hda interrupt timeouts on sa1110 + sa1111 board

hda interrupt timeouts on sa1110 + sa1111 board

[Kristoffer Ericson]

Greetings,

First off I havent bisected this yet, but sounds more than likely
that it comes with the sa1111.c changes. It works fine in v2.6.34 also.

As you can see it progresses a bit but keeps loosing interrupts.

Best wishes
Kristoffer

HP Jornada# boot
## Booting kernel from Legacy Image at c0000000 ...
   Image Name:   Linux-2.6.36-hpc+
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2185384 Bytes = 2.1 MiB
   Load Address: c0208000
   Entry Point:  c0208000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK
OK

Starting kernel ...

Uncompressing Linux... done, booting the kernel.
[    0.000000] Linux version 2.6.36-hpc+ (kristoffer at boggieman) (gcc version 4.2.2) #5 Mon Oct 25 22:35:50 CEST 2010
[    0.000000] CPU: StrongARM-1110 [6901b118] revision 8 (ARMv4), cr=c020717f
[    0.000000] CPU: VIVT data cache, VIVT instruction cache
[    0.000000] Machine: HP Jornada 720
[    0.000000] Memory policy: ECC disabled, Data cache writeback
[    0.000000] Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 32512
[    0.000000] Kernel command line: root=/dev/hda1 console=ttySA0,19200n8 console=tty1
[    0.000000] PID hash table entries: 512 (order: -1, 2048 bytes)
[    0.000000] Dentry cache hash table entries: 16384 (order: 4, 65536 bytes)
[    0.000000] Inode-cache hash table entries: 8192 (order: 3, 32768 bytes)
[    0.000000] Memory: 64MB 64MB = 128MB total
[    0.000000] Memory: 125568k/125568k available, 5504k reserved, 0K highmem
[    0.000000] Virtual kernel memory layout:
[    0.000000]     vector  : 0xffff0000 - 0xffff1000   (   4 kB)
[    0.000000]     fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
[    0.000000]     DMA     : 0xffc00000 - 0xffe00000   (   2 MB)
[    0.000000]     vmalloc : 0xc8800000 - 0xe8000000   ( 504 MB)
[    0.000000]     lowmem  : 0xc0000000 - 0xc8000000   ( 128 MB)
[    0.000000]     modules : 0xbf000000 - 0xc0000000   (  16 MB)
[    0.000000]       .init : 0xc0208000 - 0xc0222000   ( 104 kB)
[    0.000000]       .text : 0xc0222000 - 0xc060b000   (4004 kB)
[    0.000000]       .data : 0xc060c000 - 0xc062c180   ( 129 kB)
[    0.000000] Hierarchical RCU implementation.
[    0.000000] 	RCU-based detection of stalled CPUs is disabled.
[    0.000000] 	Verbose stalled-CPUs detection is disabled.
[    0.000000] NR_IRQS:120
[    0.000000] Console: colour dummy device 80x30
[    0.000000] console [tty1] enabled
[    0.000000] console [ttySA0] enabled
[   35.764397] Calibrating delay loop... 137.21 BogoMIPS (lpj=686080)
[   36.012168] pid_max: default: 32768 minimum: 301
[   36.041014] Mount-cache hash table entries: 512
[   36.070726] CPU: Testing write buffer coherency: ok
[   36.109241] NET: Registered protocol family 16
[   36.571019] SA1111 Microprocessor Companion Chip: silicon revision 1, metal revision 1
[   36.620753] sa1111-rab 0400: dmabounce: registered device
[   36.655131] sa1111-rab 0600: dmabounce: registered device
[   36.689427] sa1111-rab 0800: dmabounce: registered device
[   36.723702] sa1111-rab 0a00: dmabounce: registered device
[   36.758041] sa1111-rab 1800: dmabounce: registered device
[   36.876845] bio: create slab <bio-0> at 0
[   36.914978] NET: Registered protocol family 23
[   36.945301] cfg80211: Calling CRDA to update world regulatory domain
[   36.985597] Switching to clocksource oscr
[   37.136325] setting trigger mode 3 for irq 51 failed (sa1111_type_highirq+0x0/0x90)
[   37.182938] setting trigger mode 3 for irq 53 failed (sa1111_type_highirq+0x0/0x90)
[   37.229286] jornada720_pcmcia_configure_socket(): config socket 0 vcc 0 vpp 0
[   37.275407] setting trigger mode 3 for irq 52 failed (sa1111_type_highirq+0x0/0x90)
[   37.321979] setting trigger mode 3 for irq 54 failed (sa1111_type_highirq+0x0/0x90)
[   37.368329] jornada720_pcmcia_configure_socket(): config socket 1 vcc 0 vpp 0
[   37.415209] NET: Registered protocol family 2
[   37.442527] IP route cache hash table entries: 1024 (order: 0, 4096 bytes)
