[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAG9oJsm7GA3A9MHcMiMqxzz5a9sV+S6BCzuYRbAVawweswSOAA@mail.gmail.com>
Date: Fri, 12 May 2023 19:59:25 +1000
From: Rudi Heitbaum <rudi@...tbaum.com>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
stable@...r.kernel.org, patches@...ts.linux.dev,
linux-kernel@...r.kernel.org, akpm@...ux-foundation.org,
linux@...ck-us.net, shuah@...nel.org, patches@...nelci.org,
lkft-triage@...ts.linaro.org, pavel@...x.de, jonathanh@...dia.com,
f.fainelli@...il.com, sudipm.mukherjee@...il.com,
srw@...dewatkins.net, rwarsow@....de, ntfs3@...ts.linux.dev,
almaz.alexandrovich@...agon-software.com
Subject: Re: [PATCH 6.3 000/694] 6.3.2-rc2 review
On Fri, 12 May 2023 at 00:24, Linus Torvalds
<torvalds@...ux-foundation.org> wrote:
>
> On Thu, May 11, 2023 at 3:30 AM Rudi Heitbaum <rudi@...tbaum.com> wrote:
> >
> > I have run 6.1.28-rc2 today, and was able to trigger the error. So
> > definitely bad in both 6.3 and 6.1.
> >
> > [13812.020209] BUG: kernel NULL pointer dereference, address: 0000000000000020
> > [13812.021322] #PF: supervisor read access in kernel mode
> > [13812.022346] #PF: error_code(0x0000) - not-present page
> > [13812.023591] PGD 0 P4D 0
> > [13812.024876] Oops: 0000 [#1] SMP NOPTI
> > [13812.026088] CPU: 5 PID: 20386 Comm: .NET ThreadPool Not tainted 6.1.28-rc2 #1
> > [13812.027336] Hardware name: Intel(R) Client Systems
> > NUC12WSKi7/NUC12WSBi7, BIOS WSADL357.0085.2022.0718.1739 07/18/2022
> > [13812.028593] RIP: 0010:ntfs_lookup+0x76/0xe0 [ntfs3]
>
> I suspect this is fixed in mainline by commit 6827d50b2c43 ("fs/ntfs3:
> Refactoring of various minor issues") which changed the IS_ERR() check
> into a IS_ERR_OR_NULL().
Hi Linus,
Thanks for the direction to the commit. I can confirm that once adding
commit 6827d50b2c43 to both 6.1.28-rc2 and 6.3.2-rc2 and testing for
today - that the error is not reproducible with the below lines.
- if (!IS_ERR(inode) && inode->i_op == NULL) {
+ if (!IS_ERR_OR_NULL(inode) && !inode->i_op) {
Rudi
> But dropping the original fix from stable might be the right thing to do.
>
> Linus
Powered by blists - more mailing lists