Re: [PATCH v8] bio: limit bio max size

[Jon Hunter <jonathanh@nvidia.com>]

On 26/04/2021 14:18, Marek Szyprowski wrote:

...

> This patch landed in linux-next 20210426 as commit 42fb54fbc707 ("bio: 
> limit bio max size"). Sadly it causes the following regression during 
> boot on my test systems:
> 
> 8<--- cut here ---
> Unable to handle kernel NULL pointer dereference at virtual address 0000023c
> pgd = (ptrval)
> [0000023c] *pgd=00000000
> Internal error: Oops: 5 [#2] SMP ARM
> Modules linked in:
> CPU: 0 PID: 186 Comm: systemd-udevd Tainted: G      D 
> 5.12.0-next-20210426 #3045
> Hardware name: Generic DT based system
> PC is at bio_add_hw_page+0x58/0x1fc
> LR is at bio_add_pc_page+0x40/0x5c
> pc : [<c06c5bf0>]    lr : [<c06c5dd4>]    psr: 20000013
> sp : c3cc7de0  ip : ffffffff  fp : 00000000
> r10: 00000cc0  r9 : c20b2000  r8 : c21b5680
> r7 : dbc51b80  r6 : c30d0540  r5 : 00000014  r4 : c21b5680
> r3 : 00000000  r2 : 00000000  r1 : 00000000  r0 : c30d0540
> Flags: nzCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment none
> Control: 10c5387d  Table: 43ccc06a  DAC: 00000051
> Register r0 information: slab request_queue start c30d0540 pointer offset 0
> Register r1 information: NULL pointer
> Register r2 information: NULL pointer
> Register r3 information: NULL pointer
> Register r4 information: slab kmalloc-128 start c21b5680 pointer offset 
> 0 size 128
> Register r5 information: non-paged memory
> Register r6 information: slab request_queue start c30d0540 pointer offset 0
> Register r7 information: non-slab/vmalloc memory
> Register r8 information: slab kmalloc-128 start c21b5680 pointer offset 
> 0 size 128
> Register r9 information: slab kmalloc-4k start c20b2000 pointer offset 0 
> size 4096
> Register r10 information: non-paged memory
> Register r11 information: NULL pointer
> Register r12 information: non-paged memory
> Process systemd-udevd (pid: 186, stack limit = 0x(ptrval))
> Stack: (0xc3cc7de0 to 0xc3cc8000)
> ...
> [<c06c5bf0>] (bio_add_hw_page) from [<c06c5dd4>] (bio_add_pc_page+0x40/0x5c)
> [<c06c5dd4>] (bio_add_pc_page) from [<c06cf0ac>] 
> (blk_rq_map_kern+0x234/0x304)
> [<c06cf0ac>] (blk_rq_map_kern) from [<c0a54634>] (serial_show+0x64/0xd4)
> [<c0a54634>] (serial_show) from [<c0a228ac>] (dev_attr_show+0x18/0x48)
> [<c0a228ac>] (dev_attr_show) from [<c054721c>] (sysfs_kf_seq_show+0x88/0xf4)
> [<c054721c>] (sysfs_kf_seq_show) from [<c04d7a44>] 
> (seq_read_iter+0x10c/0x4bc)
> [<c04d7a44>] (seq_read_iter) from [<c04adf60>] (vfs_read+0x1d4/0x2e0)
> [<c04adf60>] (vfs_read) from [<c04ae47c>] (ksys_read+0x5c/0xd0)
> [<c04ae47c>] (ksys_read) from [<c03000c0>] (ret_fast_syscall+0x0/0x58)
> Exception stack(0xc3cc7fa8 to 0xc3cc7ff0)
> ...
> Code: e1520003 9a000021 e5942004 e5941020 (e592223c)
> ---[ end trace 51c4d8003ec70244 ]---


I have also noticed that an eMMC test we have started failing
today and bisect is pointing to this commit. Reverting this
change fixes it. The signature is a bit different to the
above, but nonetheless seems to be causing problems ...

[   76.675488] ------------[ cut here ]------------
[   76.680147] WARNING: CPU: 1 PID: 705 at /dvs/git/dirty/git-master_l4t-upstream/kernel/block/bio.c:1033 bio_iov_iter_get_pages+0x480/0x490
[   76.692518] Modules linked in: snd_soc_tegra30_i2s snd_soc_tegra_pcm snd_hda_codec_hdmi snd_soc_rt5640 snd_soc_tegra_rt5640 snd_soc_rl6231 snd_soc_tegra_utils snd_soc_core ac97_bus snd_pcm_dmaengine snd_soc_tegra30_ahub snd_hda_tegra snd_hda_codec snd_hda_core snd_pcm tegra_soctherm xhci_tegra snd_timer snd soundcore nouveau drm_ttm_helper ttm tegra30_devfreq tegra_wdt
[   76.725279] CPU: 1 PID: 705 Comm: dd Not tainted 5.12.0-next-20210426-g3f1fee3e7237 #1
[   76.733192] Hardware name: NVIDIA Tegra SoC (Flattened Device Tree)
[   76.739457] [<c0311628>] (unwind_backtrace) from [<c030bdd4>] (show_stack+0x10/0x14)
[   76.747203] [<c030bdd4>] (show_stack) from [<c0fe550c>] (dump_stack+0xc8/0xdc)
[   76.754423] [<c0fe550c>] (dump_stack) from [<c0345900>] (__warn+0x104/0x108)
[   76.761466] [<c0345900>] (__warn) from [<c03459b8>] (warn_slowpath_fmt+0xb4/0xbc)
[   76.768950] [<c03459b8>] (warn_slowpath_fmt) from [<c06cbb58>] (bio_iov_iter_get_pages+0x480/0x490)
[   76.777996] [<c06cbb58>] (bio_iov_iter_get_pages) from [<c0533864>] (iomap_dio_bio_actor+0x278/0x528)
[   76.787216] [<c0533864>] (iomap_dio_bio_actor) from [<c052f2d0>] (iomap_apply+0x170/0x440)
[   76.795476] [<c052f2d0>] (iomap_apply) from [<c053433c>] (__iomap_dio_rw+0x3f0/0x638)
[   76.803297] [<c053433c>] (__iomap_dio_rw) from [<c0534598>] (iomap_dio_rw+0x14/0x3c)
[   76.811043] [<c0534598>] (iomap_dio_rw) from [<c056347c>] (ext4_file_write_iter+0x550/0xa78)
[   76.819483] [<c056347c>] (ext4_file_write_iter) from [<c04b1700>] (vfs_write+0x2ec/0x3bc)
[   76.827662] [<c04b1700>] (vfs_write) from [<c04b1954>] (ksys_write+0xa8/0xd8)
[   76.834792] [<c04b1954>] (ksys_write) from [<c03000c0>] (ret_fast_syscall+0x0/0x58)
[   76.842435] Exception stack(0xc5471fa8 to 0xc5471ff0)
[   76.847485] 1fa0:                   0000006c 06400000 00000001 b0a2f000 06400000 00000000
[   76.855653] 1fc0: 0000006c 06400000 0050d0b8 00000004 0050d0b8 0050d0b8 00000000 00000000
[   76.863823] 1fe0: 00000004 beb0c9d0 b6ebdc0b b6e48206
[   76.868917] ---[ end trace d33cae3bcbc64fcb ]---


Cheers
Jon

-- 
nvpublic