2009-04-07 16:12:55 Supporting two AD1836 codecs (on separate SPORT interfaces)
Steve Strobel (UNITED STATES)
Message: 72359
I am trying to figure out how to support two AD1836 codecs, treating them as separate sound "cards". Each chip is interfaced to one of the BF537 SPORT channels. Each one works when used independently (by setting up the driver to use its SPORT and SPI chip select). I think that getting them to work at the same time should be as easy as setting up two peripherals that use the same driver, but the driver doesn't seem to be written to allow that, as the source says:
/* Only one AD1836 soundcard is supported */
static struct platform_device *device = NULL;
static struct ad1836_spi *ad1836_spi = NULL;
I would be interested in suggestions about how to "fix" the driver so it can support multiple instances, or maybe in an example of a sound driver that does.
Thanks,
Steve
P.S. My actual situation is considerably more complicated. I modified the AD1836 driver to make it work with the AD1938. Attached to the same SPI chip select is an AD5204 soft-pot which is used for input level control; I have to transfer data to both devices at the same time. The basic problem, however, would be the same with the AD1836 driver that is in the distribution.
QuoteReplyEditDelete
2009-04-07 16:52:29 Re: Supporting two AD1836 codecs (on separate SPORT interfaces)
Mike Frysinger (UNITED STATES)
Message: 72362
yes, the current driver only supports one instance ... in order to fix, all global state needs to be moved into private data and allocated on the fly in the snd_ad1836_probe() function.
QuoteReplyEditDelete
2009-04-07 16:54:02 Re: Supporting two AD1836 codecs (on separate SPORT interfaces)
Steve Strobel (UNITED STATES)
Message: 72363
I tried a different approach to solving the problem of the AD1836 driver supporting only one "soundcard". I copied the sound driver source file to a new name, gave the driver a new name (appending "-second") and tried to set up one driver for each codec chip. I made both sound drivers use the same SPI driver, but with different chip selects. I am getting error messages and a kernel panic on startup (log shown below). Is this method even worth pursuing? Any suggestions about what else I need to do to separate the drivers enough that they don't cause trouble on startup? Thanks for any ideas.
Steve
From the arch/blackfin/mach-bf537/boards/target.c file:
static struct bfin5xx_spi_chip ad1938_spi_chip_info = {
.bits_per_word = 24, // the AD1938 uses 24 rather than 16 bits like the AD1836
// the bit-banged SPI driver doesn't use .ctl_reg or .enable_dma
};
static struct bfin5xx_spi_chip ad5204_spi_chip_info = {
.bits_per_word = 11,
};
static struct bfin5xx_spi_chip ad1938_second_spi_chip_info = {
.bits_per_word = 24, // the AD1938 uses 24 rather than 16 bits like the AD1836
};
From the spi_board_info array:
{
.modalias = "ad1938-spi",
.max_speed_hz = 3125000,
.bus_num = 2, // custom bit-banged SPI driver
.chip_select = GPIO_PG3,
.controller_data = &ad1938_spi_chip_info,
.mode = SPI_MODE_0,
},
{
.modalias = "ad5204-spi",
.max_speed_hz = 3125000,
.bus_num = 2,
.chip_select = GPIO_PG8,
.controller_data = &ad5204_spi_chip_info,
.mode = SPI_MODE_0,
},
{
.modalias = "ad1938-spi",
.max_speed_hz = 3125000,
.bus_num = 2,
.chip_select = GPIO_PG4,
.controller_data = &ad1938_second_spi_chip_info,
.mode = SPI_MODE_0,
},
I am getting the following error messages on startup, which I don't yet know how to interpret.
bfin-spi bfin-spi.0: Blackfin BF5xx on-chip SPI Contoller Driver, Version 1.0, regs_base@ffc00500, dma channel@7
bfin-spi-gpio-mstr: registering platform driver
bfin-spi-gpio-mstr: bfin_spi_gpio_probe()
bits_per_word not specified in spi_bitbang_setup. Defaulting to 32.
bits_per_word not specified in spi_bitbang_setup. Defaulting to 32.
bits_per_word not specified in spi_bitbang_setup. Defaulting to 32.
Advanced Linux Sound Architecture Driver Version 1.0.12rc1 (Thu Jun 22 13:55:50 2006 UTC).
dma_alloc_init: dma_page @ 0x00392000 - 256 pages at 0x03f00000
snd_ad1938_configure
kobject_add failed for AD1938.0 with -EEXIST, don't try to register things with the same name in the same directory.
Hardware Trace:
0 Target : <0x000047f8> { _dump_stack + 0x0 }
Source : <0x00098a2c> { _kobject_shadow_add + 0xd4 }
1 Target : <0x00098a2c> { _kobject_shadow_add + 0xd4 }
Source : <0x00098a70> { _kobject_shadow_add + 0x118 }
2 Target : <0x00098a70> { _kobject_shadow_add + 0x118 }
Source : <0x0000d9b2> { _printk + 0x16 }
3 Target : <0x0000d9ae> { _printk + 0x12 }
Source : <0x0000d86c> { _vprintk + 0x1c0 }
4 Target : <0x0000d864> { _vprintk + 0x1b8 }
Source : <0xffa00d2c> { __common_int_entry + 0xcc }
5 Target : <0xffa00cca> { __common_int_entry + 0x6a }
Source : <0xffa00b12> { _return_from_int + 0x4e }
6 Target : <0xffa00b12> { _return_from_int + 0x4e }
Source : <0xffa00af2> { _return_from_int + 0x2e }
7 Target : <0xffa00ac4> { _return_from_int + 0x0 }
Source : <0xffa00cc6> { __common_int_entry + 0x66 }
8 Target : <0xffa00cc4> { __common_int_entry + 0x64 }
Source : <0xffa00f76> { _do_irq + 0x8e }
9 Target : <0xffa00f72> { _do_irq + 0x8a }
Source : <0xffa00424> { _asm_do_IRQ + 0x68 }
10 Target : <0xffa0041c> { _asm_do_IRQ + 0x60 }
Source : <0x00011f5e> { _irq_exit + 0x3e }
11 Target : <0x00011f5a> { _irq_exit + 0x3a }
Source : <0x00011c6e> { ___do_softirq + 0x96 }
12 Target : <0x00011c68> { ___do_softirq + 0x90 }
Source : <0x00011b62> { __local_bh_enable + 0x3e }
13 Target : <0x00011b24> { __local_bh_enable + 0x0 }
Source : <0x00011c64> { ___do_softirq + 0x8c }
14 Target : <0x00011c64> { ___do_softirq + 0x8c }
