Re: block, bfq: NULL pointer dereference in bfq_rq_pos_tree_lookup()

Re: block, bfq: NULL pointer dereference in bfq_rq_pos_tree_lookup()

[Paolo Valente]

[-- Attachment #1: Type: text/plain, Size: 668 bytes --]

Hi Mike,
I've waited a little bit before replying, because I've worked on a dev
patch series, for debugging another crash. I'd like to use
this series for your failure too, as the OOPS you report
unfortunately does not ring any bell :(

So, could you please try to apply this patch series?  If it doesn't
apply, I'll rebase it.  If it applies, I hope it will either solve your
failure, or at least tell us more about the cause.

BTW, the patch "Makefile: add -bfq version suffix" may fail for a
conflict on kernel-version tags.  Fix the conflict as you like, or
just remove that patch if you don't want any bfq suffix in your kernel
name.

Crossing my fingers,
Paolo


[-- Attachment #2: dev-patches-on-for-next.tgz --]
[-- Type: application/octet-stream, Size: 42405 bytes --]

[-- Attachment #3: Type: text/plain, Size: 5658 bytes --]



> Il giorno 29 apr 2021, alle ore 12:59, Mike Galbraith <efault@gmx.de> ha scritto:
> 
> Build of a distro kernel in master.today got to linking phase, and went boom.
> 
> [ 2651.337220] BUG: kernel NULL pointer dereference, address: 0000000000000030
> [ 2651.337229] #PF: supervisor read access in kernel mode
> [ 2651.337231] #PF: error_code(0x0000) - not-present page
> [ 2651.337233] PGD 0 P4D 0
> [ 2651.337236] Oops: 0000 [#1] SMP NOPTI
> [ 2651.337239] CPU: 7 PID: 18691 Comm: ld Kdump: loaded Tainted: G            E     5.12.0.gd72cd4a-master #4
> [ 2651.337242] Hardware name: MEDION MS-7848/MS-7848, BIOS M7848W08.20C 09/23/2013
> [ 2651.337244] RIP: 0010:bfq_rq_pos_tree_lookup.isra.26+0x1e/0x90
> [ 2651.337249] Code: 00 e8 56 ff ff ff 31 c0 c3 0f 1f 00 0f 1f 44 00 00 53 45 31 d2 eb 07 49 8d 71 08 4d 89 ca 4c 8b 0e 4d 85 c9 74 12 49 8b 41 28 <48> 39 50 30 72 e7 76 2c 49 8d 71 10 eb e3 31 db 4d 85 c0 4c 89 11
> [ 2651.337253] RSP: 0018:ffff8881a0cefa40 EFLAGS: 00010086
> [ 2651.337255] RAX: 0000000000000000 RBX: 0000000000000000 RCX: ffff8881a0cefa60
> [ 2651.337257] RDX: 000000003785bc88 RSI: ffff8881044449b0 RDI: ffff88816131d800
> [ 2651.337259] RBP: ffff88816131d800 R08: 0000000000000000 R09: ffff8881e11bb288
> [ 2651.337261] R10: 0000000000000000 R11: 0000000f00002204 R12: ffff8881e11bb250
> [ 2651.337263] R13: ffff88816131d9f0 R14: 000000003785bc88 R15: ffff88810413f6b0
> [ 2651.337265] FS:  00007f73d1e070c0(0000) GS:ffff88840edc0000(0000) knlGS:0000000000000000
> [ 2651.337267] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [ 2651.337269] CR2: 0000000000000030 CR3: 00000001a20c8003 CR4: 00000000001706e0
> [ 2651.337271] Call Trace:
> [ 2651.337275]  bfq_setup_cooperator+0x23c/0x3c0
> [ 2651.337279]  ? blk_attempt_req_merge+0xa/0x30
> [ 2651.337283]  bfq_insert_requests+0x1bc/0x14c0
> [ 2651.337287]  ? submit_bio+0x35/0x100
> [ 2651.337290]  ? submit_bio+0x35/0x100
> [ 2651.337292]  blk_mq_sched_insert_requests+0x5b/0xf0
> [ 2651.337295]  blk_mq_flush_plug_list+0x104/0x180
> [ 2651.337298]  blk_flush_plug_list+0xb4/0xd0
> [ 2651.337302]  blk_finish_plug+0x27/0x40
> [ 2651.337304]  read_pages+0x106/0x1e0
> [ 2651.337308]  ? xas_load+0x8/0x80
> [ 2651.337311]  ? page_cache_ra_unbounded+0x113/0x1b0
