Search String: Display: Description: Sort:

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*direct\-IO\s+writes\s+strange\s+behavior\s*$/: 30 ]

Total 30 documents matching your query.

1. direct-IO writes strange behavior (score: 1)
Author: Alberto Nava <beto@xxxxxxxxxxx>
Date: Fri, 21 Nov 2003 18:52:46 -0800
I'm seeing a very strange behavior when I write to a file using 512k direct-io writes. The filesystem is XFS and the kernel is 2.6.0-test9. The first time I do it the 512k writes are broken down int
/archives/xfs/2003-11/msg00183.html (9,314 bytes)

2. Re: direct-IO writes strange behavior (score: 1)
Author: Alberto Nava <beto@xxxxxxxxxxx>
Date: Sat, 22 Nov 2003 15:21:14 -0800
I've done some more digging on this issue. The reason the request is going in 4k pages is that the direct-io code is giving up in do_direct_IO() and the request is issued as buffer IO :-(. The reaso
/archives/xfs/2003-11/msg00186.html (11,799 bytes)

3. Re: direct-IO writes strange behavior (score: 1)
Author: Steve Lord <lord@xxxxxxx>
Date: Sun, 23 Nov 2003 11:35:38 -0600
I've done some more digging on this issue. The reason the request is going in 4k pages is that the direct-io code is giving up in do_direct_IO() and the request is issued as buffer IO :-(. The reason
/archives/xfs/2003-11/msg00189.html (11,907 bytes)

4. Re: direct-IO writes strange behavior (score: 1)
Author: Nathan Scott <nathans@xxxxxxx>
Date: Mon, 24 Nov 2003 14:38:18 +1100
Hi Alberto, Hmm... I don't see a way that we would pass back a buffer that is not mapped for a direct write. linvfs_get_block_core is the place to be concentrating if you're digging in the source, th
/archives/xfs/2003-11/msg00193.html (10,652 bytes)

5. Re: direct-IO writes strange behavior (score: 1)
Author: Alberto Nava <beto@xxxxxxxxxxx>
Date: Mon, 24 Nov 2003 09:55:14 -0800
This may be an interaction with the unwritten extent handling code. Try this on a file system made with the mkfs option -d unwritten=0. Thanks. I tried it but made not difference. We got an unmapped
/archives/xfs/2003-11/msg00194.html (9,229 bytes)

6. Re: direct-IO writes strange behavior (score: 1)
Author: Alberto Nava <beto@xxxxxxxxxxx>
Date: Mon, 24 Nov 2003 11:37:56 -0800
Alberto Nava wrote: This may be an interaction with the unwritten extent handling code. Try this on a file system made with the mkfs option -d unwritten=0. Thanks. I tried it but made not difference.
/archives/xfs/2003-11/msg00195.html (9,899 bytes)

7. Re: direct-IO writes strange behavior (score: 1)
Author: Steve Lord <lord@xxxxxxxx>
Date: 24 Nov 2003 14:36:24 -0600
unwritten = 0 was the only supported mode on linux (and originally irix) for a long time. The issue was that using preallocation meant you could go read old data off the disk. On linux now an unwritt
/archives/xfs/2003-11/msg00196.html (10,191 bytes)

8. Re: direct-IO writes strange behavior (score: 1)
Author: Nathan Scott <nathans@xxxxxxx>
Date: Thu, 27 Nov 2003 11:30:56 +1100
Hi Alberto, I spent some time this morning looking into this. Firstly, the fs/direct-io.c code has moved on from what you were looking at in test9 -- I've been testing using test11 this morning and t
/archives/xfs/2003-11/msg00232.html (13,799 bytes)

9. Re: direct-IO writes strange behavior (score: 1)
Author: Alberto Nava <beto@xxxxxxxxxxx>
Date: Wed, 26 Nov 2003 16:42:37 -0800
Nathan Scott wrote: Hi Alberto, I spent some time this morning looking into this. Firstly, the fs/direct-io.c code has moved on from what you were looking at in test9 -- I've been testing using test1
/archives/xfs/2003-11/msg00233.html (11,183 bytes)