Source : <0x00011c44> { ___do_softirq + 0x6c }
15 Target : <0x00011c38> { ___do_softirq + 0x60 }
Source : <0x00011d4c> { _tasklet_action + 0x7c }
Stack from 00275d54:
00275d58 00275d78 00098a30 002b6e00 ffffffef 00000000 00397a74 0000001f
00397a74 00275da0 00098abe 002b6e00 00397a08 001c6554 0019de10 00000000
00397a08 0019de10 00000000 00275dc8 000b2450 00000000 00275de4 00397aac
00397aac 00397a00 001c6554 00397a00 00000000 00275e00 000b6040 002b6e00
00397a00 001c6554 00000000 00000000 001a0290 00000000 00275e04 000b619e
00397a00 00397b5c 00000000 00275e24 000b6222 002b6e00 001a0290 00397a00
Call Trace:
[<00069a84>] _sysfs_create_link+0xc4/0x128
[<000e66b0>] _snd_ad1938_spi_probed+0x28/0x74
[<0000a1ff>] _show_state_filter+0x157/0x15c
[<000e49f8>] _ad1938_spi_probe+0x34/0x48
[<000cb3b8>] _spi_drv_probe+0x18/0x2c
[<000b4520>] _driver_probe_device+0x64/0x17c
[<000b4536>] _driver_probe_device+0x7a/0x17c
[<00139a56>] _klist_next+0x36/0x60
[<000b4812>] ___driver_attach+0xfa/0x104
[<000b3a0a>] _bus_for_each_dev+0x36/0x64
[<000b4718>] ___driver_attach+0x0/0x104
[<000b43ea>] _driver_attach+0x1a/0x20
[<000b4718>] ___driver_attach+0x0/0x104
[<000b3d44>] _bus_add_driver+0x64/0x164
[<000b4a1a>] _driver_register+0x42/0x74
[<000cb22a>] _spi_register_driver+0x4e/0x54
[<000081a4>] _l1_data_sram_free+0x8/0x28
[<000e4a34>] _ad1938_spi_init+0x10/0x18
[<001b1f52>] _snd_ad1938_init+0x26/0x50
[<001a66b6>] _kernel_init+0x8a/0x238
[<001a662c>] _kernel_init+0x0/0x238
[<00001676>] _kernel_thread_helper+0x6/0xc
NULL pointer access (probably)
Kernel OOPS in progress
Defered Exception context
CURRENT PROCESS:
COMM=swapper PID=1
invalid mm
return address: [0x00139a8a]; contents of:
0x00139a60: 3045 0064 e801 0000 3006 05b3 0010 ac5a
0x00139a70: 3042 0807 17eb 6006 2fef ac5a 9117 2fe3
0x00139a80: 04c4 3228 e800 0003 916a [acd4] e3ff ffbe
0x00139a90: 0c00 1c05 0c44 1803 3045 0064 e801 0000
SEQUENCER STATUS: Not tainted
SEQSTAT: 00000027 IPEND: 8030 SYSCFG: 0006
HWERRCAUSE: 0x0
EXCAUSE : 0x27
physical IVG15 asserted : <0xffa00e80> { _evt_system_call + 0x0 }
logical irq 6 mapped : <0xffa00288> { _timer_interrupt + 0x0 }
logical irq 12 mapped : <0x000e4e6c> { _rx_handler + 0x0 }
logical irq 13 mapped : <0x000e4e0c> { _tx_handler + 0x0 }
logical irq 24 mapped : <0x000bb124> { _bf537mac_interrupt + 0x0 }
logical irq 45 mapped : <0x000e522c> { _err_handler + 0x0 }
RETE: <0x00000000> /* Maybe null pointer? */
RETN: <0x00275cf4> /* unknown address */
RETX: <0x00139a8a> { _klist_del + 0xa }
RETS: <0x00139b12> { _klist_remove + 0xe }
PC : <0x00139a8a> { _klist_del + 0xa }
DCPLB_FAULT_ADDR: <0x0000000c> /* Maybe null pointer? */
ICPLB_FAULT_ADDR: <0x00139a8a> { _klist_del + 0xa }
PROCESSOR STATE:
R0 : 001a02e4 R1 : 001a0290 R2 : 0000001f R3 : 00000000
R4 : 00000000 R5 : 001a035c R6 : 000000d8 R7 : 001a02f4
P0 : 001753cd P1 : 002b766c P2 : 00000000 P3 : 0019f630
P4 : 001a0290 P5 : 001a02e4 FP : 00275d00 SP : 00275c18
LB0: ffa01d7c LT0: ffa01d7a LC0: 00000000
LB1: 0009bfe0 LT1: 0009bfd6 LC1: 00000000
B0 : 00000000 L0 : 00000000 M0 : 00000000 I0 : 00397a7c
B1 : 00000000 L1 : 00000000 M1 : 00000000 I1 : 00000000
B2 : 00000000 L2 : 00000000 M2 : 00000000 I2 : fffedbaf
B3 : 00000000 L3 : 00000000 M3 : 00000000 I3 : 00000000
A0.w: 0000002d A0.x: 00000000 A1.w: 0000002d A1.x: 00000000
USP : 00000000 ASTAT: 00002020
Hardware Trace:
0 Target : <0x0000491c> { _trap_c + 0x0 }
Source : <0xffa007ac> { _exception_to_level5 + 0xb4 }
1 Target : <0xffa006f8> { _exception_to_level5 + 0x0 }
Source : <0xffa00654> { _ex_trap_c + 0x68 }
2 Target : <0xffa00824> { _trap + 0x0 }
Source : <0x00139a88> { _klist_del + 0x8 }
3 Target : <0x00139a80> { _klist_del + 0x0 }
Source : <0x00139b0e> { _klist_remove + 0xa }
4 Target : <0x00139b04> { _klist_remove + 0x0 }
Source : <0x000b3ea2> { _bus_remove_driver + 0x5e }
5 Target : <0x000b3e9a> { _bus_remove_driver + 0x56 }
Source : <0x000b3e76> { _bus_remove_driver + 0x32 }
6 Target : <0x000b3e6e> { _bus_remove_driver + 0x2a }
Source : <0x000b4a9c> { _driver_remove_file + 0x28 }
7 Target : <0x000b4a96> { _driver_remove_file + 0x22 }
Source : <0x000b4a5a> { _put_driver + 0xe }
8 Target : <0x000b4a56> { _put_driver + 0xa }
Source : <0x00098652> { _kobject_put + 0x1a }
9 Target : <0x0009864e> { _kobject_put + 0x16 }
Source : <0x00099256> { _kref_put + 0x42 }
10 Target : <0x00099214> { _kref_put + 0x0 }
Source : <0x0009864a> { _kobject_put + 0x12 }
11 Target : <0x00098638> { _kobject_put + 0x0 }
Source : <0x000b4a52> { _put_driver + 0x6 }
12 Target : <0x000b4a4c> { _put_driver + 0x0 }
Source : <0x000b4a92> { _driver_remove_file + 0x1e }
13 Target : <0x000b4a90> { _driver_remove_file + 0x1c }
Source : <0x00068040> { _sysfs_remove_file + 0x14 }
14 Target : <0x0006803c> { _sysfs_remove_file + 0x10 }
Source : <0x00067982> { _sysfs_hash_and_remove + 0x16 }
15 Target : <0x0006797a> { _sysfs_hash_and_remove + 0xe }
Source : <0x000679f2> { _sysfs_hash_and_remove + 0x86 }
Stack from 00275bf4:
003d08ba 00275d00 ffa007b0 00196560 00196560 0019655c 00000026 00000000
00000001 00139a8a 00008030 00000027 00000000 00275cf4 00139a8a 00139a8a
00139b12 001a02e4 00002020 0009bfe0 ffa01d7c 0009bfd6 ffa01d7a 00000000
00000000 0000002d 00000000 0000002d 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 fffedbaf 00000000 00397a7c 00000000 00275d00 001a02e4
Call Trace:
[<000b4a56>] _put_driver+0xa/0x10
[<00139b12>] _klist_remove+0xe/0x1c
[<000b3ea6>] _bus_remove_driver+0x62/0x84
[<000e5fb8>] _snd_ad1938_set_register+0x24/0x30
[<000b49d0>] _driver_unregister+0x8/0x10
[<00139a1a>] _klist_dec_and_del+0x12/0x18
[<000e4a1c>] _ad1938_spi_done+0x10/0x18
