Re: [PATCH v9 2/9] selftests/livepatch: Adapt atomic replace tests to provides/obsoletes

On Tue, Sep 29, 2026 at 4:43 PM Miroslav Benes [off-list ref] wrote:

quoted

quoted

quoted

 # check_result() - verify dmesg output
 #    TODO - better filter, out of order msgs, etc?
 function check_result {

@@ -365,6 +374,27 @@ function does_sysfs_exist() {
      [[ -f "$SYSFS_KLP_DIR/$mod/$attr" ]]
 }

+# detect_provides_attr() - detect whether the running kernel supports the
+# livepatch "provides" attribute and set HAS_PROVIDES_ATTR accordingly.
+# The provides/obsoletes based tests are only run when HAS_PROVIDES_ATTR
+# is set.
+function detect_provides_attr() {
+     HAS_PROVIDES_ATTR=0
+
+     if [[ -r /proc/config.gz ]] &&
+        zgrep -q "CONFIG_KLP_HAS_PROVIDES=y" /proc/config.gz 2>/dev/null; then
+             HAS_PROVIDES_ATTR=1
+             return 0
+     fi
+
+     load_lp test_klp_cmdline_lp
+     if does_sysfs_exist test_klp_cmdline_lp "provides"; then
+             HAS_PROVIDES_ATTR=1
+     fi
+     disable_lp test_klp_cmdline_lp
+     unload_lp test_klp_cmdline_lp
+}

Why is there does_sysfs_exist() fallback? I mean, we use that currently
for exactly this purpose. My understanding is that
CONFIG_KLP_HAS_PROVIDES is also used elsewhere in the code so using it
here makes some sense, but do we need both?

CONFIG_IKCONFIG_PROC may not be enabled, so if /proc/config.gz doesn't
exist, we must fall back to an alternative one.

Ok. Let's keep it like that because it is definitely cheaper than loading
and unloading a module. Could you add a comment just before
does_sysfs_exist() fallback about CONFIG_IKCONFIG_PROC, please?

I will update it.

-- 
Regards
Yafang