Re: DriveReady SeekComplete Error

[Bartlomiej Zolnierkiewicz]

On Friday 04 of June 2004 13:56, Jens Axboe wrote:
> On Fri, Jun 04 2004, Bartlomiej Zolnierkiewicz wrote:
> > On Friday 04 of June 2004 12:22, Jens Axboe wrote:
> > > damnit, don't trim the cc list!
> > >
> > > On Fri, Jun 04 2004, mattia wrote:
> > > > I have the following error (kernel 2.6.6):
> > > >
> > > > Jun  4 08:05:43 blink kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
> > > > Jun  4 08:05:43 blink kernel: hdc: Maxtor 6Y160P0, ATA DISK drive
> > > > Jun  4 08:05:43 blink kernel: hdd: Maxtor 6Y120L0, ATA DISK drive
> > > > Jun  4 08:05:43 blink kernel: ide1 at 0x170-0x177,0x376 on irq 15
> > > > Jun  4 08:05:43 blink kernel: hda: max request size: 128KiB
> > > > Jun  4 08:05:43 blink kernel: hda: 78177792 sectors (40027 MB)
> > > > w/1819KiB Cache, CHS=65535/16/63, UDMA(100)
> > > > Jun  4 08:05:43 blink kernel:  hda: hda1 hda2 hda3
> > > > Jun  4 08:05:43 blink kernel: hdc: max request size: 1024KiB
> > > > Jun  4 08:05:43 blink kernel: hdc: 320173056 sectors (163928 MB)
> > > > w/7936KiB Cache, CHS=19929/255/63, UDMA(100)
> > > > Jun  4 08:05:43 blink kernel:  hdc: hdc1
> > > > Jun  4 08:05:43 blink kernel: hdd: max request size: 128KiB
> > > > Jun  4 08:05:43 blink kernel: hdd: 240121728 sectors (122942 MB)
> > > > w/2048KiB Cache, CHS=65535/16/63, UDMA(100)
> > > > Jun  4 08:05:43 blink kernel:  hdd: hdd1 hdd2 hdd3
> > > > Jun  4 08:05:43 blink kernel: hdd: task_no_data_intr: status=0x51 {
> > > > DriveReady SeekComplete Error }
> > > > Jun  4 08:05:43 blink kernel: hdd: task_no_data_intr: error=0x04 {
> > > > DriveStatusError }
> > > > Jun  4 08:05:43 blink kernel: hdd: Write Cache FAILED Flushing!
> > > >
> > > >
> > > > I found somewhere that's something wrong with that maxtor drive.
> > > > However, everything works fine.
> > >
> > > There's nothing wrong with the drive technically, it's just odd (lba48
> > > without FLUSH_CACHE_EXT). It's really a linux ide bug that's fixed in
> > > newer kernels. 2.6.7 will fix your problem.
> >
> > Wrong.
> >
> > Bug is a combination of a very minor firmware quirk
> > and lack of strict checking in Linux IDE driver.
> >
> > FLUSH_CACHE_EXT bit is set but it is not supported
> > (but it is not a problem because LBA48 is not supported also).
>
> Ah my bad, I didn't realize this bit was actually set correctly (you
> mean (cfs_enable_2 & 0x2400) == 0x2400 is actually true?).

Yes but only on unaffected drives. ;-)

0x2000 is FLUSH_CACHE_EXT support
0x0400 is LBA48 support

Affected drives have 0x2000 set incorrectly so we have to check also 0x0400.
(== 0x2400 is needed because & 0x2400 is true if & 0x2000 OR if & 0x0400).

Everything is explained in the patch changelog:

| - many Maxtor disks incorrectly claim CACHE FLUSH EXT command support,
|   fix it by checking both CACHE FLUSH EXT command and LBA48 support
|   (thanks to Eric D. Mudama for help in fixing this)

and in the patch itself:

+/* some Maxtor disks have bit 13 defined incorrectly so check bit 10 too */
+#define ide_id_has_flush_cache_ext(id)	\
+	(((id)->cfs_enable_2 & 0x2400) == 0x2400)

> > It is fixed in 2.6.7-rc1 but your IDE barrier patch has this problem
> > (just reminding you that it is still not fixed in 2.6.7-rc2-mm2).
>
> So where's the bug? I don't see it...

ide_fill_flush_cmd()

+	if (drive->id->cfs_enable_2 & 0x2400)
+		rq->buffer[0] = WIN_FLUSH_CACHE_EXT;

Cheers,
Bartlomiej

-
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/