[<000e67a4>] _snd_ad1938_remove+0x18/0x40
[<000e67aa>] _snd_ad1938_remove+0x1e/0x40
[<00018130>] _blocking_notifier_call_chain+0x10/0x18
[<000b5c80>] _platform_drv_remove+0x14/0x1c
[<00067a54>] _sysfs_hash_and_remove+0xe8/0x104
[<000b43aa>] ___device_release_driver+0x66/0x8c
[<000b4984>] _driver_detach+0xf0/0x114
[<00139b18>] _klist_remove+0x14/0x1c
[<000b3eac>] _bus_remove_driver+0x68/0x84
[<000b1afc>] _put_device+0x14/0x1c
[<000b619e>] _platform_device_alloc+0x3e/0x68
[<000b49d0>] _driver_unregister+0x8/0x10
[<000b622e>] _platform_device_register_simple+0x2a/0x58
[<000b5e26>] _platform_driver_unregister+0xa/0x10
[<00069a84>] _sysfs_create_link+0xc4/0x128
[<000e66b0>] _snd_ad1938_spi_probed+0x28/0x74
[<000e66f8>] _snd_ad1938_spi_probed+0x70/0x74
[<0000a1ff>] _show_state_filter+0x157/0x15c
[<000e49f8>] _ad1938_spi_probe+0x34/0x48
[<000cb3b8>] _spi_drv_probe+0x18/0x2c
[<000b4520>] _driver_probe_device+0x64/0x17c
[<000b4536>] _driver_probe_device+0x7a/0x17c
[<00139a56>] _klist_next+0x36/0x60
[<000b4812>] ___driver_attach+0xfa/0x104
[<000b3a0a>] _bus_for_each_dev+0x36/0x64
[<000b4718>] ___driver_attach+0x0/0x104
[<000b43ea>] _driver_attach+0x1a/0x20
[<000b4718>] ___driver_attach+0x0/0x104
[<000b3d44>] _bus_add_driver+0x64/0x164
[<000b4a1a>] _driver_register+0x42/0x74
[<000cb22a>] _spi_register_driver+0x4e/0x54
[<000081a4>] _l1_data_sram_free+0x8/0x28
[<000e4a34>] _ad1938_spi_init+0x10/0x18
[<001b1f52>] _snd_ad1938_init+0x26/0x50
[<001a66b6>] _kernel_init+0x8a/0x238
[<001a662c>] _kernel_init+0x0/0x238
[<00001676>] _kernel_thread_helper+0x6/0xc
Modules linked in:
Kernel panic - not syncing: Kernel exception
QuoteReplyEditDelete
2009-04-07 17:15:47 Re: Supporting two AD1836 codecs (on separate SPORT interfaces)
Mike Frysinger (UNITED STATES)
Message: 72367
you'll have to also change the DRIVER_NAME in the file ... that method may be the shortest route to success ...
QuoteReplyEditDelete
2009-04-07 17:25:55 Re: Supporting two AD1836 codecs (on separate SPORT interfaces)
Steve Strobel (UNITED STATES)
Message: 72370
Mike Frysinger wrote:
> yes, the current driver only supports one instance ...
> in order to fix, all global state needs to be moved into
> private data and allocated on the fly in the
> snd_ad1836_probe() function.
I think I understand how to go about doing that. I am still fuzzy, however, on how to go about creating the second instance of the sound driver. I see how the various stamp_devices are set up in arch/blackfin/mach-bf537/boards/target.c, but I don't see the sound interfaces there (although the SPI and SPORT which are used to communicate with the codec chips are). I also see how sound/blackfin/Kconfig and sound/blackfin/Makefile control whether the driver is compiled. I just don't know what actually causes the sound device to be created. What am I missing?
Steve
QuoteReplyEditDelete
2009-04-07 19:34:40 Re: Supporting two AD1836 codecs (on separate SPORT interfaces)
Mike Frysinger (UNITED STATES)
Message: 72373
the ad1836.c driver is a good example of how to not architect things ... it's why it's deprecated in favor of the alsa soc framework (not that there's a driver replacement for the ad1836 codec yet)
the platform resources declare the spi resources for the ad1836
snd_ad1836_init() is always called via module_init()
- registers the driver "AD1836"
- calls ad1836_spi_init()
ad1836_spi_init() registers the spi driver "ad1836-spi"
the common spi framework detects the match between the platform resources and the ad1836-spi driver
ad1836_spi_probe() is called which turns around and calls snd_ad1836_spi_probed()
snd_ad1836_spi_probed() dynamically registers platform resources for "AD1836"
the common platform framework detects the match between the platform resources and the AD1836 driver
snd_ad1836_probe() is called
i think you can take it from there ...
QuoteReplyEditDelete
2009-04-08 13:02:05 Re: Supporting two AD1836 codecs (on separate SPORT interfaces)
Steve Strobel (UNITED STATES)
Message: 72425
Thanks, Mike, for all of the info about how drivers get registered. I just read (re)read driver-model/platform.txt; I think I mostly understand how the stamp_devices array is used to create platform devices and how they are bound to drivers. I don't see anything in there about creating the sound "card", though, that I could duplicate to create the second one. I am guessing that sound devices are not created as platform devices, but I haven't been able to find where they are created. ALSA-Configuration says that the device nodes can be created by calling /dev/MAKEDEV or using the snddevices script, but I haven't found either of them.
It appears to create the device when it initializes ALSA, as the following gets printed at startup:
Advanced Linux Sound Architecture Driver Version 1.0.12rc1 (Thu Jun 22 13:55:50 2006 UTC).
dma_alloc_init: dma_page @ 0x003a0000 - 256 pages at 0x03f00000
snd_ad1836_configure
ALSA device list:
#0: ADI ad1836 at PF4 SPORT1 rx/tx dma 5/6 err irq 46
and it shows up in /proc/asound:
root:~> cd /proc/asound
root:/proc/asound> ls -l
lrwxrwxrwx 1 root root 5 Dec 31 16:00 ad1836 -> card0
dr-xr-xr-x 4 root root 0 Dec 31 16:00 card0
-r--r--r-- 1 root root 0 Dec 31 16:00 cards
-r--r--r-- 1 root root 0 Dec 31 16:00 devices
dr-xr-xr-x 2 root root 0 Dec 31 16:00 oss
-r--r--r-- 1 root root 0 Dec 31 16:00 pcm
-r--r--r-- 1 root root 0 Dec 31 16:00 timers
-r--r--r-- 1 root root 0 Dec 31 16:00 version
root:/proc/asound> cat cards
0 [ad1836 ]: AD1836 - ADI ad1836
ADI ad1836 at PF4 SPORT1 rx/tx dma 5/6 err irq 46
Thanks for any pointers to where the sound "card" is created.
