From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx2.suse.de ([195.135.220.15]:47774 "EHLO mx1.suse.de" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1726757AbeJ2H3N (ORCPT ); Mon, 29 Oct 2018 03:29:13 -0400 From: NeilBrown To: Jeff Layton , Alexander Viro Date: Mon, 29 Oct 2018 09:43:01 +1100 Cc: "J. Bruce Fields" , Martin Wilck , linux-fsdevel@vger.kernel.org, Frank Filz , linux-kernel@vger.kernel.org Subject: Re: [PATCH 0/9 v3] locks: avoid thundering-herd wake-ups In-Reply-To: <41db4227fdc5deb7703c6425c2f36589b355a7ca.camel@kernel.org> References: <154033435272.29542.13985568983074440924.stgit@noble> <41db4227fdc5deb7703c6425c2f36589b355a7ca.camel@kernel.org> Message-ID: <87y3ah3d6y.fsf@notabene.neil.brown.name> MIME-Version: 1.0 Content-Type: multipart/signed; boundary="=-=-="; micalg=pgp-sha256; protocol="application/pgp-signature" Sender: linux-fsdevel-owner@vger.kernel.org List-ID: --=-=-= Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Fri, Oct 26 2018, Jeff Layton wrote: > On Wed, 2018-10-24 at 09:43 +1100, NeilBrown wrote: >> This took longer that I had wanted, due to various reasons - sorry. >> And I'm now posting it in a merge window, which is not ideal. I don't >> expect it to be included in this merge window and I won't be at all >> impatient for review, but I didn't want to delay it further. >>=20 >> Testing found some problems, particularly the need to use >> locks_copy_lock in NFS. And there was a small thinko in there that >> effectively removed all the speed gains :-( >>=20 >> But this version: >> - shows excellent scalability with lots of threads on lots of CPUs >> contending on a single file >> - avoids the problems that Bruce found >> - seems to work. >>=20 >> Thanks, >> NeilBrown >>=20 >>=20 >> --- >>=20 >> NeilBrown (9): >> fs/locks: rename some lists and pointers. >> fs/locks: split out __locks_wake_up_blocks(). >> NFS: use locks_copy_lock() to copy locks. >> fs/locks: allow a lock request to block other requests. >> fs/locks: always delete_block after waiting. >> fs/locks: change all *_conflict() functions to return bool. >> fs/locks: create a tree of dependent requests. >> locks: merge posix_unblock_lock() and locks_delete_block() >> VFS: locks: remove unnecessary white space. >>=20 >>=20 >> fs/cifs/file.c | 4 - >> fs/lockd/svclock.c | 2=20 >> fs/locks.c | 231 ++++++++++++++++++++++----------= ------- >> fs/nfs/nfs4proc.c | 6 + >> fs/nfsd/nfs4state.c | 6 + >> include/linux/fs.h | 11 +- >> include/trace/events/filelock.h | 16 +-- >> 7 files changed, 153 insertions(+), 123 deletions(-) >>=20 >> -- >> Signature >>=20 > > > I built a kernel with these patches and ran the cthon04 lock tests and > got this on lock test 1 after a long hang (the test passed though): Thanks! I think locks_remove_posix() needs a call to locks_init_lock(&lock); just like locks_mandatory_area() has. I cannot immediately see anywhere else it is needed but I should probably look a bit harder. Thanks, NeilBrown > > [ 1694.787367] BUG: unable to handle kernel NULL pointer dereference at 0= 00000000000002c > [ 1694.789546] PGD 118ff0067 P4D 118ff0067 PUD 135915067 PMD 0=20 > [ 1694.790772] Oops: 0000 [#1] SMP NOPTI > [ 1694.791749] CPU: 7 PID: 1514 Comm: tlocklfs Not tainted 4.19.0+ #56 > [ 1694.792876] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIO= S ?