lkml.org 
[lkml]   [2003]   [Jan]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectProblem in IDE Disks cache handling in kernel 2.4.XX
Date

Hello to everybody and a Happy New Year.
Thanks a lot for all the great job you are all doing.

Having EXT2 file system corruption after suspend on a notebook I investigate
kernel code and it seems to me that IDE Disk cache handling is wrong in
kernel 2.4.XX.

Sources extracts are from kernel 2.4.20.

If you look at struct hd_driveid in hdreg.h you will find :

unsigned short command_set_1; /* (word 82) supported
* 6: look-ahead
* 5: write cache ==== this means the Disk has a disk cache =====

unsigned short command_set_2; /* (word 83)
* 13: FLUSH CACHE EXT ==== Those fields were RESERVED in ATA/ATAPI 5
* 12: FLUSH CACHE ==== and are used ONLY in ATA/ATAPI 6 ====
* 10: 48-bit Address Feature Set

unsigned short cfs_enable_1; /* (word 85)
* 6: look-ahead
* 5: write cache ==== this means the Disk cache is enabled or disabled
*/
unsigned short cfs_enable_2; /* (word 86)
* 13: FLUSH CACHE EXT ==== Those fields were RESERVED in ATA/ATAPI 5
* 12: FLUSH CACHE ==== and are used ONLY in ATA/ATAPI 6 ====
* 10: 48-bit Address Feature Set

In ide-disk.c you have

static int write_cache (ide_drive_t *drive, int arg)
{
struct hd_drive_task_hdr taskfile;
struct hd_drive_hob_hdr hobfile;
memset(&taskfile, 0, sizeof(struct hd_drive_task_hdr));
memset(&hobfile, 0, sizeof(struct hd_drive_hob_hdr));
taskfile.feature = (arg) ? SETFEATURES_EN_WCACHE : SETFEATURES_DIS_WCACHE;
taskfile.command = WIN_SETFEATURES;

if (!(drive->id->cfs_enable_2 & 0x3000)) <==== WRONG ! ONLY FOR ATA/ATAPI 6
return 1;

(void) ide_wait_taskfile(drive, &taskfile, &hobfile, NULL);
drive->wcache = arg;
return 0;
}


In fact for ATA/ATAPI 5 cfs_enable_2 has no meaning. The fields to test are
write cache bits in command_set_1 and cfs_enable_1.
And in both cases the FLUSH CACHE command ALWAYS EXISTS !
The test of cfs_enable_2 must only be used for ATA/ATAPI 6 drives to use
FLUSH CACHE or FLUSH CACHE EXT in case of 48 bit addressing mode.

And it seems to me that when an IDE drive has a cache enabled wcache must be
initialized to say so ? Or you have to do a STANDY or SLEEP before APM
suspend or power off to be sure that the cache has been written to the disk.

I had a look at patch 2.4.21pre3 and the code looks the same.

And by the way how are powered off the IDE drives ?
Because a FLUSH CACHE or STANDY or SLEEP is MANDATORY before powering off the
drive with cache enabled or you will enjoy lost data.

I am not on the list so thank you to CC me.

Best regards to all.

Francis Verscheure

-
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/

\
 
 \ /
  Last update: 2005-03-22 13:32    [W:0.043 / U:2.740 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site