[PATCH 1/2] test: Allow more variations in expected fbt kernel stacks

[PATCH 1/2] test: Allow more variations in expected fbt kernel stacks

[eugene.loh]

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

This test checks the call stack upon entry to vfs_write().
Unfortunately, these checks require some maintenance since the call
stack can vary -- slightly or greatly -- depending on processor or
kernel.  There is a competition between ease of test maintenance and
strictness of correctness checks.

Adapt post processing of output to allow new variations in stacks
seen in UEK 8 (currently Linux 6.12).

Orabug: 37459289
Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
---
 test/unittest/stack/tst.stack_fbt.sh | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/test/unittest/stack/tst.stack_fbt.sh b/test/unittest/stack/tst.stack_fbt.sh
index c365e14d6..8a0b7999c 100755
--- a/test/unittest/stack/tst.stack_fbt.sh
+++ b/test/unittest/stack/tst.stack_fbt.sh
@@ -37,9 +37,18 @@ if [ $? -ne 0 ]; then
 	exit 1
 fi
 
-# Strip out blank lines and pointer values.
-
-awk 'NF != 0 { sub(/+0x[0-9a-f]*$/, "+{ptr}"); print }' dtrace.out > dtrace.post
+# Strip out
+# - blank lines
+# - "constprop"
+# - "isra"
+# - "_after_hwframe"    (x86 starting with UEK8)
+# - pointer values
+
+awk 'NF != 0 { sub("\\.constprop\\.[0-9]", "");
+               sub("\\.isra\\.[0-9]", "");
+               sub("_after_hwframe\\+", "+");
+               sub(/+0x[0-9a-f]*$/, "+{ptr}");
+               print }' dtrace.out > dtrace.post
 if [ $? -ne 0 ]; then
 	echo ERROR: awk failed
 	cat dtrace.out
-- 
2.43.5

[PATCH 2/2] test: Account for pid:::entry ustack() being correct

[eugene.loh]

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

The pid:::entry uprobe fires so early in the function preamble
that the frame pointer is not yet set and the caller is not (yet)
correctly identified.  In Linux 6.11, x86-specific heuristics
address this problem.

Post process results from this test to accommodate both cases --
missing caller and not missing caller.

Orabug: 37459289
Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
---
 test/unittest/ustack/tst.ustack25_pid.r   |  1 +
 test/unittest/ustack/tst.ustack25_pid.r.p | 37 +++++++++++++++++++++++
 2 files changed, 38 insertions(+)
 create mode 100755 test/unittest/ustack/tst.ustack25_pid.r.p

diff --git a/test/unittest/ustack/tst.ustack25_pid.r b/test/unittest/ustack/tst.ustack25_pid.r
index 89b35c028..e7732fb81 100644
--- a/test/unittest/ustack/tst.ustack25_pid.r
+++ b/test/unittest/ustack/tst.ustack25_pid.r
@@ -1,5 +1,6 @@
 
               ustack-tst-basic`myfunc_z
+              ustack-tst-basic`myfunc_y+{ptr}
               ustack-tst-basic`myfunc_x+{ptr}
               ustack-tst-basic`myfunc_w+{ptr}
               ustack-tst-basic`myfunc_v+{ptr}
diff --git a/test/unittest/ustack/tst.ustack25_pid.r.p b/test/unittest/ustack/tst.ustack25_pid.r.p
new file mode 100755
index 000000000..6486f7ec6
--- /dev/null
+++ b/test/unittest/ustack/tst.ustack25_pid.r.p
@@ -0,0 +1,37 @@
+#!/bin/sh
+
+# A pid entry probe places a uprobe on the first instruction of a function.
+# Unfortunately, this is so early in the function preamble that the function
+# frame pointer has not yet been established and the actual caller of the
+# traced function is missed.
+#
+# In Linux 6.11, x86-specific heuristics are introduced to fix this problem.
+# See commit cfa7f3d
+# ("perf,x86: avoid missing caller address in stack traces captured in uprobe")
+# for both a description of the problem and an explanation of the heuristics.
+#
+# Add post processing to these test results to allow for both cases:
+# caller frame is missing or not missing.
+
+missing_caller=1
+if [ $(uname -m) == "x86_64" ]; then
+        read MAJOR MINOR <<< `uname -r | grep -Eo '^[0-9]+\.[0-9]+' | tr '.' ' '`
+
+        if [ $MAJOR -ge 6 ]; then
+                if [ $MAJOR -gt 6 -o $MINOR -ge 11 ]; then
+                        missing_caller=0
+                fi
+        fi
+fi
+
+if [ $missing_caller -eq 1 ]; then
+        # Add the missing caller function after the current function.
+        awk '{ print }
+             /myfunc_z/ { print "              ustack-tst-basic`myfunc_y+{ptr}" }'
+else
+        # The .r results file has an extra frame at the end in case
+        # the caller frame is missing and the 25-frame limit goes
+        # "too far."  If the caller is not missing, fake that extra frame.
+        awk '{ print }
+             /myfunc_b/ { print "              ustack-tst-basic`myfunc_a+{ptr}" }'
+fi
-- 
2.43.5

Re: [PATCH 1/2] test: Allow more variations in expected fbt kernel stacks

[Kris Van Hees]

On Sat, Feb 08, 2025 at 06:21:05PM -0500, eugene.loh@oracle.com wrote:
> From: Eugene Loh <eugene.loh@oracle.com>
> 
> This test checks the call stack upon entry to vfs_write().
> Unfortunately, these checks require some maintenance since the call
> stack can vary -- slightly or greatly -- depending on processor or
> kernel.  There is a competition between ease of test maintenance and
> strictness of correctness checks.
> 
> Adapt post processing of output to allow new variations in stacks
> seen in UEK 8 (currently Linux 6.12).
> 
> Orabug: 37459289
> Signed-off-by: Eugene Loh <eugene.loh@oracle.com>

