Re: 5.2.0-rc1 UBIFS bad unlock balance

[Richard Weinberger <richard.weinberger@xxxxxxxxx>]

On Tue, Jun 4, 2019 at 10:21 PM Han Xu <han.xu@xxxxxxx> wrote:
>
> Hi Richard,
>
> Found such issue with lock debugging. Running bonnie++ on QSPI with UBIFS.
>
> Erasing 64 Kibyte @ 1ff0000 -- 100 % complete
> [  129.575631] ubi0: attaching mtd0
> [  129.620067] ubi0: scanning is finished
> [  129.623888] ubi0: empty MTD device detected
> [  129.723031] ubi0: attached mtd0 (name "30bb0000.spi", size 32 MiB)
> [  129.729289] ubi0: PEB size: 65536 bytes (64 KiB), LEB size: 65408 bytes
> [  129.735994] ubi0: min./max. I/O unit sizes: 1/256, sub-page size 1
> [  129.742267] ubi0: VID header offset: 64 (aligned 64), data offset: 128
> [  129.748915] ubi0: good PEBs: 512, bad PEBs: 0, corrupted PEBs: 0
> [  129.755026] ubi0: user volume: 0, internal volumes: 1, max. volumes count: 128
> [  129.762365] ubi0: max/mean erase counter: 0/0, WL threshold: 4096, image sequence number: 445061296
> [  129.771508] ubi0: available PEBs: 508, total reserved PEBs: 4, PEBs reserved for bad PEB handling: 0
> [  129.781182] ubi0: background thread "ubi_bgt0d" started, PID 516
> UBI device number 0, total 512 LEBs (33488896 bytes, 31.9 MiB), available 508 LEBs (33227264 bytes, 31.6 MiB), LEB size 65408 bytes (63.8 KiB)
> Set volume size to 33227264
> Volume ID 0, size 508 LEBs (33227264 bytes, 31.6 MiB), LEB size 65408 bytes (63.8 KiB), dynamic, name "test", alignment 1
> [  132.036227] UBIFS (ubi0:0): default file-system created
> [  132.046036] UBIFS (ubi0:0): Mounting in unauthenticated mode
> [  132.052526] UBIFS (ubi0:0): background thread "ubifs_bgt0_0" started, PID 524
> [  132.095082] UBIFS (ubi0:0): UBIFS: mounted UBI device 0, volume 0, name "test"
> [  132.102377] UBIFS (ubi0:0): LEB size: 65408 bytes (63 KiB), min./max. I/O unit sizes: 8 bytes/256 bytes
> [  132.111870] UBIFS (ubi0:0): FS size: 32573184 bytes (31 MiB, 498 LEBs), journal size 1635200 bytes (1 MiB, 25 LEBs)
> [  132.122423] UBIFS (ubi0:0): reserved for root: 1538511 bytes (1502 KiB)
> [  132.129123] UBIFS (ubi0:0): media format: w5/r0 (latest is w5/r0), UUID 2866BEC7-B74D-4793-8A66-E113C824711F, small LPT model
> Using uid:0, gid:0.
> Writing with putc()...done
> Writing intelligently...done
> Rewriting...done
> Reading with getc()...done
> Reading intelligently...done
> start 'em...done...done...done...
> Create files in sequential order...done.
> Stat files in sequential order...done.
> Delete files in sequential order...[  162.282301]
> [  162.283809] =====================================
> [  162.288514] WARNING: bad unlock balance detected!
> [  162.293221] 5.2.0-rc1-next-20190524-45490-g89d8c504132e-dirty #24 Not tainted
> [  162.300356] -------------------------------------
> [  162.305061] bonnie++/525 is trying to release lock (&(&c->orphan_lock)->rlock) at:
> [  162.312643] [<ffff0000104a0774>] ubifs_delete_orphan+0xec/0xfc
> [  162.318476] but there are no more locks to release!
> [  162.323353]
> [  162.323353] other info that might help us debug this:
> [  162.329881] 2 locks held by bonnie++/525:
> [  162.333890]  #0: 00000000a21aa66e (sb_writers#10){.+.+}, at: mnt_want_write+0x20/0x4c
> [  162.341734]  #1: 000000002fe597e5 (&c->commit_sem){++++}, at: make_reservation+0x84/0x3b4
> [  162.349919]
> [  162.349919] stack backtrace:
> [  162.354281] CPU: 1 PID: 525 Comm: bonnie++ Not tainted 5.2.0-rc1-next-20190524-45490-g89d8c504132e-dirty #24

I see the bug, thanks a lot for reporting!
But it is not in 5.2-rc1, the bad commit is so far only in linux-next.

-- 
Thanks,
//richard

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/