Messages in this thread Patch in this message |  | | From | "Jason A. Donenfeld" <> | Subject | [PATCH] random: do not check for system_wq initialization in wait event | Date | Sat, 30 Apr 2022 21:56:20 +0200 |
| |
We previously checked for !system_wq, because when this was called at boot time, it could wind up hitting credit_entropy_bits()->crng_reseed(), which in turn would fire up numa pool initialization in a workqueue. However, we've now gotten rid of those numa pools, and hence do not make use of these workqueues during early init. So we can get rid of the check.
Cc: Dominik Brodowski <linux@dominikbrodowski.net> Fixes: a9412d510ab9 ("random: get rid of secondary crngs") Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> --- drivers/char/random.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/drivers/char/random.c b/drivers/char/random.c index 02e880a2c51e..77cee120974d 100644 --- a/drivers/char/random.c +++ b/drivers/char/random.c @@ -1162,9 +1162,8 @@ void add_hwgenerator_randomness(const void *buffer, size_t count, * CRNG_RESEED_INTERVAL has elapsed. */ wait_event_interruptible_timeout(random_write_wait, - !system_wq || kthread_should_stop() || - input_pool.entropy_count < POOL_MIN_BITS, - CRNG_RESEED_INTERVAL); + kthread_should_stop() || input_pool.entropy_count < POOL_MIN_BITS, + CRNG_RESEED_INTERVAL); mix_pool_bytes(buffer, count); credit_entropy_bits(entropy); } -- 2.35.1
|  |