crypto: ccree - fix resume race condition on init
authorGilad Ben-Yossef <gilad@benyossef.com>
Thu, 7 Feb 2019 13:36:11 +0000 (15:36 +0200)
committerHerbert Xu <herbert@gondor.apana.org.au>
Fri, 8 Feb 2019 07:28:02 +0000 (15:28 +0800)
commit1358c13a48c43f5e4de0c1835291837a27b9720c
treee5ec75ed516856e6e08a1cbe1dc5f6415c20a810
parent356690d029e10f1aadebc49819d3908d5f6389fb
crypto: ccree - fix resume race condition on init

We were enabling autosuspend, which is using data set by the
hash module, prior to the hash module being inited, casuing
a crash on resume as part of the startup sequence if the race
was lost.

This was never a real problem because the PM infra was using low
res timers so we were always winning the race, until commit 8234f6734c5d
("PM-runtime: Switch autosuspend over to using hrtimers") changed that :-)

Fix this by seperating the PM setup and enablement and doing the
latter only at the end of the init sequence.

Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: stable@kernel.org # v4.20
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
drivers/crypto/ccree/cc_driver.c
drivers/crypto/ccree/cc_pm.c
drivers/crypto/ccree/cc_pm.h