Steve
QuoteReplyEditDelete
2009-04-08 13:11:34 Re: Supporting two AD1836 codecs (on separate SPORT interfaces)
Steve Strobel (UNITED STATES)
Message: 72427
I think I can answer my own question. It appears that the AD1836 driver not only registers itself as a driver, but creates the device too. There is no platform_device or platform_data structure with the data for the device, so the driver just uses the configuration options (like SND_BLACKFIN_SPORT) directly. Cleaning all of that up so the driver can support multiple devices would be complicated. I think I will pursue the copy-the-driver method, as it will automatically create another device. Did I miss anything?
Steve
QuoteReplyEditDelete
2009-04-08 16:55:47 Re: Supporting two AD1836 codecs (on separate SPORT interfaces)
Mike Frysinger (UNITED STATES)
Message: 72437
right, snd_ad1836_spi_probed() registers platform resources dynamically
you'll have to copy ad1836.c and ad1836_spi.c and make sure to change the driver name used in both (look for ".name" fields)
QuoteReplyEditDelete
2009-04-08 22:29:17 Re: Supporting two AD1836 codecs (on separate SPORT interfaces)
Cliff Cai (CHINA)
Message: 72441
Steve,
You can refer to AD73322 driver,it does support two instances.It supports one card on SPORT0 the other on SPORT1 at the same time.See snd_ad73322_init() and snd_ad73322_probe()
Cliff
QuoteReplyEditDelete
2009-04-09 17:39:47 Serial/SPORT DMA conflict? (was Supporting two AD1836 codecs...)
Steve Strobel (UNITED STATES)
Message: 72480
Mike Frysinger wrote:
> ...you'll have to copy ad1836.c and ad1836_spi.c and make sure to
> change the driver name used in both (look for ".name" fields)
It took a little tweaking to avoid link-time and run-time conflicts, but that seems to be working well. Thanks for the advice. On startup I now see:
snd_ad1938_configure
snd_ad1938_configure
ALSA device list:
#0: ADI ad1938 at PF3 SPORT0 rx/tx dma 3/4 err irq 45
#1: ADI ad1938 at PF4 SPORT1 rx/tx dma 5/6 err irq 46
After boot, I can telnet to it and inspect codec registers using cat /proc/asound/card0/registers; that works fine. I changed to a dummy startup script, so it isn't running anything except the drivers on startup.
I now have a new problem that may be related to the DMA resources used by the second codec driver; after booting if I press any key in my RS-232 serial terminal program it causes a stack dump (shown below). I can, however, successfully telnet to it, inspect codec registers using cat /proc/asound/card0/registers, etc. I haven't tried running audio through it yet. It will run for a long time if I don't send it any serial, but as soon as I do, it dumps as shown below.
Linux version 2.6.22.19-ADI-2008R1.5-svn477 (stevestrobel@dev1-ubuntu) (gcc version 4.1.2 (ADI svn)) #30 Thu Apr 9 14:37
:24 MDT 2009
Hardware Trace Active and Enabled
Blackfin support (C) 2004-2007 Analog Devices, Inc.
Compiled for ADSP-BF537 Rev 0.3
Blackfin Linux support by blackfin.uclinux.org/
Processor Speed: 500 MHz core clock and 100 MHz System Clock
Board Memory: 64MB
Kernel Managed Memory: 64MB
Memory map:
text = 0x00001000-0x0013b980
rodata = 0x0013c000-0x00190ddc
data = 0x00191000-0x001a6000
stack = 0x00192000-0x00194000
init = 0x001a6000-0x001b9000
bss = 0x001b9000-0x001c9f10
available = 0x001c9f10-0x02700000
rootfs = 0x02700000-0x03f00000
DMA Zone = 0x03f00000-0x04000000
NOMPU: setting up cplb tables for global access
Instruction Cache Enabled
Data Cache Enabled (write-back)
Built 1 zonelists. Total pages: 9906
Kernel command line: root=/dev/mtdblock0 rw console=ttyBF0,115200
Configuring Blackfin Priority Driven Interrupts
PID hash table entries: 256 (order: 8, 1024 bytes)
Dentry cache hash table entries: 8192 (order: 3, 32768 bytes)
Inode-cache hash table entries: 4096 (order: 2, 16384 bytes)
Memory available: 37528k/65536k RAM, (76k init code, 1258k kernel code, 493k data, 1024k dma, 584k reserved)
Blackfin Scratchpad data SRAM: 4 KB
Blackfin Data A SRAM: 16 KB (15 KB free)
Blackfin Data B SRAM: 16 KB (16 KB free)
Blackfin Instruction SRAM: 48 KB (39 KB free)
Security Framework v1.0.0 initialized
Capability LSM initialized
Mount-cache hash table entries: 512
NET: Registered protocol family 16
Blackfin GPIO Controller
Blackfin DMA Controller
tcb_406_board_init(): registering device resources
tcb_406_board_init(): registering Blackfin SPI resources
tcb_406_board_init(): finished registering Blackfin SPI resources
Generic PHY: Registered new driver
SCSI subsystem initialized
NET: Registered protocol family 2
IP route cache hash table entries: 1024 (order: 0, 4096 bytes)
TCP established hash table entries: 2048 (order: 2, 16384 bytes)
TCP bind hash table entries: 2048 (order: 1, 8192 bytes)
TCP: Hash tables configured (established 2048 bind 2048)
TCP reno registered
JFFS2 version 2.2. (NAND) Â
2001-2006 Red Hat, Inc.
