On Tue, Feb 01, 2022 at 08:13:14AM -1000, Tejun Heo wrote:

Hello,

On Fri, Jan 28, 2022 at 09:39:12AM +0200, Linus Torvalds wrote:

quoted

However, we've done this for *so* long that I wonder if there might be
situations that have ended up depending on the lack of synchronization
for pure performance reasons.

If *this* module loading process started the async work, then we'd
wait for it, but what if there's other async work that was started by
others? This revert would now make us wait for that async work too,
and that might be a big deal slowing things down at boot time.

Looking at it, this is all under the 'module_mutex', so I guess we are
already single-threaded at least wrt loading other modules, so the
amount of unrelated async work going on is presumably fairly low and
that isn't an issue.

Looks like we're multi-threaded while running the mod inits which launch the
async jobs and single-threaded while waiting for them to finish. Greg should
know a lot better than me but according to my hazy memory and cursory code
reading udev is multi-processed when loading modules, which makes it a lot
less likely that this will impact boot time in most cases.

I think userspace is multi-processed here, which should help with the
reading of the modules from disk at boot while others are actually being
loaded due to the kernel lock.

quoted

Anyway, I think this patch is the right thing to do, but just the fact
that we've avoided that async wait for so long makes me a bit nervous
about fallout from the revert.

Comments? Maybe this is a "just apply it, see if somebody screams" situation?

So, yeah, I think the risk is pretty low and even in the unlikely case that
someone is affected, the workaround is pretty straight-forward - not waiting
for the module loading to finish if appropriate.

I agree with Linus, let's see if anyone notices :)

thanks,

greg k-h