Re: [PATCH] btrfs: remove spurious WARN_ON(ref->count) in find_parent_nodes

[Edmund Nadolski <enadolski@suse.de>]

On 01/23/2018 12:36 AM, Lu Fengqi wrote:
> On Mon, Jan 22, 2018 at 08:35:43PM -0700, Edmund Nadolski wrote:
>> On 1/22/18 5:58 AM, Nikolay Borisov wrote:
>>>
>>>
>>> On 21.01.2018 21:08, Zygo Blaxell wrote:
>>>> This warning appears during execution of the LOGICAL_INO ioctl and
>>>> appears to be spurious:
>>>>
>>>> 	------------[ cut here ]------------
>>>> 	WARNING: CPU: 3 PID: 18172 at fs/btrfs/backref.c:1391 find_parent_nodes+0xc41/0x14e0
>>>> 	Modules linked in: ib_iser rdma_cm iw_cm ib_cm ib_core configfs iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi overlay r8169 ufs qnx4 hfsplus hfs minix ntfs vfat msdos fat jfs xfs cpuid rpcsec_gss_krb5 nfsv4 nfsv3 nfs fscache algif_skcipher af_alg softdog nfsd auth_rpcgss nfs_acl lockd grace sunrpc bnep cpufreq_userspace cpufreq_powersave cpufreq_conservative nfnetlink_queue nfnetlink_log nfnetlink bluetooth rfkill snd_seq_dummy snd_hrtimer snd_seq_midi snd_seq_oss snd_seq_midi_event snd_rawmidi snd_seq snd_seq_device binfmt_misc fuse nbd xt_REDIRECT nf_nat_redirect ipt_REJECT nf_reject_ipv4 xt_nat xt_conntrack xt_tcpudp nf_log_ipv4 nf_log_common xt_LOG ip6table_nat nf_conntrack_ipv6 nf_defrag_ipv6 nf_nat_ipv6 iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack
>>>> 	 ip6table_mangle iptable_mangle ip6table_filter ip6_tables iptable_filter ip_tables x_tables tcp_cubic dummy lp dm_crypt edac_mce_amd edac_core snd_hda_codec_hdmi ppdev kvm_amd kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel snd_hda_codec_via pcbc amdkfd snd_hda_codec_generic amd_iommu_v2 aesni_intel snd_hda_intel radeon snd_hda_codec aes_x86_64 snd_hda_core snd_hwdep crypto_simd glue_helper sg snd_pcm_oss cryptd input_leds joydev pcspkr serio_raw snd_mixer_oss rtc_cmos snd_pcm parport_pc parport shpchp wmi acpi_cpufreq evdev snd_timer asus_atk0110 k10temp fam15h_power snd soundcore sp5100_tco hid_generic ipv6 af_packet crc_ccitt raid10 raid456 async_raid6_recov async_memcpy async_pq async_xor async_tx libcrc32c raid0 multipath linear dm_mod raid1 md_mod ohci_pci ide_pci_generic
>>>> 	 sr_mod cdrom pdc202xx_new ohci_hcd crc32c_intel atiixp ehci_pci psmouse ide_core i2c_piix4 ehci_hcd xhci_pci mii xhci_hcd [last unloaded: r8169]
>>>> 	CPU: 3 PID: 18172 Comm: bees Tainted: G      D W    L  4.11.9-zb64+ #1
>>>> 	Hardware name: System manufacturer System Product Name/M5A78L-M/USB3, BIOS 2101    12/02/2014
>>>> 	Call Trace:
>>>> 	 dump_stack+0x85/0xc2
>>>> 	 __warn+0xd1/0xf0
>>>> 	 warn_slowpath_null+0x1d/0x20
>>>> 	 find_parent_nodes+0xc41/0x14e0
>>>> 	 __btrfs_find_all_roots+0xad/0x120
>>>> 	 ? extent_same_check_offsets+0x70/0x70
>>>> 	 iterate_extent_inodes+0x168/0x300
>>>> 	 iterate_inodes_from_logical+0x87/0xb0
>>>> 	 ? iterate_inodes_from_logical+0x87/0xb0
>>>> 	 ? extent_same_check_offsets+0x70/0x70
>>>> 	 btrfs_ioctl+0x8ac/0x2820
>>>> 	 ? lock_acquire+0xc2/0x200
>>>> 	 do_vfs_ioctl+0x91/0x700
>>>> 	 ? __fget+0x112/0x200
>>>> 	 SyS_ioctl+0x79/0x90
>>>> 	 entry_SYSCALL_64_fastpath+0x23/0xc6
>>>> 	RIP: 0033:0x7f727b20be07
>>>> 	RSP: 002b:00007f7279f1e018 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
>>>> 	RAX: ffffffffffffffda RBX: ffffffff9c0f4d7f RCX: 00007f727b20be07
>>>> 	RDX: 00007f7279f1e118 RSI: 00000000c0389424 RDI: 0000000000000003
>>>> 	RBP: 0000000000000035 R08: 00007f72581bf340 R09: 0000000000000000
>>>> 	R10: 0000000000000020 R11: 0000000000000246 R12: 0000000000000040
>>>> 	R13: 00007f725818d230 R14: 00007f7279f1b640 R15: 00007f7258000020
>>>> 	 ? trace_hardirqs_off_caller+0x1f/0x140
>>>> 	---[ end trace 5de243350f6762c6 ]---
>>>> 	------------[ cut here ]------------
>>>>
>>>> ref->count can be below zero under normal conditions (for delayed refs),
>>>> so there is no need to spam dmesg when it happens.
>>>
>>> Why do you think it's normal for this to be a negative value under
>>> normal conditions? There should be some rationale about that otherwise
>>> you are pampering over a bug.
>>
>>
>> The ref->count in the prelim_ref can be <0 for a delayed ref that
>> has a node->action of BTRFS_DROP_DELAYED_REF.  The prelim_ref_insert()
>> relies on this when merging identical refs to keep the overall
>> count correct.  So it looks to me like it should be OK to remove
>> the WARN.
> 
> The call graph of find_parent_nodes：
> add_delayed_refs
> add_inline_refs
> add_keyed_refs
> add_missing_keys
> -merge_refs (MERGE_IDENTICAL_KEYS)
> resolve_indirect_refs
> -merge_refs (MERGE_IDENTICAL_PARENTS)
> WARN_ON(ref->count < 0)
> 
> Yes, I agree that the ref->count in the prelim_ref can be less than 0
> between add_delayed_refs and add_inline_refs. However, prelim_ref_insert
> (or merge_refs before commit 86d5f9944252 ("btrfs: convert prelimary
> reference tracking to use rbtrees")) have merged all refs for the same
> block before this WARN_ON, so I'm still confused about why there is the
> independent negative delayed ref.

