On Jun 29, 2012, at 10:25 PM, Tony Zelenoff wrote: > 30/06/12 12:20 AM, James Bottomley пишет: >> On Fri, 2012-06-29 at 20:13 +0000, Love, Robert W wrote: >>> On 12-06-26 05:17 AM, Tony Zelenoff wrote: >>> >>> Fix this message: >>> WARNING: drivers/scsi/fcoe/libfcoe.o(.init.text+0x21): Section mismatch >>> in reference from the function init_module() to the function >>> .exit.text:fcoe_transport_exit() >>> The function __init init_module() references >>> a function __exit fcoe_transport_exit(). >>> >>> Signed-off-by: Tony Zelenoff <antonz@xxxxxxxxxxxxx><mailto:antonz@xxxxxxxxxxxxx> >>> >>> >>> Thank you for the patch. However, this was fixed a few weeks ago. The patch is in James' 'for-next' branch here: http://git.kernel.org/?p=linux/kernel/git/jejb/scsi.git;a=commit;h=43b8a1613a1c8707d0e86316a7d6999fd767a5cf > MAke this patch from linux branch is faster than cloning scsi-next. :) Anyway, if it fixed - it is nice. > >>> I'm not sure if it's queue'd up for the current RC phase or not. I don't know how the scsi.git branches are managed. I think the patch should go for the current RC. >> >> Show me the bug it actually produces and I'll queue it for -rc ... > It mostly cosmetic issue, there is no reason to port it to RC branches. In some ways it is cosmetic, but the build warning results in people sending patches to fix it, because it is easy to fix (your patch is the third). It also could result in a crash, if someone built fcoe into the kernel and experienced the right kind of failure during initialization. That is because exit sections are discarded for entities built into the kernel since they can't be removed. I grant you, that is much less likely than seeing the build warnings simply because I expect that it is usually being built as a module. It could bite some embedded developer however. I am more concerned about the patch I sent on June 21st titled "scsi: Avoid NULL dereferences in scsi_cmd_to_driver" since it does fix a crashing kernel bug that is positively seen with a large number of luns. It was already discussed on the open-fcoe mailing list, but had no response on linux-scsi so far. -- Mark Rustad, LAN Access Division, Intel Corporation ��.n��������+%������w��{.n�����{������ܨ}���Ơz�j:+v�����w����ޙ��&�)ߡ�a����z�ޗ���ݢj��w�f