io scheduler noop registered
io scheduler anticipatory registered (default)
io scheduler cfq registered
ISA-BlackFin-CAN CAN Driver 3.4.6_AD_BLACKFIN (c) Apr 8 2009
BlackFin port by H.J. Oertel (oe@port.de)
bfin-wdt: initialized: timeout=20 sec (nowayout=0)
RLC-DSP4 cors driver
MSAT-IP init_switch driver
RLC-DSP4 ptts driver
Serial: Blackfin serial driver
bfin-uart.1: ttyBF0 at MMIO 0xffc00400 (irq = 18) is a BFIN-UART
bfin-uart.1: ttyBF1 at MMIO 0xffc02000 (irq = 20) is a BFIN-UART
RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
bfin_mac_mdio: probed
bfin_mac: attached PHY driver [Generic PHY] (mii_bus:phy_addr=0:01, irq=-1, mdc_clk=2500000Hz(mdc_div=19)@sclk=100MHz)
bfin_mac: Version 1.1, Blackfin BF53[67] BF527 on-chip Ethernet MAC driver
uclinux[mtd]: RAM probe address=0x2700000 size=0x1800000
Creating 1 MTD partitions on "RAM":
0x00000000-0x01800000 : "ROMfs"
m25p80 spi0.1: m25p128 (16384 Kbytes)
Creating 4 MTD partitions on "m25p80":
0x00000000-0x00040000 : "bootloader"
0x00040000-0x00080000 : "bootloader-environment"
0x00080000-0x00e00000 : "kernel"
0x00e00000-0x01000000 : "file system"
bfin-spi bfin-spi.0: Blackfin BF5xx on-chip SPI Contoller Driver, Version 1.0, regs_base@ffc00500, dma channel@7
bfin-spi-gpio-mstr: registering platform driver
bfin-spi-gpio-mstr: bfin_spi_gpio_probe()
bits_per_word not specified in spi_bitbang_setup. Defaulting to 32.
bits_per_word not specified in spi_bitbang_setup. Defaulting to 32.
bits_per_word not specified in spi_bitbang_setup. Defaulting to 32.
bits_per_word not specified in spi_bitbang_setup. Defaulting to 32.
Advanced Linux Sound Architecture Driver Version 1.0.12rc1 (Thu Jun 22 13:55:50 2006 UTC).
dma_alloc_init: dma_page @ 0x02690000 - 256 pages at 0x03f00000
snd_ad1938_configure
snd_ad1938_configure
ALSA device list:
#0: ADI ad1938 at PF3 SPORT0 rx/tx dma 3/4 err irq 45
#1: ADI ad1938 at PF4 SPORT1 rx/tx dma 5/6 err irq 46
TCP cubic registered
NET: Registered protocol family 1
NET: Registered protocol family 17
VFS: Mounted root (ext2 filesystem).
Freeing unused kernel memory: 76k freed
_____________________________________
a8888b. / Welcome to the uClinux distribution \
d888888b. / _ _ \
8P"YP"Y88 / | | |_| __ __ (TM) |
8|o||o|88 _____/ | | _ ____ _ _ \ \/ / |
8' .88 \ | | | | _ \| | | | \ / |
8`._.' Y8. \ | |__ | | | | | |_| | / \ |
d/ `8b. \ \____||_|_| |_|\____|/_/\_\ |
dP . Y8b. \ For embedded processors including |
d8:' " `::88b \ the Analog Devices Blackfin /
d8" 'Y88b \___________________________________/
:8P ' :888
8a. : _a88P For further information, check out:
._/"Yaa_: .| 88P| - blackfin.uclinux.org/
\ YP" `| 8P `. - docs.blackfin.uclinux.org/
/ \.___.d| .' - www.uclinux.org/
`--..__)8888P`._.' jgs/a:f - www.analog.com/blackfin
Have a lot of fun...
This is /mnt/permanent/scripts/init-passwd
Checking for password file
Password file already exists
Password file exists - using it
BusyBox v1.4.1 (2009-04-08 18:09:09 MDT) Built-in shell (msh)
Enter 'help' for a list of built-in commands.
root:~>
To prevent start-all from being executed automatically, reboot and as
soon as you see the uClinux banner above (before it gets here), type
"stop" and press ENTER.
Dummy startup script...
PHY: 0:01 - Link is Up - 100/Full
[It will run here until I send it a serial character, after which...]
NULL pointer access (probably)
Kernel OOPS in progress
Defered Exception context
No Valid process in current context
return address: [0x000b13d0]; contents of:
0x000b13b0: 081a 303a 185a 3067 3028 67f4 680b 204e
0x000b13c0: 6801 600f e522 0025 a1e1 3001 4c08 5a91
0x000b13d0: [9952] ae22 5430 0c00 9155 141b e6ba ffff
0x000b13e0: e6bd fffe e528 0044 0c40 1874 0000 acc1
SEQUENCER STATUS: Not tainted
SEQSTAT: 00000027 IPEND: c030 SYSCFG: 0006
HWERRCAUSE: 0x0
EXCAUSE : 0x27
physical IVG14 asserted : <0xffa00b20> { _evt14_softirq + 0x0 }
physical IVG15 asserted : <0xffa00e80> { _evt_system_call + 0x0 }
logical irq 6 mapped : <0xffa00288> { _timer_interrupt + 0x0 }
logical irq 12 mapped : <0x000e4e90> { _rx_handler + 0x0 }
logical irq 13 mapped : <0x000e4e30> { _tx_handler + 0x0 }
logical irq 14 mapped : <0x000e4e90> { _rx_handler + 0x0 }
logical irq 15 mapped : <0x000e4e30> { _tx_handler + 0x0 }
logical irq 18 mapped : <0x000b128c> { _bfin_serial_dma_rx_int + 0x0 }
logical irq 19 mapped : <0x000b11f0> { _bfin_serial_dma_tx_int + 0x0 }
logical irq 24 mapped : <0x000bb124> { _bf537mac_interrupt + 0x0 }
logical irq 45 mapped : <0x000e5250> { _err_handler + 0x0 }
logical irq 46 mapped : <0x000e5250> { _err_handler + 0x0 }
RETE: <0x00000000> /* Maybe null pointer? */
RETN: <0x00193dbc> /* unknown address */
RETX: <0x000b13d0> { _bfin_serial_rx_dma_timeout + 0xe0 }
RETS: <0x000b130e> { _bfin_serial_rx_dma_timeout + 0x1e }
PC : <0x000b13d0> { _bfin_serial_rx_dma_timeout + 0xe0 }
DCPLB_FAULT_ADDR: <0x00000000> /* Maybe null pointer? */
ICPLB_FAULT_ADDR: <0x000b13d0> { _bfin_serial_rx_dma_timeout + 0xe0 }
PROCESSOR STATE:
R0 : 00000000 R1 : 00000000 R2 : 00000fff R3 : 00000072
R4 : 00193dd2 R5 : 00000000 R6 : 00000000 R7 : 00000001
P0 : 0000000f P1 : 00000000 P2 : 00000000 P3 : 00000001
P4 : 001c63a4 P5 : 001bff60 FP : 00193dd4 SP : 00193ce0
LB0: ffa01d7c LT0: ffa01d7a LC0: 00000000
LB1: 025f9aed LT1: 025f9aec LC1: 00000000
B0 : 00000000 L0 : 00000000 M0 : 00000000 I0 : 00000001
B1 : 00000000 L1 : 00000000 M1 : 00000000 I1 : 00000000