prelim_ref_insert() will merge only those refs which compare
identically. However any refs having e.g. different offsets [1]
would not be merged.

I agree with Nikolay that it's a good idea to explain that the
WARN is not needed for ref->count < 0.  I'd prefer to include
a few lines of code comments since a single line deletion
is harder to track via the history.

Ed

[1] see this patch set:
    https://www.spinics.net/lists/linux-btrfs/msg69472.html

which added the ignore_offset behavior to the backref search and gives
a case for multiple extent refs at different offsets.


>> (However the ref_mod in the btrfs_delayed_ref_node evidently cannot
>> go <0).
>>
>>
>>>> On kernel v4.14 this warning occurs 100-1000 times more frequently than
>>>> on kernels v4.2..v4.12.  In the worst case, one test machine had 59020
>>>> warnings in 24 hours on v4.14.14 compared to 55 on v4.12.14.
>>>>
>>>> Signed-off-by: Zygo Blaxell <ce3g8jdj@umail.furryterror.org>
>>>> ---
>>>>  fs/btrfs/backref.c | 1 -
>>>>  1 file changed, 1 deletion(-)
>>>>
>>>> diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
>>>> index 7d0dc100a09a..57e8d2562ed5 100644
>>>> --- a/fs/btrfs/backref.c
>>>> +++ b/fs/btrfs/backref.c
>>>> @@ -1263,7 +1263,6 @@ static int find_parent_nodes(struct btrfs_trans_handle *trans,
>>>>  	while (node) {
>>>>  		ref = rb_entry(node, struct prelim_ref, rbnode);
>>>>  		node = rb_next(&ref->rbnode);
>>>> -		WARN_ON(ref->count < 0);
>>>>  		if (roots && ref->count && ref->root_id && ref->parent == 0) {
>>>>  			if (sc && sc->root_objectid &&
>>>>  			    ref->root_id != sc->root_objectid) {
>>>>
>>
>> Reviewed-by: Edmund Nadolski <enadolski@suse.com>
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>>
>