Hi,

I'm working on a custom hardware running Linux kernel 2.6.32.3. The hardware is based on the AT91SAM9263-ek evaluation board and have touch screen attached to it. The touch screen controller is ads7846 and is connected to SPI.

I'm in some trouble using the SPI (driver: atmel_spi).

ads7846 have two extra inputs that I want to use. The problem is that when reading these inputs from sysfs I get a lot of zeroes when it should read ( when no sensor is connected) ~3300 mV, e.g.
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0
~# cat /sys/bus/spi/devices/spi0.1/in0_input
3299 <-- This one is correct.
~# cat /sys/bus/spi/devices/spi0.1/in0_input
0

I have measured the input and it's stable at 3.3V.

Do anyone know why I read so many zeroes?


The second problem is a kernel krasch. The process adreader is reading the sys/bus/spi/devices/spi0.1/in0_input and /sys/bus/spi/devices/spi0.1/in1_input every second. After about 5 days the kernel krasched dumping the following message to the consol:

[425375.800000] Unable to handle kernel NULL pointer dereference at virtual address 00000016
[425375.810000] pgd = c3084000
[425375.810000] [00000016] *pgd=23051031, *pte=00000000, *ppte=00000000
[425375.810000] Internal error: Oops: 17 [#1] PREEMPT
[425375.810000] last sysfs file: /sys/devices/platform/atmel_spi.0/spi0.1/in0_input
[425375.810000] Modules linked in: atmel_pwm_bl
[425375.810000] CPU: 0 Not tainted (2.6.32.3E1219BSpitfire #1)
[425375.810000] PC is at atmel_spi_interrupt+0x64/0x218
[425375.810000] LR is at atmel_spi_interrupt+0x14/0x218
[425375.810000] pc : [<c017b188>] lr : [<c017b138>] psr: 20000013
[425375.810000] sp : c3065d70 ip : c3065db8 fp : c3065e4c
[425375.810000] r10: c389a350 r9 : c3064000 r8 : 00000000
[425375.810000] r7 : c389a2a0 r6 : c389a370 r5 : 00000000 r4 : c389a358
[425375.810000] r3 : 00000202 r2 : c4890000 r1 : c4890000 r0 : c389a358
[425375.810000] Flags: nzCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment user
[425375.810000] Control: 0005317f Table: 23084000 DAC: 00000015
[425375.810000] Process adreader (pid: 1050, stack limit = 0xc3064270)
[425375.810000] Stack: (0xc3065d70 to 0xc3066000)
[425375.810000] 5d60: 00000000 00000041 c03031dc c39036a0
[425375.810000] 5d80: 0000000e 00000001 00000000 00000000 00000002 c0061f34 c02ee444 c39036a0
[425375.810000] 5da0: 0000000e 00000001 00000000 c0064630 0000000e 00000000 0000000e c002506c
[425375.810000] 5dc0: c3065dd4 ffffffff fefff000 c0025aa8 7fffffff c3065ed0 c38603e0 00000000
[425375.810000] 5de0: 00000000 00000000 7fffffff c3064018 c3065ecc c3258c08 00000002 c3065e4c
[425375.810000] 5e00: c3065e50 c3065e18 c02317a8 c0230e18 60000013 ffffffff c39036a0 0000000e
[425375.810000] 5e20: 00000002 00000000 00000000 7fffffff 7fffffff c3065ecc c3258c08 00000002
[425375.810000] 5e40: c3065ec4 c3065e50 c02317a8 c0230e14 23914000 00000018 c4890000 c3064000
[425375.810000] 5e60: c3258bec 60000013 c3258c08 c389a358 c389a0a0 00000000 00000000 c3064000
[425375.810000] 5e80: 7fffffff c023155c c002bcd4 00000001 c38603e0 c0039790 c3065ed0 c3065ed0
[425375.810000] 5ea0: c3258c08 c3065ecc 00000000 00000002 c389a0a0 c3258c00 c3add800 c3065ec8
[425375.810000] 5ec0: c017a124 c023146c c02ee994 00000001 c3065e98 c3065e98 c3258cbc c3258ce0
[425375.810000] 5ee0: c3258d04 c01a3a5c c39cd560 c389a0a0 c3add800 c3227000 c02f87f0 c3065f88
[425375.810000] 5f00: c39cd578 00000000 00001fff c01a3b34 c02fdf74 c39cd560 c3adcbd8 c0157d34
[425375.810000] 5f20: 00001fff c389a0a8 c39cd560 c00cdc08 c39cd560 00000000 c02f1dd4 00019258
[425375.810000] 5f40: 00000000 c32c0640 00019258 c3065f88 00000003 00001fff c3064000 00000000
[425375.810000] 5f60: be992bc4 c008bb0c 00000000 c32c0640 00000000 00000000 c32c0640 00000003
[425375.810000] 5f80: c0026048 c008befc 00000000 00000000 00017410 00000001 be992a9c 00001fff
[425375.810000] 5fa0: 00019258 c0025ea0 be992a9c 00001fff 00000007 00019258 00001fff fbad2488
[425375.810000] 5fc0: be992a9c 00001fff 00019258 00000003 be992a9c 00000f30 4013a000 be992bc4
[425375.810000] 5fe0: 00000000 be992980 40068aa0 40069084 80000010 00000007 00000000 00000000
[425375.810000] [<c017b188>] (atmel_spi_interrupt+0x64/0x218) from [<c0061f34>] (handle_IRQ_event+0x40/0x10c)
[425375.810000] [<c0061f34>] (handle_IRQ_event+0x40/0x10c) from [<c0064630>] (handle_level_irq+0xa0/0x168)
[425375.810000] [<c0064630>] (handle_level_irq+0xa0/0x168) from [<c002506c>] (asm_do_IRQ+0x6c/0x98)
[425375.810000] [<c002506c>] (asm_do_IRQ+0x6c/0x98) from [<c0025aa8>] (__irq_svc+0x48/0x8c)
[425375.810000] Exception stack(0xc3065dd0 to 0xc3065e18)
[425375.810000] 5dc0: 7fffffff c3065ed0 c38603e0 00000000
[425375.810000] 5de0: 00000000 00000000 7fffffff c3064018 c3065ecc c3258c08 00000002 c3065e4c
[425375.810000] 5e00: c3065e50 c3065e18 c02317a8 c0230e18 60000013 ffffffff
[425375.810000] [<c0025aa8>] (__irq_svc+0x48/0x8c) from [<c0230e18>] (schedule+0x14/0x370)
[425375.810000] [<c0230e18>] (schedule+0x14/0x370) from [<c02317a8>] (schedule_timeout+0x18/0x1c4)
[425375.810000] [<c02317a8>] (schedule_timeout+0x18/0x1c4) from [<c023155c>] (wait_for_common+0x100/0x1cc)
[425375.810000] [<c023155c>] (wait_for_common+0x100/0x1cc) from [<c017a124>] (spi_sync+0x44/0x58)
[425375.810000] [<c017a124>] (spi_sync+0x44/0x58) from [<c01a3a5c>] (ads7846_read12_ser+0x160/0x1b8)
[425375.810000] [<c01a3a5c>] (ads7846_read12_ser+0x160/0x1b8) from [<c01a3b34>] (in0_input_show+0x20/0x48)
[425375.810000] [<c01a3b34>] (in0_input_show+0x20/0x48) from [<c0157d34>] (dev_attr_show+0x24/0x4c)
[425375.810000] [<c0157d34>] (dev_attr_show+0x24/0x4c) from [<c00cdc08>] (sysfs_read_file+0xb4/0x130)
[425375.810000] [<c00cdc08>] (sysfs_read_file+0xb4/0x130) from [<c008bb0c>] (vfs_read+0xac/0x158)
[425375.810000] [<c008bb0c>] (vfs_read+0xac/0x158) from [<c008befc>] (sys_read+0x3c/0x68)
[425375.810000] [<c008befc>] (sys_read+0x3c/0x68) from [<c0025ea0>] (ret_fast_syscall+0x0/0x28)
[425375.810000] Code: e5813018 e5902000 e59f31a4 e5823120 (e1d501b6)
[425376.240000] ---[ end trace 5f66d043b00bc2c9 ]---
[425376.250000] Kernel panic - not syncing: Fatal exception in interrupt