> [ 2651.337314]  page_cache_ra_unbounded+0x113/0x1b0
> [ 2651.337317]  filemap_get_pages+0xb8/0x4a0
> [ 2651.337321]  filemap_read+0x99/0x2d0
> [ 2651.337323]  ? __alloc_pages_nodemask+0xff/0x260
> [ 2651.337326]  ? page_add_new_anon_rmap+0x50/0xf0
> [ 2651.337329]  ? __handle_mm_fault+0xaae/0x1060
> [ 2651.337333]  new_sync_read+0x105/0x180
> [ 2651.337336]  vfs_read+0x168/0x190
> [ 2651.337340]  ksys_read+0x7d/0xb0
> [ 2651.337343]  do_syscall_64+0x39/0x80
> [ 2651.337347]  entry_SYSCALL_64_after_hwframe+0x44/0xae
> [ 2651.337351] RIP: 0033:0x7f73d0ef00c1
> [ 2651.337354] Code: fe ff ff 48 8d 3d 27 a1 09 00 48 83 ec 08 e8 d6 03 02 00 66 0f 1f 44 00 00 8b 05 ea d2 2c 00 48 63 ff 85 c0 75 13 31 c0 0f 05 <48> 3d 00 f0 ff ff 77 57 f3 c3 0f 1f 44 00 00 55 53 48 89 d5 48 89
> [ 2651.337357] RSP: 002b:00007ffcd479d5f8 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
> [ 2651.337359] RAX: ffffffffffffffda RBX: 00005589d27348a0 RCX: 00007f73d0ef00c1
> [ 2651.337361] RDX: 0000000000001000 RSI: 00005589d273d9e0 RDI: 0000000000000007
> [ 2651.337363] RBP: 0000000000000d68 R08: 0000000000000000 R09: 0000000000000003
> [ 2651.337365] R10: ffffffffffffffb0 R11: 0000000000000246 R12: 00007f73d11b87a0
> [ 2651.337366] R13: 00007f73d11b92e0 R14: 0000000000000d68 R15: 0000000000000008
> [ 2651.337369] Modules linked in: fuse(E) msr(E) af_packet(E) xt_tcpudp(E) ip6t_REJECT(E) nf_reject_ipv6(E) ipt_REJECT(E) nf_reject_ipv4(E) xt_conntrack(E) nf_conntrack(E) nf_defrag_ipv6(E) nf_defrag_ipv4(E) libcrc32c(E) nfnetlink(E) bridge(E) stp(E) llc(E) iscsi_ibft(E) ip6table_mangle(E) iscsi_boot_sysfs(E) ip6table_raw(E) iptable_raw(E) ebtable_filter(E) ebtables(E) rfkill(E) ip6table_filter(E) ip6_tables(E) iptable_filter(E) ip_tables(E) x_tables(E) bpfilter(E) usblp(E) joydev(E) intel_rapl_msr(E) iTCO_wdt(E) at24(E) intel_pmc_bxt(E) intel_rapl_common(E) mei_hdcp(E) regmap_i2c(E) iTCO_vendor_support(E) snd_hda_codec_realtek(E) nls_iso8859_1(E) snd_hda_codec_generic(E) ledtrig_audio(E) nls_cp437(E) x86_pkg_temp_thermal(E) snd_hda_codec_hdmi(E) intel_powerclamp(E) coretemp(E) snd_hda_intel(E) kvm_intel(E) snd_intel_dspcfg(E) snd_hda_codec(E) kvm(E) snd_hwdep(E) snd_hda_core(E) irqbypass(E) crct10dif_pclmul(E) crc32_pclmul(E) snd_pcm(E) ghash_clmulni_intel(E) snd_timer(E) r8169(E)
> [ 2651.337399]  aesni_intel(E) snd(E) realtek(E) crypto_simd(E) mei_me(E) i2c_i801(E) mdio_devres(E) lpc_ich(E) cryptd(E) pcspkr(E) i2c_smbus(E) mei(E) libphy(E) soundcore(E) mfd_core(E) intel_smartconnect(E) fan(E) thermal(E) nfsd(E) auth_rpcgss(E) nfs_acl(E) lockd(E) sch_fq_codel(E) grace(E) sunrpc(E) nfs_ssc(E) hid_logitech_hidpp(E) hid_logitech_dj(E) uas(E) usb_storage(E) hid_generic(E) usbhid(E) nouveau(E) wmi(E) drm_ttm_helper(E) ttm(E) i2c_algo_bit(E) ahci(E) drm_kms_helper(E) syscopyarea(E) sysfillrect(E) sysimgblt(E) fb_sys_fops(E) cec(E) xhci_pci(E) ehci_pci(E) libahci(E) rc_core(E) xhci_hcd(E) ehci_hcd(E) libata(E) drm(E) usbcore(E) video(E) button(E) sd_mod(E) t10_pi(E) vfat(E) fat(E) virtio_blk(E) virtio_mmio(E) virtio_ring(E) virtio(E) ext4(E) crc32c_intel(E) crc16(E) mbcache(E) jbd2(E) loop(E) sg(E) dm_multipath(E) dm_mod(E) scsi_dh_rdac(E) scsi_dh_emc(E) scsi_dh_alua(E) scsi_mod(E) efivarfs(E) autofs4(E)
> [ 2651.337447] Dumping ftrace buffer:
> [ 2651.337450]    (ftrace buffer empty)
> [ 2651.337452] CR2: 0000000000000030
> 

