linux-kernel - Re: [PATCH RESEND] scsi: ultrastor: Use correct format identifier for kernel pointer

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <20160501231831.GA17680@sophia>
Date:	Sun, 1 May 2016 19:18:31 -0400
From:	William Breathitt Gray <vilhelm.gray@...il.com>
To:	James Bottomley <jejb@...ux.vnet.ibm.com>
Cc:	martin.petersen@...cle.com, linux-scsi@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH RESEND] scsi: ultrastor: Use correct format identifier
 for kernel pointer

On Sun, May 01, 2016 at 04:08:25PM -0700, James Bottomley wrote:
>On Sun, 2016-05-01 at 17:57 -0400, William Breathitt Gray wrote:
>> The 'bios_segment' member of a struct ultrastor_config is passed to the
>> sprintf function with a respective %05X format identifier. The
>> 'bio_segment' member is a kernel pointer, but the %X format identifier
>> expects an int data type. A cast to int is correctly used to satisfy the
>> format identifier, but this assumes that the int data type is the same
>> size as the kernel pointer, which is not the case on several
>> architectures such as X86_64. This patch removes the int cast and
>> replaces the %05X format identifier with %pK in order to print the
>> 'bio_segment' member regardless of architecture.
>
>But this statement is wrong: ultrastor only has seven possible hard
>coded bios segment values, all under 20 bits, as it happens, that's why
>the printk is %05X.  Unless you're proposing to alter values in the
>static table, there's no point in changing the print, is there?  Are
>you proposing to alter the static table values?
>
>James

I'm sorry, I didn't realize bios_segment values were all under 20 bits;
in which case they should all fit within the original 5 digit
hexadecimal print. Please ignore this patch then as it was a
misunderstanding on my part.

Thanks,

William Breathitt Gray