Have anyone experienced similar problem or can help me in the right direction to fix this issue?

Please let me know if I should attach more information.

Best regards,
Niclas

Check what it's doing in atmel_spi_interrupt() about 18-20% into the code.

I suspect you'll find it's accessing a structure who's pointer is a NULL. Add a sanity check for the pointer, and try to figure out why it would be zero.

Update:
The pointer which is NULL and causing the kernel to crash is xfer in atmel_spi_interrupt(). This is the same pointer as as->current_transfer.

The crash occur when derefering at line 459:

/* REVISIT: udelay in irq is unfriendly */
if (xfer->delay_usecs)
udelay(xfer->delay_usecs);

To get to this part of the code a SPI overrun must have taken place.

Havn't found out why as->current_transfer is NULL.

Regarding all the zeros I get when reading the in0_input and in1_input, e.g. cat /sys/bus/spi/devices/spi0.1/in0_input
In the driver for the touchscreen ADC (ads7846.c), function ads7846_read12_ser(...) is called when doing cat on the sysfs file. It seems like status = spi_sync(spi, &req->msg); returns before the PDC has moved data from the SPI to the supplied &req->sample. Hence the sample is zero. If I add a printk() or a small delay after the call to spi_sync() the value is read correct (the PDC get som extra time to get the data to the req->sample buffer). Any hints how to fix this?