Re: mmap problem in localedef

From: John David Anglin <dave.anglin@bell.net>
To: John David Anglin <dave.anglin@bell.net>
Cc: Aaro Koskinen <aaro.koskinen@iki.fi>,
	Helge Deller <deller@gmx.de>,
	Rolf Eike Beer <eike-kernel@sf-tec.de>,
	Meelis Roos <mroos@linux.ee>,
	"linux-parisc@vger.kernel.org List"
	<linux-parisc@vger.kernel.org>,
	Carlos O'Donell <carlos@redhat.com>
Subject: Re: mmap problem in localedef
Date: Sun, 2 Apr 2017 17:45:34 -0400	[thread overview]
Message-ID: <AF2BAE69-BC0C-4FDE-9B65-D15AC5FB0960@bell.net> (raw)
In-Reply-To: <C5E87DDC-4792-483B-835F-C396A4067174@bell.net>

On 2017-04-02, at 3:49 PM, John David Anglin wrote:

> On 2017-04-02, at 3:12 PM, Aaro Koskinen wrote:
> 
>> Hi,
>> 
>> On Sun, Apr 02, 2017 at 10:08:59AM -0400, John David Anglin wrote:
>>> On 2017-04-01, at 3:10 PM, Aaro Koskinen wrote:
>>>> It was attached to Dave's original mail. Inlined below:
>>>> 
>>>> Index: eglibc-2.18/sysdeps/unix/sysv/linux/hppa/bits/shm.h
>>>> ===================================================================
>>>> --- eglibc-2.18.orig/sysdeps/unix/sysv/linux/hppa/bits/shm.h	2013-01-09 19:28:48.000000000 -0500
>>>> +++ eglibc-2.18/sysdeps/unix/sysv/linux/hppa/bits/shm.h	2014-04-09 21:08:06.516202583 -0400
>>>> @@ -36,7 +36,7 @@
>>>> #define SHM_UNLOCK	12		/* unlock segment (root only) */
>>>> 
>>>> /* Segment low boundary address multiple.  */
>>>> -#define SHMLBA 0x00400000		/* address needs to be 4 Mb aligned */
>>>> +#define SHMLBA		(__getpagesize ())
>>>> 
>>>> /* Type to count number of attaches.  */
>>>> typedef unsigned long int shmatt_t;
>>> 
>>> This patch goes back to the following discussion:
>>> https://patchwork.kernel.org/patch/3933841/
>>> 
>>> As far as I can tell, it was not submitted to the glibc list, nor is
>>> there a bug report regarding localedef.
>> 
>> glibc bug report: https://sourceware.org/bugzilla/show_bug.cgi?id=19838
> 
> 
> Helge, would you send this change to glibc list?  If you already sent it, would you do a ping?

Actually, I see from further research that I'm to blame:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=725508

Your bug report was here:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=741243

> 
> You are in the best position to explain why this value should be page size and not 4MB.  It would be good to note
> in BZ where patch is on list.

Given the inequivalent alias messages which may have come from glibc testsuite, I'm not sure we have this right.
It seems like the kernel is not sharing pages in ld-2.24.so correctly.  The text and data in the file appear correctly
aligned on page boundaries.

Dave
--
John David Anglin	dave.anglin@bell.net