Re: sata_sil problems (lockup at boottime)

[Oliver Bausinger <bausi@bausi.org>]

On Sunday 05 December 2004 17:10, Bartlomiej Zolnierkiewicz wrote:
> On Sun, 5 Dec 2004 09:08:32 +0100, Oliver Bausinger <bausi@bausi.org> wrote:
> > Hi,
> >
> > On Sunday 05 December 2004 00:34, you wrote:
> > > Hello everyone,
> > >
> > > i'm having some really strange problems with the sata_sil driver.
> > >
> > > My hardware:
> > > Asus A7N8X Deluxe Rev 1.x with onboard Sil3112 Controller
> > > - Samsung SP1614C SATA Harddrive
> > > - for further testing : PATA to SATA Adapter that came with an ABIT
> > > Mainboard
> >
> > I have the same problem with this hardware. The harddisk is a SAMSUNG
> > SP1614C again and the controller is this one:
> >
> > 0000:00:0a.0 Unknown mass storage controller: Silicon Image, Inc.
> > (formerly CMD Technology Inc) SiI 3112 [SATALink/SATARaid] Serial ATA
> > Controller (rev 01)
> >         Subsystem: Silicon Image, Inc. (formerly CMD Technology Inc) SiI
> > 3112 [SATALink/SATARaid] Serial ATA Controller
> >         Flags: bus master, 66MHz, medium devsel, latency 32, IRQ 5
> >         I/O ports at a000 [size=8]
> >         I/O ports at 9800 [size=4]
> >         I/O ports at 9400 [size=8]
> >         I/O ports at 9000 [size=4]
> >         I/O ports at 8800 [size=16]
> >         Memory at e2000000 (32-bit, non-prefetchable)
> >
> > Trying to insert the sata_sil module on 2.6.{7,8,9,10-rc3} gives
> >
> > libata version 1.10 loaded.
> > sata_sil version 0.8
> > PCI: Found IRQ 5 for device 0000:00:0a.0
> > ata1: SATA max UDMA/100 cmd 0xD0800080 ctl 0xD080008A bmdma 0xD0800000
> > irq 5 ata2: SATA max UDMA/100 cmd 0xD08000C0 ctl 0xD08000CA bmdma
> > 0xD0800008 irq 5 irq 5: nobody cared!
> >  [<c012e5aa>] __report_bad_irq+0x2a/0x90
> >  [<c012df60>] handle_IRQ_event+0x30/0x70
> >  [<c012e69c>] note_interrupt+0x6c/0xd0
> >  [<c012e0d0>] __do_IRQ+0x130/0x160
> >  [<c01042b9>] do_IRQ+0x19/0x30
> >  [<c0102896>] common_interrupt+0x1a/0x20
> >  [<c0117ab0>] __do_softirq+0x30/0x90
> >  [<c0117b36>] do_softirq+0x26/0x30
> >  [<c012df24>] irq_exit+0x34/0x40
> >  [<c01042be>] do_IRQ+0x1e/0x30
> >  [<c0102896>] common_interrupt+0x1a/0x20
> >  [<c01005f0>] default_idle+0x0/0x30
> >  [<c0100613>] default_idle+0x23/0x30
> >  [<c010068a>] cpu_idle+0x3a/0x60
> >  [<c032678d>] start_kernel+0x16d/0x1b0
> >  [<c0326360>] unknown_bootoption+0x0/0x1b0
> > handlers:
> > [<d0962490>] (ata_interrupt+0x0/0x1c0 [libata])
> > Disabling IRQ #5
> > ata1: dev 0 cfg 49:2f00 82:346b 83:7f01 84:4003 85:3c68 86:3c01 87:4003
> > 88:20ff
> > ata1: dev 0 ATA, max UDMA7, 312581808 sectors: lba48
>
> Looks like some other device is using the same IRQ
> (it is not handled correctly and gets disabled).
>

No, IRQ 5 is unshared if I understand correctly. Look here:

root@camino:~# lspci -v |grep IRQ
        Flags: bus master, medium devsel, latency 32, IRQ 9
        Flags: bus master, medium devsel, latency 32, IRQ 9
        Flags: medium devsel, IRQ 9
        Flags: bus master, slow devsel, latency 32, IRQ 9
        Flags: bus master, 66MHz, medium devsel, latency 32, IRQ 5
        Flags: bus master, medium devsel, latency 32, IRQ 11
        Flags: bus master, medium devsel, latency 32, IRQ 10
        Flags: bus master, medium devsel, latency 64, IRQ 11

Only the SiI 3112 uses IRQ 5.

> > [The other possible thing to do is to use the siimage driver in
> > drivers/ide, but this one stopped working with 2.6.8 and further. The
> > siimage driver works quite good with 2.6.7, so this is another proof for
> > no hardware problems]
>
> Please try narrowing problem down to specific 2.6.7-bk snapshot.  Thanks.

Forget this. My config was broken as I didn't enable CONFIG_BLK_DEV_IDE_SATA.
That's why siimage loaded but did nothing.
I'm currently running 2.6.10-rc3 with siimage and it works.

If I really were a hardware problem with interrupt handling, the siimage 
driver should be broken as well? But it's not.

Greetings,
Bausi