linux-kernel - Re: [PATCH v3 00/16] Introduce a store type enum for the Maple tree

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <80926b22-a8d2-9992-eb5e-27e2c99cf460@google.com>
Date: Tue, 25 Jun 2024 22:37:25 -0700 (PDT)
From: Hugh Dickins <hughd@...gle.com>
To: Sidhartha Kumar <sidhartha.kumar@...cle.com>
cc: linux-kernel@...r.kernel.org, maple-tree@...ts.infradead.org, 
    linux-mm@...ck.org, akpm@...ux-foundation.org, liam.howlett@...cle.com, 
    willy@...radead.org, chuck.lever@...cle.com, hughd@...gle.com
Subject: Re: [PATCH v3 00/16] Introduce a store type enum for the Maple
 tree

On Tue, 18 Jun 2024, Sidhartha Kumar wrote:

> This series is rebased on top of mm-unstable + the patch:
> maple_tree: modified return type of mas_wr_store_entry()[1]. Andrew could
> you add that patch to mm-unstable before merging this series.
> 
> v2[2] -> v3:
>   - fix new line issues throughout the series
>   - remove use of helper function in patch 13 

Please give tmpfs a try on the latest mm-unstable, with
CONFIG_DEBUG_ATOMIC_SLEEP=y (maybe some of the messages below come from
other config options like PROVE_LOCKING, but ATOMIC_SLEEP the main one).

To the un-maple-trained eye, this series simply replaces a working maple
tree preallocation scheme by a broken one, doing GFP_KERNEL allocations
while holding spinlock.  But I doubt that was the intention: maybe a line
of code has gone missing or something, and you can quickly unbreak it.

 BUG: sleeping function called from invalid context at include/linux/sched/mm.h:337
 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 63, name: kdevtmpfs
 preempt_count: 1, expected: 0
 RCU nest depth: 0, expected: 0
 3 locks held by kdevtmpfs/63:
  #0: ffff8880008473f0 (sb_writers){.+.+}-{0:0}, at: mnt_want_write+0x19/0x40
  #1: ffff8880008a0888 (&type->i_mutex_dir_key/1){+.+.}-{3:3}, at: filename_create+0x8a/0x120
  #2: ffff8880008a0650 (&simple_offset_lock_class){+.+.}-{2:2}, at: mtree_alloc_cyclic+0x72/0xb0
 Preemption disabled at:
 [<ffffffff81138b94>] preempt_count_add+0x54/0x60
 CPU: 4 UID: 0 PID: 63 Comm: kdevtmpfs Not tainted 6.10.0-rc5-m25 #2
 Hardware name: LENOVO 20XXS3LA00/20XXS3LA00, BIOS N32ET91W (1.67 ) 02/02/2024
 Call Trace:
  <TASK>
  dump_stack_lvl+0x5d/0x80
  ? preempt_count_add+0x54/0x60
  dump_stack+0x10/0x20
  __might_resched+0x23b/0x260
  ? mas_alloc_nodes+0x71/0x160
  __might_sleep+0x56/0x60
  might_alloc+0x2a/0x40
  kmem_cache_alloc_noprof+0x28/0x190
  mas_alloc_nodes+0x71/0x160
  ? lock_is_held+0xc/0x10
  mas_node_count_gfp+0x2e/0x30
  mas_wr_preallocate+0x43/0x60
  mas_insert.isra.0+0x49/0xa0
  mas_alloc_cyclic+0x9c/0x100
  mtree_alloc_cyclic+0x92/0xb0
  simple_offset_add+0x3c/0x60
  shmem_mknod+0x55/0xb0
  vfs_mknod+0x9c/0xc0
  devtmpfs_work_loop+0x1c4/0x2a0
  ? trace_hardirqs_on+0x37/0x40
  ? _raw_spin_unlock_irqrestore+0x39/0x50
  ? complete_with_flags+0x40/0x50
  ? dmar_validate_one_drhd+0xa0/0xa0
  devtmpfsd+0x25/0x30
  kthread+0x100/0x110
  ? list_del_init+0x30/0x30
  ret_from_fork+0x22/0x40
  ? list_del_init+0x30/0x30
  ret_from_fork_asm+0x11/0x20

and lots more like that.

Thanks,
Hugh