]> git.proxmox.com Git - mirror_ubuntu-kernels.git/commit
Revert "firmware: add sanity check on shutdown/suspend"
authorLinus Torvalds <torvalds@linux-foundation.org>
Mon, 11 Sep 2017 04:19:06 +0000 (21:19 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Mon, 11 Sep 2017 04:19:06 +0000 (21:19 -0700)
commitf007cad159e99fa2acd3b2e9364fbb32ad28b971
tree3347b090b27c27082414070a9cbf08a7bb75cbc6
parent64414e5f9896805c2e80583345e9b1745be73aa9
Revert "firmware: add sanity check on shutdown/suspend"

This reverts commit 81f95076281fdd3bc382e004ba1bce8e82fccbce.

It causes random failures of firmware loading at resume time (well,
random for me, it seems to be more reliable for others) because the
firmware disabling is not actually synchronous with any particular
resume event, and at least the btusb driver that uses a workqueue to
load the firmware at resume seems to occasionally hit the "firmware
loading is disabled" logic because the firmware loader hasn't gotten the
resume event yet.

Some kind of sanity check for not trying to load firmware when it's not
possible might be a good thing, but this commit was not it.

Greg seems to have silently suffered the same issue, and pointed to the
likely culprit, and Gabriel C verified the revert fixed it for him too.

Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Pointed-at-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tested-by: Gabriel C <nix.or.die@gmail.com>
Cc: Luis R. Rodriguez <mcgrof@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Documentation/driver-api/firmware/request_firmware.rst
drivers/base/firmware_class.c