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

> ---
>  test/unittest/stack/tst.stack_fbt.sh | 15 ++++++++++++---
>  1 file changed, 12 insertions(+), 3 deletions(-)
> 
> diff --git a/test/unittest/stack/tst.stack_fbt.sh b/test/unittest/stack/tst.stack_fbt.sh
> index c365e14d6..8a0b7999c 100755
> --- a/test/unittest/stack/tst.stack_fbt.sh
> +++ b/test/unittest/stack/tst.stack_fbt.sh
> @@ -37,9 +37,18 @@ if [ $? -ne 0 ]; then
>  	exit 1
>  fi
>  
> -# Strip out blank lines and pointer values.
> -
> -awk 'NF != 0 { sub(/+0x[0-9a-f]*$/, "+{ptr}"); print }' dtrace.out > dtrace.post
> +# Strip out
> +# - blank lines
> +# - "constprop"
> +# - "isra"
> +# - "_after_hwframe"    (x86 starting with UEK8)
> +# - pointer values
> +
> +awk 'NF != 0 { sub("\\.constprop\\.[0-9]", "");
> +               sub("\\.isra\\.[0-9]", "");
> +               sub("_after_hwframe\\+", "+");
> +               sub(/+0x[0-9a-f]*$/, "+{ptr}");
> +               print }' dtrace.out > dtrace.post
>  if [ $? -ne 0 ]; then
>  	echo ERROR: awk failed
>  	cat dtrace.out
> -- 
> 2.43.5
> 

Re: [PATCH 2/2] test: Account for pid:::entry ustack() being correct

[Kris Van Hees]

On Sat, Feb 08, 2025 at 06:21:06PM -0500, eugene.loh@oracle.com wrote:
> From: Eugene Loh <eugene.loh@oracle.com>
> 
> The pid:::entry uprobe fires so early in the function preamble
> that the frame pointer is not yet set and the caller is not (yet)
> correctly identified.  In Linux 6.11, x86-specific heuristics
> address this problem.
> 
> Post process results from this test to accommodate both cases --
> missing caller and not missing caller.
> 
> Orabug: 37459289
> Signed-off-by: Eugene Loh <eugene.loh@oracle.com>

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

> ---
>  test/unittest/ustack/tst.ustack25_pid.r   |  1 +
>  test/unittest/ustack/tst.ustack25_pid.r.p | 37 +++++++++++++++++++++++
>  2 files changed, 38 insertions(+)
>  create mode 100755 test/unittest/ustack/tst.ustack25_pid.r.p
> 
> diff --git a/test/unittest/ustack/tst.ustack25_pid.r b/test/unittest/ustack/tst.ustack25_pid.r
> index 89b35c028..e7732fb81 100644
> --- a/test/unittest/ustack/tst.ustack25_pid.r
> +++ b/test/unittest/ustack/tst.ustack25_pid.r
> @@ -1,5 +1,6 @@
>  
>                ustack-tst-basic`myfunc_z
> +              ustack-tst-basic`myfunc_y+{ptr}
>                ustack-tst-basic`myfunc_x+{ptr}
>                ustack-tst-basic`myfunc_w+{ptr}
>                ustack-tst-basic`myfunc_v+{ptr}
> diff --git a/test/unittest/ustack/tst.ustack25_pid.r.p b/test/unittest/ustack/tst.ustack25_pid.r.p
> new file mode 100755
> index 000000000..6486f7ec6
> --- /dev/null
> +++ b/test/unittest/ustack/tst.ustack25_pid.r.p
> @@ -0,0 +1,37 @@
> +#!/bin/sh
> +
> +# A pid entry probe places a uprobe on the first instruction of a function.
> +# Unfortunately, this is so early in the function preamble that the function
> +# frame pointer has not yet been established and the actual caller of the
> +# traced function is missed.
> +#
> +# In Linux 6.11, x86-specific heuristics are introduced to fix this problem.
> +# See commit cfa7f3d
> +# ("perf,x86: avoid missing caller address in stack traces captured in uprobe")
> +# for both a description of the problem and an explanation of the heuristics.
> +#
> +# Add post processing to these test results to allow for both cases:
> +# caller frame is missing or not missing.
> +
> +missing_caller=1
> +if [ $(uname -m) == "x86_64" ]; then
> +        read MAJOR MINOR <<< `uname -r | grep -Eo '^[0-9]+\.[0-9]+' | tr '.' ' '`
> +
> +        if [ $MAJOR -ge 6 ]; then
> +                if [ $MAJOR -gt 6 -o $MINOR -ge 11 ]; then
> +                        missing_caller=0
> +                fi
> +        fi
> +fi
> +
> +if [ $missing_caller -eq 1 ]; then
> +        # Add the missing caller function after the current function.
> +        awk '{ print }
> +             /myfunc_z/ { print "              ustack-tst-basic`myfunc_y+{ptr}" }'
> +else
> +        # The .r results file has an extra frame at the end in case
> +        # the caller frame is missing and the 25-frame limit goes
> +        # "too far."  If the caller is not missing, fake that extra frame.
> +        awk '{ print }
> +             /myfunc_b/ { print "              ustack-tst-basic`myfunc_a+{ptr}" }'
> +fi
> -- 
> 2.43.5
>