10. Re: direct-IO writes strange behavior (score: 1)
Author: Andrew Morton <akpm@xxxxxxxx>
Date: Thu, 27 Nov 2003 00:44:20 -0800
You seem to be using a -mm kernel. It has extensive and not-quite-complete and not-really-tested-at-all-with-XFS direct-io changes. It would be best to concentrate on stock Linus direct-io for now -
/archives/xfs/2003-11/msg00235.html (9,260 bytes)

11. Re: direct-IO writes strange behavior (score: 1)
Author: Nathan Scott <nathans@xxxxxxx>
Date: Fri, 28 Nov 2003 08:18:27 +1100
hi Andrew, cheers. -- Nathan
/archives/xfs/2003-11/msg00239.html (9,712 bytes)

12. Re: direct-IO writes strange behavior (score: 1)
Author: Andrew Morton <akpm@xxxxxxxx>
Date: Thu, 27 Nov 2003 14:25:55 -0800
I know. But the code which he quoted had -mm stuff in it.
/archives/xfs/2003-11/msg00243.html (9,915 bytes)

13. Re: direct-IO writes strange behavior (score: 1)
Author: Alberto Nava <beto@xxxxxxxxxxx>
Date: Fri, 28 Nov 2003 10:10:44 -0800
Andrew Morton wrote: Alberto Nava <beto@xxxxxxxxxxx> wrote: I've done some more digging on this issue. The reason the request is going in 4k pages is that the direct-io code is giving up in do_direct
/archives/xfs/2003-11/msg00246.html (10,355 bytes)

14. Re: direct-IO writes strange behavior (score: 1)
Author: Alberto Nava <beto@xxxxxxxxxxx>
Date: Fri, 28 Nov 2003 17:14:09 -0800
Alberto Nava wrote: I'll try this again on test9 or test11. I tried this with the 2.6.0-test11 and it works fine. The 512k writes to a new file are propagaded all the way to the drives. It work witho
/archives/xfs/2003-11/msg00250.html (9,999 bytes)

15. Re: direct-IO writes strange behavior (score: 1)
Author: Nathan Scott <nathans@xxxxxxx>
Date: Sat, 29 Nov 2003 12:39:22 +1100
thanks for checking. no problem. cheers. -- Nathan
/archives/xfs/2003-11/msg00251.html (9,882 bytes)

16. direct-IO writes strange behavior (score: 1)
Author: Alberto Nava <beto@xxxxxxxxxxx>
Date: Fri, 21 Nov 2003 18:52:46 -0800
I'm seeing a very strange behavior when I write to a file using 512k direct-io writes. The filesystem is XFS and the kernel is 2.6.0-test9. The first time I do it the 512k writes are broken down int
/archives/xfs/2003-11/msg00444.html (9,314 bytes)

17. Re: direct-IO writes strange behavior (score: 1)
Author: Alberto Nava <beto@xxxxxxxxxxx>
Date: Sat, 22 Nov 2003 15:21:14 -0800
I've done some more digging on this issue. The reason the request is going in 4k pages is that the direct-io code is giving up in do_direct_IO() and the request is issued as buffer IO :-(. The reaso
/archives/xfs/2003-11/msg00447.html (11,799 bytes)

18. Re: direct-IO writes strange behavior (score: 1)
Author: Steve Lord <lord@xxxxxxx>
Date: Sun, 23 Nov 2003 11:35:38 -0600
I've done some more digging on this issue. The reason the request is going in 4k pages is that the direct-io code is giving up in do_direct_IO() and the request is issued as buffer IO :-(. The reason
/archives/xfs/2003-11/msg00450.html (11,907 bytes)

19. Re: direct-IO writes strange behavior (score: 1)
Author: Nathan Scott <nathans@xxxxxxx>
Date: Mon, 24 Nov 2003 14:38:18 +1100
Hi Alberto, Hmm... I don't see a way that we would pass back a buffer that is not mapped for a direct write. linvfs_get_block_core is the place to be concentrating if you're digging in the source, th
/archives/xfs/2003-11/msg00454.html (10,652 bytes)

20. Re: direct-IO writes strange behavior (score: 1)
Author: Alberto Nava <beto@xxxxxxxxxxx>
Date: Mon, 24 Nov 2003 09:55:14 -0800
This may be an interaction with the unwritten extent handling code. Try this on a file system made with the mkfs option -d unwritten=0. Thanks. I tried it but made not difference. We got an unmapped
/archives/xfs/2003-11/msg00455.html (9,229 bytes)


This search system is powered by Namazu