From: Eric Sandeen <sandeen@sandeen.net>
To: "Darrick J. Wong" <darrick.wong@oracle.com>
Cc: sandeen@redhat.com, linux-xfs@vger.kernel.org
Subject: Re: [PATCH 10/10] misc: fix ubsan warnings
Date: Fri, 27 Oct 2017 11:24:52 -0500 [thread overview]
Message-ID: <a70d6f9a-cee6-1d06-6637-cea07feefae7@sandeen.net> (raw)
In-Reply-To: <20171027161433.GH5483@magnolia>
On 10/27/17 11:14 AM, Darrick J. Wong wrote:
> On Fri, Oct 27, 2017 at 08:48:07AM -0500, Eric Sandeen wrote:
...
>>> diff --git a/include/command.h b/include/command.h
>>> index fb3f5c7..59eab96 100644
>>> --- a/include/command.h
>>> +++ b/include/command.h
>>> @@ -25,7 +25,7 @@
>>> * not iterate the command args function callout and so can be used
>>> * for functions like "help" that should only ever be run once.
>>> */
>>> -#define CMD_FLAG_ONESHOT (1<<31)
>>> +#define CMD_FLAG_ONESHOT (1U<<31)
>>> #define CMD_FLAG_FOREIGN_OK (1<<30) /* command not restricted to XFS */
>>> #define CMD_FLAG_LIBRARY (1<<29) /* command provided by libxcmd */
>>
>> What complained about this? cmd flags is a signed int, right, so why is
>> this a problem? (And why are we only using the top 3 bits? And let's just do 1U
>> on every flag if it's really needed, but why is it needed? I must be missing
>> something.)
>
> I think the complaint is that "1" is signed int, and ubsan doesn't like
> us shifting a value bit into the sign bit.
Oh right, duh.
> But yeah, I suppose if we 'U'ize one of them we could do it for all.
>
> (Though really, why not start with the lower bits?)
Heh, so:
io/init.h:
#define CMD_NOFILE_OK (1<<0) /* command doesn't need an open file */
#define CMD_NOMAP_OK (1<<1) /* command doesn't need a mapped region */
#define CMD_FOREIGN_OK CMD_FLAG_FOREIGN_OK
include/command.h:
#define CMD_FLAG_ONESHOT (1<<31)
#define CMD_FLAG_FOREIGN_OK (1<<30) /* command not restricted to XFS */
#define CMD_FLAG_LIBRARY (1<<29) /* command provided by libxcmd */
What could possibly go wrong?
-Eric
next prev parent reply other threads:[~2017-10-27 16:24 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-10-26 22:14 [PATCH 00/10] xfsprogs: 4.14 rollup Darrick J. Wong
2017-10-26 22:14 ` [PATCH 01/10] db: increase metadump's default overly long extent discard threshold Darrick J. Wong
2017-10-27 0:03 ` Eric Sandeen
2017-10-27 0:12 ` Darrick J. Wong
2017-10-26 22:15 ` [PATCH 02/10] xfsprogs: explicitly cast troublesome types to match printf format specifiers Darrick J. Wong
2017-10-27 0:06 ` Eric Sandeen
2017-10-26 22:15 ` [PATCH 03/10] xfs_io: add new error injection knobs to inject command Darrick J. Wong
2017-10-27 0:09 ` Eric Sandeen
2017-10-26 22:15 ` [PATCH 04/10] xfs_repair: fix bag memory overwrite problems Darrick J. Wong
2017-10-27 0:49 ` Eric Sandeen
2017-10-26 22:15 ` [PATCH 05/10] xfs_repair: clear DAX flag from non-file inodes Darrick J. Wong
2017-10-27 2:01 ` Eric Sandeen
2017-10-26 22:15 ` [PATCH 06/10] xfs_repair: fix cowextsize field checking and repairing Darrick J. Wong
2017-10-27 2:06 ` Eric Sandeen
2017-10-27 16:17 ` Darrick J. Wong
2017-10-27 16:27 ` Eric Sandeen
2017-10-26 22:15 ` [PATCH 07/10] misc: enable ubsan if it's available Darrick J. Wong
2017-10-26 22:23 ` [PATCH v2] misc: enable ubsan if the builder wants it Darrick J. Wong
2017-10-26 22:15 ` [PATCH 08/10] misc: enable gcc/clang address sanitizer Darrick J. Wong
2017-10-26 22:24 ` [PATCH v2 " Darrick J. Wong
2017-10-26 22:15 ` [PATCH 09/10] misc: enable thread Darrick J. Wong
2017-10-26 22:24 ` [PATCH v2 09/10] misc: enable thread sanitizer if requested Darrick J. Wong
2017-10-26 22:15 ` [PATCH 10/10] misc: fix ubsan warnings Darrick J. Wong
2017-10-27 13:48 ` Eric Sandeen
2017-10-27 16:14 ` Darrick J. Wong
2017-10-27 16:24 ` Eric Sandeen [this message]
2017-10-27 16:19 ` [PATCH v2 " Darrick J. Wong
2017-10-26 22:32 ` [PATCH 00/10] xfsprogs: 4.14 rollup Goldwyn Rodrigues
2017-10-27 0:01 ` Eric Sandeen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=a70d6f9a-cee6-1d06-6637-cea07feefae7@sandeen.net \
--to=sandeen@sandeen.net \
--cc=darrick.wong@oracle.com \
--cc=linux-xfs@vger.kernel.org \
--cc=sandeen@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.