[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20220623042150.762330-1-saravanak@google.com>
Date: Wed, 22 Jun 2022 21:21:50 -0700
From: Saravana Kannan <saravanak@...gle.com>
To: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
"Rafael J. Wysocki" <rafael@...nel.org>
Cc: Saravana Kannan <saravanak@...gle.com>, kernel-team@...roid.com,
linux-kernel@...r.kernel.org
Subject: [PATCH v1] driver core: Ignore driver_async_probe cmdline param for
module drivers
If the module's async_probe option isn't set, the module loading code
will do a async_synchronize_full() before returning to userspace. This
effectively negates any benefits of driver_async_probe listing the
module's driver.
If the module's async_probe is set, then we already do async probe for
that module's driver even if driver_async_probe does not list that
driver. So, again, the driver_async_probe's value doesn't matter for a
module's driver.
So this change ignores the driver_async_probe for module drivers to
avoid useless async probes.
In addition, if driver_async_probe lists a module's driver and the
driver's async probe ends up calling request_module() in the async
thread context, that's going to trigger a spurious WARNING stack dump
without any real benefits of async probing. So this will avoid that
unnecessary warning in that situation.
Signed-off-by: Saravana Kannan <saravanak@...gle.com>
---
drivers/base/dd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/base/dd.c b/drivers/base/dd.c
index e600dd2afc35..f1ac28a4ce62 100644
--- a/drivers/base/dd.c
+++ b/drivers/base/dd.c
@@ -823,7 +823,7 @@ bool driver_allows_async_probing(struct device_driver *drv)
return false;
default:
- if (cmdline_requested_async_probing(drv->name))
+ if (cmdline_requested_async_probing(drv->name) && drv->owner == NULL)
return true;
if (module_requested_async_probing(drv->owner))
--
2.37.0.rc0.161.g10f37bed90-goog
Powered by blists - more mailing lists