))
> +do_dtrace
> +echo "   " stack limit $stack_limit is large enough
> +if [ $mydepth -lt $lo ]; then
> +    echo ERROR: $mydepth is lower than low bound $lo
> +    exit 1
> +fi
> +if [ $mydepth -gt $hi ]; then
> +    echo ERROR: $mydepth is greater than high bound $hi
> +    exit 1
> +fi
> +echo "   " success:  actual depth captures full stack
> +
> +# restore the kernel parameter
> +sysctl kernel.perf_event_max_stack=$orig_maxstack
>  
>  exit 0
> -- 
> 2.47.3
> 

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2025-10-13 19:15 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-10-09 21:33 [PATCH v2] test: Account for large stack-limit kernel parameter eugene.loh
2025-10-13 19:15 ` Kris Van Hees

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox