Re: [PATCH] Marvell SATA fixes v2

[Bogdan Costescu]

On Mon, 21 Nov 2005, Bogdan Costescu wrote:

> This was a SMP kernel running on an Intel PIV with HyperThreading 
> and with 2 disks attached to the controller. I'll try to get this 
> crash again with serial console...

I seem to have some problems getting the serial console to work, but 
in the mean time I have some more results, this time using a UP 
kernel (again without MSI and libata DEBUG):

1. several badblocks tests finished fine; the speed is also fine 
(about 50Mbytes/s both read and write as reported by both iostat 
during badblocks and bonnie++ on an ext2 FS).

2. I know that we are not yet talking about hotplugging, but this 
is what happens if you do it (unplug) :-)

Badness in mv_channel_reset at /root/sata_mv/sata_mv.c:1701 (Not 
tainted)
 [<f8955282>] mv_channel_reset+0xe3/0xed [sata_mv]
 [<f89552b0>] mv_stop_and_reset+0x24/0x2d [sata_mv]
 [<f8954ae0>] mv_host_intr+0x13d/0x17c [sata_mv]
 [<f8954b92>] mv_interrupt+0x73/0xeb [sata_mv]
 [<c0141f0d>] handle_IRQ_event+0x2e/0x5a
 [<c0141fb7>] __do_IRQ+0x7e/0xd7
 [<c0104e7e>] do_IRQ+0x4a/0x82
 =======================
 [<c01038ca>] common_interrupt+0x1a/0x20
 [<c023a35f>] acpi_safe_halt+0x22/0x2c
 [<c023a472>] acpi_processor_idle+0x109/0x27a
 [<c01010c1>] cpu_idle+0x40/0x58
 [<c03f873f>] start_kernel+0x15b/0x1b2
 [<c03f831d>] unknown_bootoption+0x0/0x1b6
Badness in __msleep at /root/sata_mv/sata_mv.c:1721 (Not tainted)
 [<f8955649>] __mv_phy_reset+0x390/0x405 [sata_mv]
 [<c0111923>] delay_tsc+0xb/0x13
 [<c01e7659>] __delay+0x9/0xa
 [<f8954ae0>] mv_host_intr+0x13d/0x17c [sata_mv]
 [<f8954b92>] mv_interrupt+0x73/0xeb [sata_mv]
 [<c0141f0d>] handle_IRQ_event+0x2e/0x5a
 [<c0141fb7>] __do_IRQ+0x7e/0xd7
 [<c0104e7e>] do_IRQ+0x4a/0x82
 =======================
 [<c01038ca>] common_interrupt+0x1a/0x20
 [<c023a35f>] acpi_safe_halt+0x22/0x2c
 [<c023a472>] acpi_processor_idle+0x109/0x27a
 [<c01010c1>] cpu_idle+0x40/0x58
 [<c03f873f>] start_kernel+0x15b/0x1b2
 [<c03f831d>] unknown_bootoption+0x0/0x1b6
ata4: no device found (phy stat 00000000)

The acpi_* stuff might be due to the fact that this interrupt is taken 
over first by ACPI and later assigned also to sata_mv; this happens 
only with the UP kernel, for the SMP kernel sata_mv did not share the 
interrupt.

--
Bogdan Costescu

IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
E-mail: Bogdan.Costescu@xxxxxxxxxxxxxxxxxxxxx
-
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/