linux-ext4 - RE: [PATCH 1/4] ext4: fix wrong return value in ext4_read_inode

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 for Android: free password hash cracker in your pocket

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

Message-ID: <3ED34739A4E85E4F894367D57617CDEF011BC6D559@LAX-EX-MB2.datadirect.datadirectnet.com>
Date:   Wed, 18 Apr 2018 05:31:12 +0000
From:   Wang Shilong <wshilong@....com>
To:     Andreas Dilger <adilger@...ger.ca>,
        Wang Shilong <wangshilong1991@...il.com>
CC:     Ext4 Developers List <linux-ext4@...r.kernel.org>,
        Liu Bo <bo.liu@...ux.alibaba.com>,
        Shuichi Ihara <sihara@....com>, "tytso@....edu" <tytso@....edu>
Subject: RE: [PATCH 1/4] ext4: fix wrong return value in
 ext4_read_inode_bitmap()

Hello Andreas,

   Yup, that one still need some work from myside, and I separate it from this patch series.
One thing I am still worried a bit, ted did you agree that we change ext4_error() to ext4_warning()
for bitmap errors in default(but of course error should be recored for next e2fsck)?

or as Andreas talked we might introduce another mount options?

Thanks,
Shilong

________________________________________
From: Andreas Dilger [adilger@...ger.ca]
Sent: Wednesday, April 18, 2018 13:19
To: Wang Shilong
Cc: Ext4 Developers List; Liu Bo; Wang Shilong; Shuichi Ihara; tytso@....edu
Subject: Re: [PATCH 1/4] ext4: fix wrong return value in ext4_read_inode_bitmap()

One thing that is (IMHO) missing from this patch series, is the last
patch that removes all of the calls to ext4_error() before calling
ext4_mark_group_bitmap_corrupted(), passing the error message as a
parameter, and cals save_error_info() in that function.

That would essentially change the code from aborting on any bitmap
error to marking the filesystem and block group in error (avoiding it
for any new allocations) but allows the filesystem to continue working.

We might consider to control this with a mount option, similar to the
"errors=" option, like "bitmap_error=error" or "bitmap_error=warning"
or something similar?

Cheers, Andreas