Re: AHCI SATA DVD writing issue

[Tejun Heo <htejun@gmail.com>]

Hello,

davor emard wrote:
> I've seen your mail in kernel ahci.c and I'm not sure if I addressed
> the right place to report thos but you are my starting point :-)

Yeap, it's the right place.

> I have sata dvd writer issues. Have tried pioneer dvr215 and optiarc 7173S
> connected on intel and marvell 6145 port. I updated the motherboard
> bios (that has some marvell firmware onboard) up to the latest.
> marvell sata port gave me the better results but not too good
> both sata's are ahci compliant, so using ahci.c driver in kernel.
> 
> I compiled 2.6.24.3 on amd64 but the similar problems existed even before
> on earlier kernels from 2.6.20 when when I started trying with this
> and now I think It's my duty to start reporting this at last
> 
> I tried scsi cdrom with and without support for vendor-specific ioctls
> doesn't matter
> 
> On intel sata hardware I cannot flash optiarc drive from linux
> (drive not detected).
> 
> On marvell 6145 sata flashing works to the optiarc drive
> (program autodetects the drive from sg and sr-mod driver).
> Flash program on linux also detects the pioneer but the linux
> flashing support is not stable to this program so I haven't tried it
> 
> I can also write correct iso image from pioneer and optiarc using
> growisofs -Z /dev/scd0=image.iso
> I can mount the disc with pioneer
> mount /dev/scd0 /mnt and it works
> I cannot do the mount with optiarc, mount waits a long
> time and in kernel log appears something like
> 
> Mar 19 01:18:32 emard kernel: ata6.00: exception Emask 0x0 SAct 0x0
> SErr 0x0 action 0x2 f
> Mar 19 01:18:32 emard kernel: ata6.00: cmd
> a0/00:00:00:08:00/00:00:00:00:00/a0 tag 0 pio
> Mar 19 01:18:32 emard kernel: cdb ac 00 00 00 00 00 00 00 00
> 01 03 00 00 00 00
> Mar 19 01:18:32 emard kernel: res
> 40/00:03:00:00:00/00:00:00:00:00/a0 Emask 0x4
> Mar 19 01:18:32 emard kernel: ata6.00: status: { DRDY }

0xAC is GET PERFORMANCE.  There's no in-kernel code which issues the 
command and it's not the most likely one to get issued during mount.

> and once this appears, both optiarc and pioneer connected refuse
> to work with messages like

That is really weird.  Ports on these SATA controllers are very 
independent from each other.  Unless IRQ delivery gets hosed by IRQ 
storm, failure on one port almost never affects others.

> Mar 19 01:18:32 emard kernel: ata6: hard resetting link
> Mar 19 01:18:34 emard kernel: ata6: failed to reset engine (errno=-5)
> Mar 19 01:18:34 emard kernel: ata6: SATA link up 1.5 Gbps (SStatus 113
> SControl 300)
> Mar 19 01:19:04 emard kernel: ata6.00: qc timeout (cmd 0xa1)
> Mar 19 01:19:04 emard kernel: ata6.00: failed to IDENTIFY (I/O error,
> err_mask=0x4)
> Mar 19 01:19:04 emard kernel: ata6.00: revalidation failed (errno=-5)
> Mar 19 01:19:04 emard kernel: ata6: failed to recover some devices,
> retrying in 5 secs
> Mar 19 01:19:09 emard kernel: ata6: hard resetting link
> 
> I need to reboot to get them "working" again

Can you please post full kernel log from boot through the initial and 
following failures?  Preferably with PRINTK_TIMESTAMP turned on but just 
attaching syslog file should do too.

> Also using the dvd from wodim is particularly very broken.
> if I try to autodetect pionner
> wodim dev=/dev/scd0 -scanbus
> it will produce in the log something like
> 
> Mar 19 01:18:32 emard kernel: ata6.00: exception Emask 0x0 SAct 0x0
> SErr 0x0 action 0x2 f
> Mar 19 01:18:32 emard kernel: ata6.00: cmd
> a0/00:00:00:08:00/00:00:00:00:00/a0 tag 0 pio
> Mar 19 01:18:32 emard kernel: cdb ac 00 00 00 00 00 00 00 00
> 01 03 00 00 00 00
> Mar 19 01:18:32 emard kernel: res
> 40/00:03:00:00:00/00:00:00:00:00/a0 Emask 0x4
> Mar 19 01:18:32 emard kernel: ata6.00: status: { DRDY }

0xAC again.

> and after a long time, wodim will even print correct device and
> capabilities of pioneer, but nothing else on sata will work from
> this point (unless I reboot again).
> 
> Any following wodim dev=/dev/scd0 -scanbus
> will report nothing and of course drives will not work from this point on
> 
> if I use wodim to autodetect optiarc
> wodim dev=/dev/scd1 -scanbus
> then it will report above sata exception but after some shorter
> pause then drive will be correctly detected with wodim gives me
> a hard scsi error report, but the drive will continue being autodetectable
> so following wodim dev=/dev/scd1 -scanbus will again report the optiarc
> device correctly
> 
> What do you say, is it easy to get rid of this ata exception errors that
> hose up the sata bus?

Can't really tell at this point.  Let's see.

Thanks.

-- 
tejun