Daniel Axtens <dja@xxxxxxxxxx> writes: > Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx> writes: > >> On Wed, May 15, 2019 at 03:35:51AM +1000, Daniel Axtens wrote: >>> >>> By all means disable vmx ctr if I don't get an answer to you in a >>> timeframe you are comfortable with, but I am going to at least try to >>> have a look. >> >> I'm happy to give you guys more time. How much time do you think >> you will need? >> > Give me till the end of the week: if I haven't solved it by then I will > probably have to give up and go on to other things anyway. So as you've hopefully seen, I've nailed it down and posted a patch. (http://patchwork.ozlabs.org/patch/1099934/) I'm also seeing issues with ghash with the extended tests: [ 7.582926] alg: hash: p8_ghash test failed (wrong result) on test vector 0, cfg="random: use_final src_divs=[<reimport>9.72%@+39832, <reimport>18.2%@+65504, <reimport,nosimd>45.57%@alignmask+18, <reimport,nosimd>15.6%@+65496, 6.83%@+65514, <reimport,nosimd>1.2%@+25, <reim" It seems to happen when one of the source divisions has nosimd and the final result uses the simd finaliser, so that's interesting. Regards, Daniel > > (FWIW, it seems to happen when encoding greater than 4 but less than 8 > AES blocks - in particular with both 7 and 5 blocks encoded I can see it > go wrong from block 4 onwards. No idea why yet, and the asm is pretty > dense, but that's where I'm at at the moment.) > > Regards, > Daniel > >> Thanks, >> -- >> Email: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx> >> Home Page: http://gondor.apana.org.au/~herbert/ >> PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