-20180531_142017-buildhw-08.phx2.fedoraproject.org-1.fc28 04/01/2014 > [ 1694.795179] RIP: 0010:__locks_delete_block+0x14/0x90 > [ 1694.796203] Code: 01 a1 e9 9f 4f d8 ff 0f 1f 44 00 00 66 2e 0f 1f 84 0= 0 00 00 00 00 0f 1f 44 00 00 8b 05 29 9d 58 01 55 53 48 89 fb 85 c0 75 5a <= 48> 8b 43 20 48 85 c0 74 20 48 8b 53 18 48 89 10 48 85 d2 74 04 48 > [ 1694.799277] RSP: 0018:ffff9d21c1f63cb8 EFLAGS: 00010202 > [ 1694.800374] RAX: 0000000000000001 RBX: 000000000000000c RCX: aaaaaaaaa= aaaaaad > [ 1694.801682] RDX: 0000000000000001 RSI: ffffffff9f7b0c38 RDI: 000000000= 0000246 > [ 1694.802996] RBP: 000000000000000c R08: 0000000000000000 R09: 000000000= 0000001 > [ 1694.804317] R10: 0000000000000001 R11: ffffffffa0bdc188 R12: ffff9d21c= 1f63dd8 > [ 1694.805633] R13: ffff9d21c1f63e00 R14: ffffffff9f3241a8 R15: ffff8d0b5= aef72e0 > [ 1694.806941] FS: 00007efc8699c740(0000) GS:ffff8d0b7ba00000(0000) knlG= S:0000000000000000 > [ 1694.808380] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 > [ 1694.809550] CR2: 000000000000002c CR3: 000000011e0d8000 CR4: 000000000= 00006e0 > [ 1694.810888] Call Trace: > [ 1694.811692] __locks_wake_up_blocks+0x2d/0x80 > [ 1694.812713] locks_delete_block+0x1d/0x40 > [ 1694.813691] locks_lock_inode_wait+0x9c/0x1c0 > [ 1694.814731] nfs4_proc_lock+0x120/0x440 [nfsv4] > [ 1694.815786] ? nfs_put_lock_context+0x25/0x80 [nfs] > [ 1694.816866] ? do_unlk+0x98/0xe0 [nfs] > [ 1694.817818] locks_remove_posix+0xba/0x1d0 > [ 1694.818811] ? _cond_resched+0x15/0x30 > [ 1694.819768] ? wait_on_commit+0x38/0xb0 [nfs] > [ 1694.820787] ? process_echoes+0x60/0x60 > [ 1694.821752] ? __nfs_commit_inode+0xc2/0x1c0 [nfs] > [ 1694.822819] filp_close+0x56/0x70 > [ 1694.823712] __x64_sys_close+0x1e/0x50 > [ 1694.824661] do_syscall_64+0x60/0x1f0 > [ 1694.825599] entry_SYSCALL_64_after_hwframe+0x49/0xbe > [ 1694.826731] RIP: 0033:0x7efc8616c0a4 > [ 1694.827673] Code: eb 89 e8 af f6 01 00 66 2e 0f 1f 84 00 00 00 00 00 0= f 1f 44 00 00 8b 05 aa e7 2c 00 48 63 ff 85 c0 75 13 b8 03 00 00 00 0f 05 <= 48> 3d 00 f0 ff ff 77 44 f3 c3 66 90 48 83 ec 18 48 89 7c 24 08 e8 > [ 1694.830929] RSP: 002b:00007ffc70beb7b8 EFLAGS: 00000246 ORIG_RAX: 0000= 000000000003 > [ 1694.832371] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007efc8= 616c0a4 > [ 1694.833784] RDX: 0000000000000000 RSI: 00007efc864378a0 RDI: 000000000= 0000009 > [ 1694.835183] RBP: 00007ffc70beb7d0 R08: 00007efc864378a0 R09: 00007efc8= 699c740 > [ 1694.836560] R10: 00000000000006b4 R11: 0000000000000246 R12: 000000000= 0401000 > [ 1694.837941] R13: 00007ffc70beb990 R14: 0000000000000000 R15: 000000000= 0000000 > [ 1694.839322] Modules linked in: rpcsec_gss_krb5 nfsv4 dns_resolver nfs = fscache nfsd auth_rpcgss nfs_acl lockd grace ip6t_rpfilter ip6t_REJECT nf_r= eject_ipv6 xt_conntrack ip_set nfnetlink ebtable_nat ebtable_broute bridge = stp llc ip6table_nat nf_nat_ipv6 ip6table_mangle ip6table_raw ip6table_secu= rity iptable_nat nf_nat_ipv4 nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_i= pv4 iptable_mangle iptable_raw iptable_security ebtable_filter ebtables ip6= table_filter ip6_tables sunrpc joydev virtio_balloon i2c_piix4 pcspkr edac_= mce_amd xfs libcrc32c serio_raw virtio_blk qxl drm_kms_helper virtio_net tt= m net_failover virtio_console failover drm ata_generic pata_acpi floppy qem= u_fw_cfg > [ 1694.849736] CR2: 000000000000002c > [ 1694.850813] ---[ end trace da2f469c62deb564 ]--- > > --=20 > Jeff Layton --=-=-= Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iQIzBAEBCAAdFiEEG8Yp69OQ2HB7X0l6Oeye3VZigbkFAlvWO3UACgkQOeye3VZi gbmM5g/+KrVN+cHGw7EHL+PyftwVraBPSBtF5EQeyYbbaNOddZqL31TbMUTa/OpO LFRAXL1c71WMcu0o6r3TamPLH1N7HWOjkrEjocdTK+AYSjgnIeJa8FMIxK9h1ETU ySzH+yESiD29Ryr+ARFs00whrFREt+LRwLDq5mYM7pzXf8gmIW1IgSPQbhB2zimh aXn03Fy82Vx2w98DG4YPI3S+wnomrrxAgO0BUGVLFk9W17UQvw+O0095JGWlRJsE vtB34BCjz4zC4FhqsqicdEbPqhrPpomj1GFeGE8iFrUpZZPhUjLteS3etX0ZRdNB Y7Did8zzKenoLEgfaCDQFhJbwtO0VjvjB+UOZRJvsOXziO9sqfVc6V1HfbBsB9Bz i7wGD8iVyiTSdOGLvqoNS8f6ys6H2eFsm2zF3xtGSrC/4f9wUW8bhSfB6hQVhXJ8 uonfTEH/5IZMWwZvkSgsxCWM6Ib0IXjwZrC1q4TFK3vJiraKwFrJU8NQssg47S4q BZL8TDIlnXQsPu23O956uh9F5AVbgFE8ZC8n4wjUckB6Oxhtl/dVr3YLmu2ak0A7 sGbKSiptHiQ1TSqSGovrM+MDH9fO1uIBm7fzXIgaTcKzaM0TATeQqPFTOqdQ36Im zvM2mQGNJ0pvpKwOukKZ7A9PtMrAX4TnpBEWVLrNKlo42YQ9yT8= =nCQQ -----END PGP SIGNATURE----- --=-=-=--