[   37.486373] TCP established hash table entries: 4096 (order: 3, 32768 bytes)
[   37.530504] TCP bind hash table entries: 4096 (order: 2, 16384 bytes)
[   37.570239] TCP: Hash tables configured (established 4096 bind 4096)
[   37.608734] TCP reno registered
[   37.627990] UDP hash table entries: 256 (order: 0, 4096 bytes)
[   37.663606] UDP-Lite hash table entries: 256 (order: 0, 4096 bytes)
[   37.702723] NET: Registered protocol family 1
[   37.731009] RPC: Registered udp transport module.
[   37.759616] RPC: Registered tcp transport module.
[   37.788218] RPC: Registered tcp NFSv4.1 backchannel transport module.
[   37.833065] NetWinder Floating Point Emulator V0.97 (double precision)
[   37.883605] squashfs: version 4.0 (2009/01/31) Phillip Lougher
[   37.922102] Installing knfsd (copyright (C) 1996 okir at monad.swb.de).
[   37.962784] JFFS2 version 2.2. (NAND) ? 2001-2006 Red Hat, Inc.
[   38.001582] msgmni has been set to 245
[   38.025283] io scheduler noop registered
[   38.049163] io scheduler deadline registered
[   38.075491] io scheduler cfq registered (default)
[   38.104754] jornada720_pcmcia_configure_socket(): config socket 1 vcc 33 vpp 33
[   38.363534] HP Jornada 700 series backlight driver
[   38.394112] Epson S1D13XXX FB Driver
[   38.416551] s1d13xxxfb: chip production id 4 = S1D13506
[   38.448228] s1d13xxxfb: chip revision 1
[   38.471602] s1d13xxxfb: regs mapped at 0xc8a00000, fb 512 KiB mapped at 0xc8900000
[   38.519420] jornada720_pcmcia_configure_socket(): config socket 1 vcc 33 vpp 33
[   38.563830] jornada720_pcmcia_configure_socket(): config socket 1 vcc 33 vpp 33
[   38.690487] Console: switching to colour frame buffer device 80x30
[   38.819934] fb0: S1D13xxx frame buffer device
[   38.881851] pcmcia_socket pcmcia_socket1: pccard: PCMCIA card inserted into slot 1
[   39.003855] pcmcia 1.0: pcmcia: registering new device pcmcia1.0 (IRQ: 49)
[   39.142892] Serial: SA11x0 driver
[   39.201346] sa11x0-uart.1: ttySA1 at MMIO 0x80010000 (irq = 15) is a SA1100
[   39.285795] sa11x0-uart.3: ttySA0 at MMIO 0x80050000 (irq = 17) is a SA1100
[   39.400914] loop: module loaded
[   39.459181] nbd: registered device at major 43
[   39.585156] Uniform Multi-Platform E-IDE driver
[   39.652332] ide-gd driver 1.18
[   39.741358] mice: PS/2 mouse device common for all mice
[   39.812717] input: HP Jornada 720 keyboard as /class/input/input0
[   39.891238] input: HP Jornada 7xx Touchscreen as /class/input/input1
[   39.970397] sa1100-rtc sa1100-rtc: warning: initializing default clock divider/trim value
[   40.093835] sa1100-rtc sa1100-rtc: rtc core: registered sa1100-rtc as rtc0
[   40.174572] IPv4 over IPv4 tunneling driver
[   40.238903] GRE over IPv4 tunneling driver
[   40.302103] TCP cubic registered
[   40.355806] NET: Registered protocol family 17
[   40.416527] NET: Registered protocol family 15
[   40.478279] IrCOMM protocol (Dag Brattli)
[   40.589404] lib80211: common routines for IEEE802.11 drivers
[   40.662432] jornada720_pcmcia_configure_socket(): config socket 1 vcc 33 vpp 0
[   40.767874] jornada720_pcmcia_configure_socket(): config socket 1 vcc 33 vpp 0
[   41.220300] hda: SanDisk SDCFX3-8192, CFA DISK drive
[   42.002086] ide0 at 0xc8840000-0xc8840007,0xc884000e on irq 49
[   42.071940] isa bounce pool size: 16 pages
[   42.133597] hda: max request size: 128KiB
[   42.191639] hda: 16007040 sectors (8195 MB) w/1KiB Cache, CHS=15880/16/63
[   72.340242] hda: lost interrupt
[  102.500136] hda: lost interrupt
[  132.580134] hda: lost interrupt
[  162.660132] hda: lost interrupt
[  192.740138] hda: lost interrupt
[  192.786589]  hda: hda1 hda2
[  192.837397] ide-cs 1.0: ide-cs: hda: Vpp = 0.0
[  192.892128] sa1100-rtc sa1100-rtc: setting system clock to 1970-01-01 00:02:32 UTC (152)
[  223.060133] hda: lost interrupt
[  253.140164] hda: lost interrupt
[  283.220131] hda: lost interrupt
[  313.300134] hda: lost interrupt
[  343.380134] hda: lost interrupt
[  373.460134] hda: lost interrupt
[  403.540134] hda: lost interrupt
[  433.620132] hda: lost interrupt
[  463.700134] hda: lost interrupt

