Hey Jens, Please git pull the following branch: git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git stable/for-jens-3.13 which has a fix to the Xen block frontend driver. For the backends that don''t support persistent grants, we don''t want to use the logic that would use copying (needed for persistent grants). That logic existed there initially but with the persistent grants code added in - it was removed (so in 3.10 time-frame). This adds part of it back. Please pull! P.S. Sorry for the late pull! drivers/block/xen-blkfront.c | 125 ++++++++++++++++++++++++++++++++++--------- 1 file changed, 100 insertions(+), 25 deletions(-) Roger Pau Monne (1): xen-blkfront: restore the non-persistent data path
On 11/05/2013 09:59 AM, Konrad Rzeszutek Wilk wrote:> > Hey Jens, > > Please git pull the following branch: > > git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git stable/for-jens-3.13 > > > which has a fix to the Xen block frontend driver. For the backends that > don''t support persistent grants, we don''t want to use the logic that would > use copying (needed for persistent grants). That logic existed there initially > but with the persistent grants code added in - it was removed (so in 3.10 > time-frame). This adds part of it back. > > Please pull!Did you forget to publish it? -- Jens Axboe
On Tue, Nov 05, 2013 at 10:00:00AM -0700, Jens Axboe wrote:> On 11/05/2013 09:59 AM, Konrad Rzeszutek Wilk wrote: > > > > Hey Jens, > > > > Please git pull the following branch: > > > > git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git stable/for-jens-3.13 > > > > > > which has a fix to the Xen block frontend driver. For the backends that > > don''t support persistent grants, we don''t want to use the logic that would > > use copying (needed for persistent grants). That logic existed there initially > > but with the persistent grants code added in - it was removed (so in 3.10 > > time-frame). This adds part of it back. > > > > Please pull! > > Did you forget to publish it?No, but I gave you the wrong git tree <sad> git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip.git stable/for-jens-3.13 Sorry about that.> -- > Jens Axboe > > -- > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/
On 11/05/2013 10:17 AM, Konrad Rzeszutek Wilk wrote:> On Tue, Nov 05, 2013 at 10:00:00AM -0700, Jens Axboe wrote: >> On 11/05/2013 09:59 AM, Konrad Rzeszutek Wilk wrote: >>> >>> Hey Jens, >>> >>> Please git pull the following branch: >>> >>> git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git stable/for-jens-3.13 >>> >>> >>> which has a fix to the Xen block frontend driver. For the backends that >>> don''t support persistent grants, we don''t want to use the logic that would >>> use copying (needed for persistent grants). That logic existed there initially >>> but with the persistent grants code added in - it was removed (so in 3.10 >>> time-frame). This adds part of it back. >>> >>> Please pull! >> >> Did you forget to publish it? > > No, but I gave you the wrong git tree <sad> > > git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip.git stable/for-jens-3.13 > > Sorry about that.That did work a lot better. Thanks, pulled. -- Jens Axboe
On Tue, Nov 05, 2013 at 11:59:08AM -0500, Konrad Rzeszutek Wilk wrote:> > Hey Jens, > > Please git pull the following branch: > > git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git stable/for-jens-3.13 > > > which has a fix to the Xen block frontend driver. For the backends that > don''t support persistent grants, we don''t want to use the logic that would > use copying (needed for persistent grants). That logic existed there initially > but with the persistent grants code added in - it was removed (so in 3.10 > time-frame). This adds part of it back.I just tested this patch as part of a backported series on top of a 3.4.68 base. When this patch is included and a domU boots on a system that does not support persistent grants the kernel panics on boot with