From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <4C4A292F.5040900@domain.hid> Date: Fri, 23 Jul 2010 16:43:43 -0700 From: Bob Feretich MIME-Version: 1.0 References: <4C43C83F.5030401@domain.hid> <4C43D265.9080609@domain.hid> <4C43FC4F.3010006@domain.hid> <4C440D6F.1060204@domain.hid> <4C44E9C9.3020600@domain.hid> <4C454C0C.6000100@domain.hid> <4C4550C8.3070504@domain.hid> <4C45541C.5090507@domain.hid> <4C45CEE6.5000903@domain.hid> <4C460506.3010103@domain.hid> <4C46141F.50103@domain.hid> <4C468040.3030509@domain.hid> <4C468485.5060408@domain.hid> <4C47597F.1010406@domain.hid> <4C478313.2030606@domain.hid> <4C48C1BB.8080606@domain.hid> <4C48C2CE.1030706@domain.hid> In-Reply-To: <4C48C2CE.1030706@domain.hid> Content-Type: multipart/alternative; boundary="------------040401050406050206060908" Subject: Re: [Xenomai-help] Adeos patch prevents IRQ 384 (MMC Chip Detect) on omap-2.6.33 List-Id: Help regarding installation and common use of Xenomai List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Gilles Chanteperdrix Cc: xenomai@xenomai.org This is a multi-part message in MIME format. --------------040401050406050206060908 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit I have started a new thread with a name that better describes the failure I am seeing. This may or may not be the reason I was seeing root file system mount hang on omap-2.6.31 (previous thread). The SD card/MMC "Chip Detect" interrupt (activated when a SD card is inserted or removed) can't be tested when the root file system is on the SD card. So I rebuilt my 2.6.33 BeagleBoard system to mount the root file system via NFS. Now the failure of Xenomai to handle twl4030 interrupts can be identified and isolated. 2.6.33 without the Adeos patch... The Angstrom Distribution beagleboard ttyS2 Angstrom 2010.4-test-20100518 beagleboard ttyS2 beagleboard login: root root@domain.hid:~# cat /proc/interrupts CPU0 7: 2 INTC TWL4030-PIH 12: 203 INTC DMA 37: 1219 INTC gp timer 56: 368 INTC i2c_omap 61: 0 INTC i2c_omap 74: 125 INTC serial 77: 10694 INTC ehci_hcd:usb2 83: 106 INTC mmc0 92: 1 INTC musb_hdrc 93: 0 INTC musb_hdrc 378: 2 twl4030 twl4030_usb 379: 0 twl4030 rtc0 384: 0 twl4030 mmc0 Err: 0 root@domain.hid:~# ls /dev block ram6 tty39 ... snipped ... mem tty2 tty58 mmcblk0 tty20 tty59 mmcblk0p1 tty21 tty6 mmcblk0p2 tty22 tty60 network_latency tty23 tty61 ... snipped ... root@domain.hid:~# umount /dev/mmcblk0p1 root@domain.hid:~# umount /dev/mmcblk0p2 ****** I removed the SD card here **************************** root@domain.hid:~# [ 122.126495] mmc0: card 0007 removed root@domain.hid:~# ls /dev block ram9 tty40 ... snipped ... mem tty21 tty59 <=== mmc entries are gone network_latency tty22 tty6 ... snipped ... root@domain.hid:~# cat /proc/interrupts CPU0 7: 3 INTC TWL4030-PIH 12: 208 INTC DMA 37: 2477 INTC gp timer 56: 388 INTC i2c_omap 61: 0 INTC i2c_omap 74: 437 INTC serial 77: 12374 INTC ehci_hcd:usb2 83: 135 INTC mmc0 92: 1 INTC musb_hdrc 93: 0 INTC musb_hdrc 378: 2 twl4030 twl4030_usb 379: 0 twl4030 rtc0 384: 1 twl4030 mmc0 <===== notice the Chip Detect irpt Err: 0 ****** I reinserted the SD card here **************************** root@domain.hid:~# [ 270.691955] mmc0: new high speed SD card at address 0007 [ 270.699981] mmcblk0: mmc0:0007 SD01G 972 MiB [ 270.704742] mmcblk0: p1 p2 [ 271.448608] EXT2-fs (mmcblk0p2): warning: mounting ext3 filesystem as ext2 [ 271.455749] EXT2-fs (mmcblk0p2): warning: mounting unchecked fs, running e2fsck is recommended root@domain.hid:~# cat /proc/interrupts CPU0 7: 4 INTC TWL4030-PIH 12: 358 INTC DMA 37: 3520 INTC gp timer 56: 416 INTC i2c_omap 61: 0 INTC i2c_omap 74: 625 INTC serial 77: 14021 INTC ehci_hcd:usb2 83: 247 INTC mmc0 92: 1 INTC musb_hdrc 93: 0 INTC musb_hdrc 378: 2 twl4030 twl4030_usb 379: 0 twl4030 rtc0 384: 2 twl4030 mmc0 <===== notice second Chip Detect irpt Err: 0 2.6.33 with the Adeos patch and Xenomai enabled... The Angstrom Distribution beagleboard ttyS2 Angstrom 2010.4-test-20100518 beagleboard ttyS2 beagleboard login: root root@domain.hid:~# cat /proc/interrupts CPU0 7: 161656 INTC TWL4030-PIH 12: 176 INTC DMA 37: 2642 INTC gp timer 56: 585760 INTC i2c_omap 61: 0 INTC i2c_omap 74: 85 INTC serial 77: 11018 INTC ehci_hcd:usb2 83: 144 INTC mmc0 92: 1 INTC musb_hdrc 93: 0 INTC musb_hdrc 378: 0 twl4030 twl4030_usb 379: 0 twl4030 rtc0 384: 0 twl4030 mmc0 Err: 0 root@domain.hid:~# ls /dev block rtp11 tty29 ... snipped ... mem rtp30 tty48 mmcblk0 rtp31 tty49 mmcblk0p1 rtp4 tty5 mmcblk0p2 rtp5 tty50 network_latency rtp6 tty51 ... snipped ... root@domain.hid:~# umount /dev/mmcblk0p1 root@domain.hid:~# umount /dev/mmcblk0p2 ****** I removed the SD card here, no removal msg ***************** root@domain.hid:~# ls /dev block rtp11 tty29 ... snipped ... mem rtp30 tty48 mmcblk0 rtp31 tty49 mmcblk0p1 rtp4 tty5 mmcblk0p2 rtp5 tty50 network_latency rtp6 tty51 ... snipped ... root@domain.hid:~# cat /proc/interrupts CPU0 7: 1106318 INTC TWL4030-PIH 12: 180 INTC DMA 37: 13933 INTC gp timer 56: 4029606 INTC i2c_omap 61: 0 INTC i2c_omap 74: 440 INTC serial 77: 12416 INTC ehci_hcd:usb2 83: 364 INTC mmc0 92: 1 INTC musb_hdrc 93: 0 INTC musb_hdrc 378: 0 twl4030 twl4030_usb <=== also 0, but not sure if its meaningful 379: 0 twl4030 rtc0 384: 0 twl4030 mmc0 <===== no Chip Detect irpt Err: 0 Reinserting the SD card did not result in a message. The SD card file systems seemed to be in a half mounted state. I suspect that this bug effects all interrupts received via twl4030 family devices. Regards, Bob Feretich On 7/22/2010 3:14 PM, Gilles Chanteperdrix wrote: > ... snipped ... > As I said, the absence of IRQ 378 in 2.6.33 looks like a regression, > which should be easy to find with git bisect. Then when you know the > commit which caused the regression, tell me, I will revert it on 2.6.33, > so that I can reproduce the bug you are facing on my board, and > hopefully what happens with the Adeos patch. > > --------------040401050406050206060908 Content-Type: text/html; charset="utf-8" Content-Transfer-Encoding: 8bit I have started a new thread with a name that better describes the failure I am seeing. This may or may not be the reason I was seeing root file system mount hang on omap-2.6.31 (previous thread).

