On Tue, Feb 15, 2022 at 10:13:33PM +0100, Jason A. Donenfeld wrote: > During crng_init == 0, we never credit entropy in add_interrupt_ > randomness(), but instead dump it directly into the primary_crng. That's > fine, except for the fact that we then wind up throwing away that > entropy later when we switch to extracting from the input pool and > overwriting the primary_crng key. The two other early init sites -- > add_hwgenerator_randomness()'s use crng_fast_load() and add_device_ > randomness()'s use of crng_slow_load() -- always additionally give their > inputs to the input pool. But not add_interrupt_randomness(). > > This commit fixes that shortcoming by calling mix_pool_bytes() after > crng_fast_load() in add_interrupt_randomness(). That's partially > verboten on PREEMPT_RT, where it implies taking spinlock_t from an IRQ > handler. But this also only happens during early boot and then never > again after that. Plus it's a trylock so it has the same considerations > as calling crng_fast_load(), which we're already using. > > Cc: Theodore Ts'o <tytso@xxxxxxx> > Reviewed-by: Dominik Brodowski <linux@xxxxxxxxxxxxxxxxxxxx> > Suggested-by: Eric Biggers <ebiggers@xxxxxxxxxx> > Signed-off-by: Jason A. Donenfeld <Jason@xxxxxxxxx> > --- > v3 uses a trylock instead of a spinlock, just like all the other locks > taken in hard irq. (Incidentally, we're now talking about moving this > into the deferred stage, so that at can be a spinlock, but at least with > what we have here, this really must be a trylock.) This looks fine, though it's unfortunate that it has to be a trylock so this isn't guaranteed. Also, the commit message is a bit misleading because it talks about "overwriting" the primary_crng key, but at this point in the series the extracted entropy is still being XOR'd with the primary_crng key. It's not until the next patch that the key is simply overwritten. - Eric