]> git.ipfire.org Git - thirdparty/kernel/stable.git/commit
virtio: rng: delay hwrng_register() till driver is ready
authorAmit Shah <amit.shah@redhat.com>
Sat, 26 Jul 2014 22:04:01 +0000 (07:34 +0930)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 5 Oct 2014 20:41:08 +0000 (13:41 -0700)
commit5590b04b7a4fa2400f4d6ccec3c4d6166fab52cf
tree8ebec3b9aa63136159cc8c10755e8275c2759105
parent6ba3934df8b88f6593da479e1418c8219d4f5396
virtio: rng: delay hwrng_register() till driver is ready

commit 5c06273401f2eb7b290cadbae18ee00f8f65e893 upstream.

Instead of calling hwrng_register() in the probe routing, call it in the
scan routine.  This ensures that when hwrng_register() is successful,
and it requests a few random bytes to seed the kernel's pool at init,
we're ready to service that request.

This will also enable us to remove the workaround added previously to
check whether probe was completed, and only then ask for data from the
host.  The revert follows in the next commit.

There's a slight behaviour change here on unsuccessful hwrng_register().
Previously, when hwrng_register() failed, the probe() routine would
fail, and the vqs would be torn down, and driver would be marked not
initialized.  Now, the vqs will remain initialized, driver would be
marked initialized as well, but won't be available in the list of RNGs
available to hwrng core.  To fix the failures, the procedure remains the
same, i.e. unload and re-load the module, and hope things succeed the
next time around.

Signed-off-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/char/hw_random/virtio-rng.c