B2 : 00000000 L2 : 00000000 M2 : 00000000 I2 : 0004513c
B3 : 00000000 L3 : 00000000 M3 : 00000000 I3 : 00000000
A0.w: 000000a0 A0.x: 00000000 A1.w: 000000a0 A1.x: 00000000
USP : 00194000 ASTAT: 02002021
Hardware Trace:
0 Target : <0x0000491c> { _trap_c + 0x0 }
Source : <0xffa007ac> { _exception_to_level5 + 0xb4 }
1 Target : <0xffa006f8> { _exception_to_level5 + 0x0 }
Source : <0xffa00654> { _ex_trap_c + 0x68 }
2 Target : <0xffa00824> { _trap + 0x0 }
Source : <0x000b13ce> { _bfin_serial_rx_dma_timeout + 0xde }
3 Target : <0x000b13c4> { _bfin_serial_rx_dma_timeout + 0xd4 }
Source : <0x000b1466> { _bfin_serial_rx_dma_timeout + 0x176 }
4 Target : <0x000b145a> { _bfin_serial_rx_dma_timeout + 0x16a }
Source : <0x000b13be> { _bfin_serial_rx_dma_timeout + 0xce }
5 Target : <0x000b13b0> { _bfin_serial_rx_dma_timeout + 0xc0 }
Source : <0x000b1502> { _bfin_serial_rx_dma_timeout + 0x212 }
6 Target : <0x000b14fc> { _bfin_serial_rx_dma_timeout + 0x20c }
Source : <0x000b149c> { _bfin_serial_rx_dma_timeout + 0x1ac }
7 Target : <0x000b1496> { _bfin_serial_rx_dma_timeout + 0x1a6 }
Source : <0x000b13ac> { _bfin_serial_rx_dma_timeout + 0xbc }
8 Target : <0x000b139a> { _bfin_serial_rx_dma_timeout + 0xaa }
Source : <0x000b1380> { _bfin_serial_rx_dma_timeout + 0x90 }
9 Target : <0x000b130e> { _bfin_serial_rx_dma_timeout + 0x1e }
Source : <0x000072e2> { _get_dma_curr_xcount + 0x3e }
10 Target : <0x000072a4> { _get_dma_curr_xcount + 0x0 }
Source : <0x000b130a> { _bfin_serial_rx_dma_timeout + 0x1a }
11 Target : <0x000b1300> { _bfin_serial_rx_dma_timeout + 0x10 }
Source : <0x00007272> { _get_dma_curr_ycount + 0x3e }
12 Target : <0x00007234> { _get_dma_curr_ycount + 0x0 }
Source : <0x000b12fc> { _bfin_serial_rx_dma_timeout + 0xc }
13 Target : <0x000b12f0> { _bfin_serial_rx_dma_timeout + 0x0 }
Source : <0x00014abc> { _run_timer_softirq + 0xb4 }
14 Target : <0x00014a92> { _run_timer_softirq + 0x8a }
Source : <0x00014a72> { _run_timer_softirq + 0x6a }
15 Target : <0x00014a24> { _run_timer_softirq + 0x1c }
Source : <0x0001fe50> { _hrtimer_run_queues + 0xc8 }
Stack from 00193cbc:
000ed990 00193dd4 ffa007b0 00196560 00196560 0019655c 00193d50 0012177a
0028a940 000b13d0 0000c030 00000027 00000000 00193dbc 000b13d0 000b13d0
000b130e 00000000 02002021 025f9aed ffa01d7c 025f9aec ffa01d7a 00000000
00000000 000000a0 00000000 000000a0 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 0004513c 00000000 00000001 00194000 00193dd4 001bff60
Call Trace:
[<00014abe>] _run_timer_softirq+0xb6/0x170
[<000b12f0>] _bfin_serial_rx_dma_timeout+0x0/0x224
[<00011c38>] ___do_softirq+0x60/0xac
[<00029d46>] _handle_simple_irq+0x72/0x94
[<00011f5a>] _irq_exit+0x3a/0x40
[<0001d768>] _kthreadd+0x0/0xe0
[<00008050>] __l1_sram_free+0x6c/0xc0
[<0004512b>] _flock_lock_file_wait+0x163/0x1f4
[<0000ffff>] _do_exit+0x2ab/0x774
[<0013a192>] _rest_init+0x42/0x48
[<00040000>] ___link_path_walk+0x8ec/0xc40
[<001a6a70>] _start_kernel+0x20c/0x22c
[<001a63ac>] _unknown_bootoption+0x0/0x220
[<001a61bc>] _real_start+0x50/0x94
Modules linked in:
Kernel panic - not syncing: Kernel exception
--------------------------
If I disable the second codec driver, it boots and accepts input from the serial port normally. A log of that boot process follows for comparison:
Linux version 2.6.22.19-ADI-2008R1.5-svn477 (stevestrobel@dev1-ubuntu) (gcc version 4.1.2 (ADI svn)) #31 Thu Apr 9 15:14
:06 MDT 2009
Hardware Trace Active and Enabled
Blackfin support (C) 2004-2007 Analog Devices, Inc.
Compiled for ADSP-BF537 Rev 0.3
Blackfin Linux support by blackfin.uclinux.org/
Processor Speed: 500 MHz core clock and 100 MHz System Clock
Board Memory: 64MB
Kernel Managed Memory: 64MB
Memory map:
text = 0x00001000-0x0013a470
rodata = 0x0013b000-0x0018f738
data = 0x00190000-0x001a4000
stack = 0x00190000-0x00192000
init = 0x001a4000-0x001b7000
bss = 0x001b7000-0x001c7ef0
available = 0x001c7ef0-0x02200000
rootfs = 0x02200000-0x03f00000
DMA Zone = 0x03f00000-0x04000000
NOMPU: setting up cplb tables for global access
Instruction Cache Enabled
Data Cache Enabled (write-back)
Built 1 zonelists. Total pages: 8636
Kernel command line: root=/dev/mtdblock0 rw console=ttyBF0,115200
Configuring Blackfin Priority Driven Interrupts
PID hash table entries: 256 (order: 8, 1024 bytes)
Dentry cache hash table entries: 8192 (order: 3, 32768 bytes)
Inode-cache hash table entries: 4096 (order: 2, 16384 bytes)
Memory available: 32416k/65536k RAM, (76k init code, 1253k kernel code, 490k data, 1024k dma, 584k reserved)
Blackfin Scratchpad data SRAM: 4 KB
Blackfin Data A SRAM: 16 KB (15 KB free)
Blackfin Data B SRAM: 16 KB (16 KB free)
Blackfin Instruction SRAM: 48 KB (39 KB free)
Security Framework v1.0.0 initialized
Capability LSM initialized
Mount-cache hash table entries: 512
NET: Registered protocol family 16
Blackfin GPIO Controller
Blackfin DMA Controller
tcb_406_board_init(): registering device resources
tcb_406_board_init(): registering Blackfin SPI resources
tcb_406_board_init(): finished registering Blackfin SPI resources
Generic PHY: Registered new driver
SCSI subsystem initialized
NET: Registered protocol family 2
IP route cache hash table entries: 1024 (order: 0, 4096 bytes)
TCP established hash table entries: 2048 (order: 2, 16384 bytes)
TCP bind hash table entries: 2048 (order: 1, 8192 bytes)
TCP: Hash tables configured (established 2048 bind 2048)
TCP reno registered
JFFS2 version 2.2. (NAND) Â 2001-2006 Red Hat, Inc.