hda interrupt timeouts on sa1110 + sa1111 board

[Russell King - ARM Linux]

On Mon, Oct 25, 2010 at 11:12:39PM +0200, Kristoffer Ericson wrote:
> Greetings,
> 
> First off I havent bisected this yet, but sounds more than likely
> that it comes with the sa1111.c changes. It works fine in v2.6.34 also.
> 
> As you can see it progresses a bit but keeps loosing interrupts.

I don't see anything in the sa1111 core, soc_common, or sa11* pcmcia
code which would cause this change of behaviour.  I think it may
require a bisect to find what change broke this.

> [   37.136325] setting trigger mode 3 for irq 51 failed (sa1111_type_highirq+0x0/0x90)
> [   37.182938] setting trigger mode 3 for irq 53 failed (sa1111_type_highirq+0x0/0x90)

These are fine - the soc_common code intentionally tries to set the type
to rising, then both edges - the idea being that we get rising edge if
the hardware doesn't support both edges.

> [   40.662432] jornada720_pcmcia_configure_socket(): config socket 1 vcc 33 vpp 0
> [   40.767874] jornada720_pcmcia_configure_socket(): config socket 1 vcc 33 vpp 0
> [   41.220300] hda: SanDisk SDCFX3-8192, CFA DISK drive
> [   42.002086] ide0 at 0xc8840000-0xc8840007,0xc884000e on irq 49
> [   42.071940] isa bounce pool size: 16 pages
> [   42.133597] hda: max request size: 128KiB
> [   42.191639] hda: 16007040 sectors (8195 MB) w/1KiB Cache, CHS=15880/16/63
> [   72.340242] hda: lost interrupt

And this suggests no interrupts what so ever are coming from the drive.

hda interrupt timeouts on sa1110 + sa1111 board

[Kristoffer Ericson]

On Wed, Oct 27, 2010 at 09:27:59PM +0100, Russell King - ARM Linux wrote:
> On Mon, Oct 25, 2010 at 11:12:39PM +0200, Kristoffer Ericson wrote:
> > Greetings,
> > 
> > First off I havent bisected this yet, but sounds more than likely
> > that it comes with the sa1111.c changes. It works fine in v2.6.34 also.
> > 
> > As you can see it progresses a bit but keeps loosing interrupts.
> 
> I don't see anything in the sa1111 core, soc_common, or sa11* pcmcia
> code which would cause this change of behaviour.  I think it may
> require a bisect to find what change broke this.

Filip Zyzniewski did a bisect of this issue a couple of weeks ago (for v2.6.34)
and made 3 patches to get it going again. With these it works for me again.
Unfortunantly they are reverts rather than fixes, so it doesnt solve much
with the current layout.

Ill just drop them to the mailinglist and we can take it from there.

> 
> > [   37.136325] setting trigger mode 3 for irq 51 failed (sa1111_type_highirq+0x0/0x90)
> > [   37.182938] setting trigger mode 3 for irq 53 failed (sa1111_type_highirq+0x0/0x90)
> 
> These are fine - the soc_common code intentionally tries to set the type
> to rising, then both edges - the idea being that we get rising edge if
> the hardware doesn't support both edges.
> 
> > [   40.662432] jornada720_pcmcia_configure_socket(): config socket 1 vcc 33 vpp 0
> > [   40.767874] jornada720_pcmcia_configure_socket(): config socket 1 vcc 33 vpp 0
> > [   41.220300] hda: SanDisk SDCFX3-8192, CFA DISK drive
> > [   42.002086] ide0 at 0xc8840000-0xc8840007,0xc884000e on irq 49
> > [   42.071940] isa bounce pool size: 16 pages
> > [   42.133597] hda: max request size: 128KiB
> > [   42.191639] hda: 16007040 sectors (8195 MB) w/1KiB Cache, CHS=15880/16/63
> > [   72.340242] hda: lost interrupt
> 
> And this suggests no interrupts what so ever are coming from the drive.