the trace below. Full boot log: http://paste.ubuntu.com/6367329/ It''s possible we missed something in the backport series. I''ll test this patch on top of 3.11.4 to see how it fares. general protection fault: 0000 [#1] SMP CPU 14 Modules linked in: aesni_intel cryptd aes_x86_64 aes_generic dm_mirror dm_region_hash dm_log dm_mod Pid: 2192, comm: blkid Not tainted 3.4.68-3.8.amzn1.x86_64 #1 Xen HVM domU RIP: 0010:[<ffffffff812dda71>] [<ffffffff812dda71>] get_grant+0x51/0x110 RSP: 0018:ffff881d07b5d978 EFLAGS: 00010006 RAX: 025c68fd93b116ed RBX: ffff880e72fc6b50 RCX: 0000000000000000 RDX: 1e8134cf7b38b76c RSI: 0000000001d1a05f RDI: ffff881d07b5da64 RBP: ffff881d07b5d998 R08: e200000000000000 R09: dead000000100100 R10: dead000000200200 R11: ffff8b9207578571 R12: 0000000001d1a05f R13: ffff880e72efc000 R14: ffff881d07b5da64 R15: ffff881d1822b0b0 FS: 00007fe197073760(0000) GS:ffff881d5bcc0000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fe196c4a3b0 CR3: 0000001d079c6000 CR4: 00000000000407e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process blkid (pid: 2192, threadinfo ffff881d07b5c000, task ffff881d07ae1690) Stack: 0000000000000246 ffff880e72963400 0000000000000000 ffff881d1822b0b0 ffff881d07b5da98 ffffffff812ddf60 0000000000000001 0000000000002000 00ff881d07b5d9e8 ffff881d07b5dfd8 ffff881d1822b000 ffff880e72d560b0 Call Trace: [<ffffffff812ddf60>] do_blkif_request+0x410/0x7c0 [<ffffffff811882b7>] ? submit_bh+0xf7/0x120 [<ffffffff81212026>] queue_unplugged+0x56/0xf0 [<ffffffff81216573>] blk_flush_plug_list+0x1c3/0x220 [<ffffffff812165e8>] blk_finish_plug+0x18/0x50 [<ffffffff8110ec11>] __do_page_cache_readahead+0x1c1/0x250 [<ffffffff8110efb1>] force_page_cache_readahead+0x71/0xa0 [<ffffffff8110f333>] page_cache_sync_readahead+0x43/0x50 [<ffffffff81104f58>] generic_file_aio_read+0x4d8/0x760 [<ffffffff811673e0>] ? do_last+0x470/0x8c0 [<ffffffff8118ed98>] blkdev_aio_read+0x58/0x80 [<ffffffff81158e1a>] do_sync_read+0xda/0x120 [<ffffffff811e22f3>] ? security_file_permission+0x93/0xb0 [<ffffffff811592a1>] ? rw_verify_area+0x61/0xf0 [<ffffffff81159780>] vfs_read+0xb0/0x180 [<ffffffff8115989a>] sys_read+0x4a/0x90 [<ffffffff813f0979>] system_call_fastpath+0x16/0x1b Code: 48 8d 82 b0 12 00 00 48 39 c3 0f 84 ca 00 00 00 49 b9 00 01 10 00 00 00 ad de 48 8b 13 49 ba 00 02 20 00 00 00 ad de 48 8b 43 08 <48> 89 42 08 48 89 10 44 8b 5b f0 4c 89 0b 4c 89 53 08 45 85 db RIP [<ffffffff812dda71>] get_grant+0x51/0x110 RSP <ffff881d07b5d978> ---[ end trace 3eba56aed2b98597 ]--- --msw> Please pull! > > P.S. > Sorry for the late pull! > > drivers/block/xen-blkfront.c | 125 ++++++++++++++++++++++++++++++++++--------- > 1 file changed, 100 insertions(+), 25 deletions(-) > > Roger Pau Monne (1): > xen-blkfront: restore the non-persistent data path >
On Tue, Nov 05, 2013 at 02:52:29PM -0800, Matt Wilson wrote:> On Tue, Nov 05, 2013 at 11:59:08AM -0500, Konrad Rzeszutek Wilk wrote: > > > > Hey Jens, > > > > Please git pull the following branch: > > > > git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git stable/for-jens-3.13 > > > > > > which has a fix to the Xen block frontend driver. For the backends that > > don''t support persistent grants, we don''t want to use the logic that would > > use copying (needed for persistent grants). That logic existed there initially > > but with the persistent grants code added in - it was removed (so in 3.10 > > time-frame). This adds part of it back. > > I just tested this patch as part of a backported series on top of a > 3.4.68 base. When this patch is included and a domU boots on a system > that does not support persistent grants the kernel panics on boot with > the trace below. Full boot log: http://paste.ubuntu.com/6367329/ > > It''s possible we missed something in the backport series. I''ll test > this patch on top of 3.11.4 to see how it fares.Hi Konrad, Looks like there''s something amiss in the backport. Testing your git branch worked fine. Sorry for the false alarm. --msw