io scheduler noop registered
io scheduler anticipatory registered (default)
io scheduler cfq registered
ISA-BlackFin-CAN CAN Driver 3.4.6_AD_BLACKFIN (c) Apr 8 2009
BlackFin port by H.J. Oertel (oe@port.de)
bfin-wdt: initialized: timeout=20 sec (nowayout=0)
RLC-DSP4 cors driver
MSAT-IP init_switch driver
RLC-DSP4 ptts driver
Serial: Blackfin serial driver
bfin-uart.1: ttyBF0 at MMIO 0xffc00400 (irq = 18) is a BFIN-UART
bfin-uart.1: ttyBF1 at MMIO 0xffc02000 (irq = 20) is a BFIN-UART
RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
bfin_mac_mdio: probed
bfin_mac: attached PHY driver [Generic PHY] (mii_bus:phy_addr=0:01, irq=-1, mdc_clk=2500000Hz(mdc_div=19)@sclk=100MHz)
bfin_mac: Version 1.1, Blackfin BF53[67] BF527 on-chip Ethernet MAC driver
uclinux[mtd]: RAM probe address=0x2200000 size=0x1d00000
Creating 1 MTD partitions on "RAM":
0x00000000-0x01d00000 : "ROMfs"
m25p80 spi0.1: m25p128 (16384 Kbytes)
Creating 4 MTD partitions on "m25p80":
0x00000000-0x00040000 : "bootloader"
0x00040000-0x00080000 : "bootloader-environment"
0x00080000-0x00e00000 : "kernel"
0x00e00000-0x01000000 : "file system"
bfin-spi bfin-spi.0: Blackfin BF5xx on-chip SPI Contoller Driver, Version 1.0, regs_base@ffc00500, dma channel@7
bfin-spi-gpio-mstr: registering platform driver
bfin-spi-gpio-mstr: bfin_spi_gpio_probe()
bits_per_word not specified in spi_bitbang_setup. Defaulting to 32.
bits_per_word not specified in spi_bitbang_setup. Defaulting to 32.
Advanced Linux Sound Architecture Driver Version 1.0.12rc1 (Thu Jun 22 13:55:50 2006 UTC).
dma_alloc_init: dma_page @ 0x003a2000 - 256 pages at 0x03f00000
snd_ad1938_configure
ALSA device list:
#0: ADI ad1938 at PF3 SPORT0 rx/tx dma 3/4 err irq 45
TCP cubic registered
NET: Registered protocol family 1
NET: Registered protocol family 17
VFS: Mounted root (ext2 filesystem).
Freeing unused kernel memory: 76k freed
_____________________________________
a8888b. / Welcome to the uClinux distribution \
d888888b. / _ _ \
8P"YP"Y88 / | | |_| __ __ (TM) |
8|o||o|88 _____/ | | _ ____ _ _ \ \/ / |
8' .88 \ | | | | _ \| | | | \ / |
8`._.' Y8. \ | |__ | | | | | |_| | / \ |
d/ `8b. \ \____||_|_| |_|\____|/_/\_\ |
dP . Y8b. \ For embedded processors including |
d8:' " `::88b \ the Analog Devices Blackfin /
d8" 'Y88b \___________________________________/
:8P ' :888
8a. : _a88P For further information, check out:
._/"Yaa_: .| 88P| - blackfin.uclinux.org/
\ YP" `| 8P `. - docs.blackfin.uclinux.org/
/ \.___.d| .' - www.uclinux.org/
`--..__)8888P`._.' jgs/a:f - www.analog.com/blackfin
Have a lot of fun...
This is /mnt/permanent/scripts/init-passwd
Checking for password file
Password file already exists
Password file exists - using it
BusyBox v1.4.1 (2009-04-08 18:09:09 MDT) Built-in shell (msh)
Enter 'help' for a list of built-in commands.
root:~>
To prevent start-all from being executed automatically, reboot and as
soon as you see the uClinux banner above (before it gets here), type
"stop" and press ENTER.
Dummy startup script...
PHY: 0:01 - Link is Up - 100/Full
root:~>
-------------------------------------
Am I getting serial RX DMA timeouts because of a conflict between the DMA resources used by the second AD1938 driver and the serial port driver? Any other suggestions about what I should check? Thanks.
Steve
QuoteReplyEditDelete
2009-04-09 19:21:49 Re: Serial/SPORT DMA conflict? (was Supporting two AD1836 codecs...)
Mike Frysinger (UNITED STATES)
Message: 72484
hmm, if a dma channel was not available, the driver should have complained and disabled itself rather than crash
can you telnet into the board and cat /proc/dma with and without the second codec enabled ?
if you set the UART to PIO mode in the kernel config, does that work as well ?
QuoteReplyEditDelete
2009-04-10 11:28:11 Re: Serial/SPORT DMA conflict? (was Supporting two AD1836 codecs...)
Steve Strobel (UNITED STATES)
Message: 72510
Mike Frysinger wrote:
> hmm, if a dma channel was not available, the driver should have complained and disabled itself rather than crash
Looking in the databook, it appears that the SPORT interfaces are using the default DMA channels. Assuming the UART is also, there is no reason they would have a conflict. I was just guessing that it was DMA related because of the stack dump.
> can you telnet into the board and cat /proc/dma with and without the second codec enabled ?
I don't seem to have that (virtual) file:
root:~> cat /proc/dma
cat: /proc/dma: No such file or directory
root:~> cd proc
root:/proc> ls
1 25 99 filesystems maps sram
136 3 asound fs meminfo stat
152 37 buddyinfo gpio misc sys
154 38 bus interrupts modules sysvipc
157 39 cmdline iomem mounts timer_list
158 4 cplbinfo ioports mtd tty
159 40 cpuinfo irq net uptime
171 5 devices kallsyms partitions version
172 6 diskstats kmsg scsi vmstat
174 77 driver loadavg self zoneinfo
2 83 execdomains locks slabinfo
root:/proc> find -name dma
root:/proc>
> if you set the UART to PIO mode in the kernel config, does that work as well ?
Yes, it works fine in PIO mode. I didn't realize that selection was avaliable. I could just ignore the apparent DMA problem and go on using PIO mode, but it makes me kind of nervous. Is there something else I should check? Thanks for any suggestions.
Steve
QuoteReplyEditDelete
2009-04-10 11:34:37 Re: Serial/SPORT DMA conflict? (was Supporting two AD1836 codecs...)
Mike Frysinger (UNITED STATES)
Message: 72511
hmm, /proc/dma is only in trunk and it doesnt seem to be complete anyways. i'll have to dig into this.
QuoteReplyEditDelete
2009-04-10 13:44:08 Re: Serial/SPORT DMA conflict? (was Supporting two AD1836 codecs...)