Re: block, bfq: NULL pointer dereference in bfq_rq_pos_tree_lookup()

[Mike Galbraith]

On Mon, 2021-05-03 at 11:52 +0200, Paolo Valente wrote:
> Hi Mike,
> I've waited a little bit before replying, because I've worked on a dev
> patch series, for debugging another crash. I'd like to use
> this series for your failure too, as the OOPS you report
> unfortunately does not ring any bell :(
>
> So, could you please try to apply this patch series?  If it doesn't
> apply, I'll rebase it.

This bug isn't deterministic, but I can wedge your set into my devel
trees, and see if anything falls out.

	-Mike

Re: block, bfq: NULL pointer dereference in bfq_rq_pos_tree_lookup()

[Mike Galbraith]

On Mon, 2021-05-03 at 14:41 +0200, Mike Galbraith wrote:
> On Mon, 2021-05-03 at 11:52 +0200, Paolo Valente wrote:
> > Hi Mike,
> > I've waited a little bit before replying, because I've worked on a dev
> > patch series, for debugging another crash. I'd like to use
> > this series for your failure too, as the OOPS you report
> > unfortunately does not ring any bell :(
> >
> > So, could you please try to apply this patch series?  If it doesn't
> > apply, I'll rebase it.
>
> This bug isn't deterministic, but I can wedge your set into my devel
> trees, and see if anything falls out.

What fell out was not the least bit useful.  After days of box working
just fine despite bug being given ample enticement, it didn't take the
bait.  I then build master sans patch set, which exploded on its very
first distro build, after which bug (snickered mightily and) went back
to into hiding.

	-Mike

Re: block, bfq: NULL pointer dereference in bfq_rq_pos_tree_lookup()

[Paolo Valente]

> Il giorno 8 mag 2021, alle ore 06:49, Mike Galbraith <efault@gmx.de> ha scritto:
> 
> On Mon, 2021-05-03 at 14:41 +0200, Mike Galbraith wrote:
>> On Mon, 2021-05-03 at 11:52 +0200, Paolo Valente wrote:
>>> Hi Mike,
>>> I've waited a little bit before replying, because I've worked on a dev
>>> patch series, for debugging another crash. I'd like to use
>>> this series for your failure too, as the OOPS you report
>>> unfortunately does not ring any bell :(
>>> 
>>> So, could you please try to apply this patch series?  If it doesn't
>>> apply, I'll rebase it.
>> 
>> This bug isn't deterministic, but I can wedge your set into my devel
>> trees, and see if anything falls out.
> 
> What fell out was not the least bit useful.  After days of box working
> just fine despite bug being given ample enticement, it didn't take the
> bait.  I then build master sans patch set, which exploded on its very
> first distro build, after which bug (snickered mightily and) went back
> to into hiding.
> 

Not easy to debug this way :)

At any rate, I've just posted the fix contained in my debug patch series.

Thanks,
Paolo

> 	-Mike
>