Content-Type: multipart/mixed; boundary="===============2898254615109202021==" MIME-Version: 1.0 From: Thierry Reding To: lkp@lists.01.org Subject: Re: [futex] 8019ad13ef: will-it-scale.per_process_ops -97.8% regression Date: Wed, 11 Mar 2020 09:20:00 +0100 Message-ID: <20200311082000.GA724@ulmo> In-Reply-To: <87h7yy90ve.fsf@nanos.tec.linutronix.de> List-Id: --===============2898254615109202021== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable On Sun, Mar 08, 2020 at 07:07:17PM +0100, Thomas Gleixner wrote: > Linus Torvalds writes: > = > > [ Just a re-send without html crud that makes all the lists unhappy. > > I'm still on the road, the flight I was supposed to be on yesterday > > got cancelled.. ] > > > > I do note that the futex hashing seems to be broken by that commit. Or > > at least it's questionable. It keeps hashing on "both.word", and > > doesn't use the u64 field at all for hashing. > > > > Maybe I'm mis-reading it - I didn't apply the patch, I just looked at > > the patch and my source base separately. > > > > But the 98% regression sure says something went wrong ;) > = > Right you are. The pointer needs to be the starting point as it moved > ahead of word, which means it starts at word and hashes word and > offset and an extra u32 beyond the end of the key. > = > Thanks, > = > tglx > ---- > diff --git a/kernel/futex.c b/kernel/futex.c > index e14f7cd45dbd..9f3251349f65 100644 > --- a/kernel/futex.c > +++ b/kernel/futex.c > @@ -385,8 +385,8 @@ static inline int hb_waiters_pending(struct futex_has= h_bucket *hb) > */ > static struct futex_hash_bucket *hash_futex(union futex_key *key) > { > - u32 hash =3D jhash2((u32*)&key->both.word, > - (sizeof(key->both.word)+sizeof(key->both.ptr))/4, > + u32 hash =3D jhash2((u32*)&key->both.ptr, > + (sizeof(key->both.ptr) + sizeof(key->both.word)) / 4, > key->both.offset); > return &futex_queues[hash & (futex_hashsize - 1)]; > } I was running into an issue with sshfs mounts and git bisect pointed to the "futex: Fix inode life-time issue" commit as being the culprit. The issue I was seeing is that an sshfs mount would succeed, but trying to then access the mount ("ls -l /mnt" would be enough) would cause a soft lock-up. Applying this fix on top of next-20200310 fixes the issue for me, so: Tested-by: Thierry Reding I still need that fix on next-20200311, but it'll probably show up in linux-next sometime soon. Thierry --===============2898254615109202021== Content-Type: application/pgp-signature MIME-Version: 1.0 Content-Transfer-Encoding: base64 Content-Disposition: attachment; filename="signature.asc" LS0tLS1CRUdJTiBQR1AgU0lHTkFUVVJFLS0tLS0KCmlRSXpCQUFCQ0FBZEZpRUVpT3JEQ0FGSnpQ ZkFqY2lmM1NPczEzOCtzNkVGQWw1b255d0FDZ2tRM1NPczEzOCsKczZFOTh4QUFqQjVVbTE5N0U5 SkZ5aFp0UzZNRFF0WlBaZ3JGY2Fhc2F2aFVQdzloMTlobEVEaS9hZjlvR3JMTQpoSmQwS3Z5bWt3 L2NrWm1wTUdDamkrKzhyTkE4RFUxYlBybUxyNVFJRlM0aUNESlFWVFd5eG5kcUZtdVAwYWlGClps VDhyaThteTV5NWtmWHRiZjAvQ1NNTVV0MGExU1RIVksrUVc1ZkZSNXpVK29XSlNReWE2U1pXclF2 MHFZN1QKbGZkYjJNS0E4Uml1cjJFWWpaSDlEZHR0OTdBd0xVNnpZVW8yNjc5RVR5c1lYemdROFIr QXhYNGZZaUMrWjBjWgp2aXNYeHhoNkJLMk9MaGgrRFMyUnplYXhjNStiUTRmcHpBUFM5Q0NFdGNE WjNISGhOSDZRbVUzUWs0VWtyRVYyCktDaGFSdytEUUJWZ21UVFhPbllNVURXaTczZGRMZFk3b0lj OEZZMUNEcDhZSkluTzdFRk9MelBJQjhUekxiL3YKYUdQbVVYOGxYRlB2aHVJeXUzVDNQNVJrL0VI N2hoY1NQZ00wMWV3QVEyWE9OcThKL000VlRCU3g2aWpRaUl2Ugo2NUk5cXdhRHNVTGVRa1czeXVj MkJZMDMrM3dPM2xNWjRmSkRIK2NMV2NzQnZyczJiZEZDVSt1a1Y2Rk0wcldvCktRSDZVQzdOSytW TUhQSVhheHF0VEo5ZkJ1VVBtNVdDeE45eDlpMUIwK0Z0NUloYlZRVGdDYXp5R05sTFpvUGsKMFhW Tk9PcitJeHo2czloeGJLTXd3MjNPWlk2WWZMRCt2cGxKMlVoNlBVeWh0bzNOUENFWFZlVjRzZ0F2 aGxCVwpZakp0OFZJWUxwbmpTYzJKbnRYLzlFVWJ2c0R6ZnVLazV0ZUpLZHJpcGV1UXlvZGhXUXM9 Cj0zekplCi0tLS0tRU5EIFBHUCBTSUdOQVRVUkUtLS0tLQo= --===============2898254615109202021==--