On Fri, Jul 6, 2012 at 3:15 AM, Henrik Rydberg [off-list ref] wrote:
On Thu, Jul 05, 2012 at 10:22:30AM -0700, Dmitry Torokhov wrote:
quoted
Hi Daniel, Henrik,
On Thu, Jun 28, 2012 at 09:08:05PM +0800, Daniel Kurtz wrote:
quoted
If sysfs entry creation fails, the driver is still usable, so don't
just abort probe. Just warn and continue.
I understand that I am a bit late to the party :) but I do not agree
with this change. Failure to create attributes is not sometihng that
user could cause (at least not easily) and thus would not be a setup
issue but something more severe. I believe we should fail loading the
driver so sysfs attribute breakage will be noticed as soon as possible,
instead of discovering it much much later in the process.
So I think we should revert this one.
Ok, I agree. Applied and pushed.
Daniel, I should probably have asked this before: what was the reason
for this patch in the first place?
Thanks,
Henrik
No additional reason beyond what is stated in the commit message -
that the driver is still usable without its sysfs entries, so no need
to abort probe.
I'm fine with reverting back to the fail fast approach, too.
Thanks,
Daniel