Re: [syzbot] KMSAN: uninit-value in asix_mdio_read (2)

On 12/19/21 00:42, Andrew Lunn wrote:

On Sun, Dec 19, 2021 at 12:14:30AM +0300, Pavel Skripkin wrote:

quoted

On 12/18/21 14:07, syzbot wrote:

quoted

Hello,

syzbot found the following issue on:

HEAD commit:    b0a8b5053e8b kmsan: core: add dependency on DEBUG_KERNEL
git tree:       https://github.com/google/kmsan.git master
console output: https://syzkaller.appspot.com/x/log.txt?x=13a4d133b00000
kernel config:  https://syzkaller.appspot.com/x/.config?x=46a956fc7a887c60
dashboard link: https://syzkaller.appspot.com/bug?extid=f44badb06036334e867a
compiler:       clang version 14.0.0 (/usr/local/google/src/llvm-git-monorepo 2b554920f11c8b763cd9ed9003f4e19b919b8e1f), GNU ld (GNU Binutils for Debian) 2.35.2
syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=149fddcbb00000
C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=17baef25b00000

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+f44badb06036334e867a@syzkaller.appspotmail.com

The last unhanded case is asix_read_cmd() == 0. Let's handle it...

That does not look correct, i think.

asix_read_cmd() == 0 means no error from the read itself. If there is
no error, we should look at the value of smsr and test for bit

As far as I understand asix_read_cmd() == 0 means 0 bytes was read, since it just returns usb_control_msg() return value. I don't know if usb_control_msg() can ever return 0, but looks like it's the only unhanded case

AX_HOST_EN. Doing a continue means we just ignore the value returned
by the good read.

I think the correct fix is to look at the value of i. If we have
exceeded 30, we should return -ETIMEDOUT.

Sorry, I don't get how it can help. Report says, that on following line

	else if (smsr & AX_HOST_EN)

smsr is uninitialized. How value of i can help here?

Can you, please, explain if you don't mind :)

Thanks


With regards,
Pavel Skripkin