Mike Frysinger (UNITED STATES)
Message: 72513
erp, i lied ... /proc/dma is complete. the board i was looking at had UART set to PIO mode for different reasons
please try the attached patch ... it should create a /proc/dma for you. i didnt actually run-time test it though ...
i looked at both drivers and they seem to be doing the right thing (requesting dma and checking the return)
bfin-2008R1-proc-dma.patch
QuoteReplyEditDelete
2009-04-10 18:41:26 ALSA setup for two AD1336 codecs (was Supporting two AD1836 codecs)
Steve Strobel (UNITED STATES)
Message: 72527
Mike Frysinger wrote:
> you'll have to copy ad1836.c and ad1836_spi.c and make sure to
> change the driver name used in both (look for ".name" fields)
I think all of that is set up (I haven't gotton back to trying DMA for the UARTs yet), but I am having trouble figuring out how to refer to the second sound interface. My code was doing the following:
if( snd_pcm_open( &pcm_input_handle, "default", SND_PCM_STREAM_CAPTURE, 0 ) < 0 )
I plan to call snd_pcm_open() twice, once for each AD1836 (actually AD1938), but I obviously can't refer to both of them as "default". I have tried to figure out other ways to refer to each card, but the only other thing I have found that works is "pcm.hw" (from alsa.conf) and I don't know how to refer to a second one of those either. Two sound "cards" exist in /proc/asound:
root:/proc/asound> ls
ad1938 card0 cards oss timers
ad1938_1 card1 devices pcm version
root:/proc/asound> cat cards
0 [ad1938 ]: AD1938 - ADI ad1938
ADI ad1938 at PF3 SPORT0 rx/tx dma 3/4 err irq 45
1 [ad1938_1 ]: AD1938_SECOND - ADI ad1938
ADI ad1938 at PF4 SPORT1 rx/tx dma 5/6 err irq 46
but I can't figure out how to open the PCM interface of the second one (or the first one by any name except "default" or "pcm.hw"). Maybe I need to create a second copy of something else?
root:/proc> cat /proc/devices
Character devices:
[snip]
116 alsa
apparently corresponds with the following from modprobe.conf:
# ALSA native device support
alias char-major-116 snd
options snd major=116 cards_limit=1
alias snd-card-0 snd-ad1836
Since my second driver is not called "snd-ad1836", I presume that I can't just change cards_limit to 2, but that I should copy that section and assign a different major number. Even then, I don't know how that will help with the call to snd_pcm_open(). I would be happy to RTM if someone would point me to one with the answers Thanks for any suggestions.
Steve
QuoteReplyEditDelete
2009-04-10 18:54:33 Re: ALSA setup for two AD1336 codecs (was Supporting two AD1836 codecs)
Mike Frysinger (UNITED STATES)
Message: 72528
you'll probably have to copy AD1836.conf to AD1836_SECOND.conf in the romfs tree
QuoteReplyEditDelete
2009-04-10 19:24:05 Re: ALSA setup for two AD1336 codecs (was Supporting two AD1836 codecs)
Steve Strobel (UNITED STATES)
Message: 72529
Mike Frysinger wrote:
> you'll probably have to copy AD1836.conf to AD1836_SECOND.conf in the romfs tree
I have done that. Those files look like this:
AD1938.pcm.default {
type plug
slave {
pcm {
type mmap_emul
slave {
pcm {
type hw
card 0
format S32_LE
rate 16000
}
}
}
}
}
and
AD1938_SECOND.pcm.default { [snip]
They show up in /usr/share/alsa/cards. Referring to AD1938.pcm.default when calling snd_pcm_open() doesn't seem to work, though. After reading alsa.opensrc.org/.asoundrc, I decided to try "hw:0,0" and "hw:1,0"; that seems to access the two cards. I am now getting the messages, "Error setting HW params. Input audio config: Cannot allocate memory" but at least it is looking at the "cards". Thanks for your help. More troubleshooting on Monday
Steve
QuoteReplyEditDelete
2009-04-12 23:51:53 Re: ALSA setup for two AD1336 codecs (was Supporting two AD1836 codecs)
Cliff Cai (CHINA)
Message: 72562
see: docs.blackfin.uclinux.org/doku.php?id=linux-kernel:drivers:ad73322
it gives an example of config file for multi cards case.
Cliff
QuoteReplyEditDelete
2009-04-13 13:25:22 Re: ALSA setup for two AD1336 codecs (was Supporting two AD1836 codecs)
Steve Strobel (UNITED STATES)
Message: 72602
Cliff Cai wrote:
> see: docs.blackfin.uclinux.org/doku.php?id=linux-kernel:drivers:ad73322
> it gives an example of config file for multi cards case.
I got that working. Thanks.
Steve
QuoteReplyEditDelete
2009-04-13 13:44:04 snd_pcm_hw_params() cannot allocate memory (was Re: ALSA setup for two AD1336 codecs)
Steve Strobel (UNITED STATES)
Message: 72603
Thanks to many suggestions in this tread, I have two AD1938 codec drivers loading on startup. It prints:
ALSA device list:
#0: ADI ad1938 at PF3 SPORT0 rx/tx dma 3/4 err irq 45
#1: ADI ad1938 at PF4 SPORT1 rx/tx dma 5/6 err irq 46
I access them from my userland code using hw:0,0 and hw:1,0 and ALSA finds them. When I call snd_pcm_hw_params() for either one, it returns -12, which snd_strerror() interprets as "Cannot allocate memory". I have about 7MB free, and this happens right after a reset when the memory shouldn't be fragmented, so I doubt that it is having trouble with allocating from the system heap. Is there some kind of ALSA memory pool or anything that I might need to resize? I have been trying to find where snd_pcm_hw_params() allocates memory, but haven't found it yet.
If I enable only one sound card driver, I can point it at either AD1938 (by changing the selected SPORT and PF pin) and either one will work fine using exactly the same userland code (and hw:0,0). The userland code does not yet try to open both at the same time; it fails as described above when the only change is enabling the second driver.
If I enable both drivers, everything seems fine until I try to open the sound interface. I can run amixer to talk to either card, view the register contents using /proc/asound/card?/registers, etc.
Thanks for any suggestions.
Steve
QuoteReplyEditDelete
2009-04-13 17:48:42 Re: snd_pcm_hw_params() cannot allocate memory (was Re: ALSA setup for two AD1336 codecs)
Mike Frysinger (UNITED STATES)
Message: 72610
does it allocate with kmalloc or a dma func ? i'm pretty sure it's the latter which means you need to increase the dma buffer size in your kernel configuration (look under the Blackfin options).
QuoteReplyEditDelete
2009-04-14 13:40:07 Re: snd_pcm_hw_params() cannot allocate memory (was Re: ALSA setup for two AD1336 codecs)
Steve Strobel (UNITED STATES)
Message: 72646
Mike Frysinger wrote:
> does it allocate with kmalloc or a dma func ? i'm pretty sure it's the latter which means
> you need to increase the dma buffer size in your kernel configuration (look under the Blackfin options).
You are right. Increasing the DMA region fixed that problem. Thanks a bunch.
For anyone that might be looking for that option someday, it can be found in the kernel config > Blackfin Processor Options > Enable DMA Support > Uncached SDRAM region > Enable 2M DMA Region (changed from 1M to 2M).