Re: sshd segmentation fault on ia64 on 28-rc7-mmotm-081203 - bisected

[Franck Bui-Huu]

Hello,

[ Sorry for the late reply ]

Lee Schermerhorn wrote:
> On Wed, 2008-12-03 at 11:43 -0500, Lee Schermerhorn wrote:
>> I couldn't find anything on the list about this.
>>
>> I've built and booted 28-rc7-mmotm-081203 on both x86_64 and ia64.  I
>> can ssh into the x86_64 platform fine.  On the ia64, I'm seeing sshd
>> [from RHEL5 distro] seg faulting.  [Same sshd worked on the
>> 28-rc6-mmotm-081201 kernel.]
>>
>> Has anyone else seen this?
>>
> 
> I finally got around to bisecting this--sshd fails with SEGV on ia64--on
> mmotm-081208.  It appears to be caused by the patch duo:
> 
> do_mpage_readpage-remove-first_logical_block-parameter.patch
> do_mpage_readpage-remove-first_logical_block-parameter-fix.patch

Unfortunately I don't see why this patch causes such breakage.

Could you apply the following patch with the previous patch applied
and see if sshd still segfaults ?

It basically reverts:

   do_mpage_readpage-remove-first_logical_block-parameter-fix.patch

Therefore I can see which part of the patch is broken.

Also can you tell me what the block size used by the file system
hosting sshd ?

Thanks !

                Franck

--- 8< ---

diff --git a/fs/mpage.c b/fs/mpage.c
index ca33b28..b5ac1ff 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -163,10 +163,7 @@ map_buffer_to_page(struct page *page, struct buffer_head *bh, int page_block)
 */
 static int last_mapped_page(struct buffer_head *map_bh, struct page *page)
 {
-	pgoff_t map_index = map_bh->b_page->index;
-
-	return buffer_boundary(map_bh) &&
-		page->index >= map_index + (map_bh->b_size>>PAGE_CACHE_SHIFT);
+	return buffer_boundary(map_bh);
 }
 
 /*
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/