The SD card/MMC  "Chip Detect" interrupt (activated when a SD card is inserted or removed) can't be tested when the root file system is on the SD card. So I rebuilt my 2.6.33 BeagleBoard system to mount the root file system via NFS. Now the  failure of Xenomai to handle twl4030 interrupts can be identified and isolated.

2.6.33 without the Adeos patch...
The Angstrom Distribution beagleboard ttyS2

Angstrom 2010.4-test-20100518 beagleboard ttyS2

beagleboard login: root
root@domain.hid:~# cat /proc/interrupts
           CPU0
  7:          2        INTC  TWL4030-PIH
 12:        203        INTC  DMA
 37:       1219        INTC  gp timer
 56:        368        INTC  i2c_omap
 61:          0        INTC  i2c_omap
 74:        125        INTC  serial
 77:      10694        INTC  ehci_hcd:usb2
 83:        106        INTC  mmc0
 92:          1        INTC  musb_hdrc
 93:          0        INTC  musb_hdrc
378:          2     twl4030  twl4030_usb
379:          0     twl4030  rtc0
384:          0     twl4030  mmc0
Err:          0
root@domain.hid:~# ls /dev
block               ram6                tty39
... snipped ...
mem                 tty2                tty58
mmcblk0             tty20               tty59
mmcblk0p1           tty21               tty6
mmcblk0p2           tty22               tty60
network_latency     tty23               tty61
... snipped ...
root@domain.hid:~# umount /dev/mmcblk0p1
root@domain.hid:~# umount /dev/mmcblk0p2

****** I removed the SD card here ****************************

root@domain.hid:~# [  122.126495] mmc0: card 0007 removed
root@domain.hid:~# ls /dev
block               ram9                tty40
... snipped ...
mem                 tty21               tty59  <=== mmc entries are gone
network_latency     tty22               tty6
... snipped ...
root@domain.hid:~# cat /proc/interrupts
           CPU0
  7:          3        INTC  TWL4030-PIH
 12:        208        INTC  DMA
 37:       2477        INTC  gp timer
 56:        388        INTC  i2c_omap
 61:          0        INTC  i2c_omap
 74:        437        INTC  serial
 77:      12374        INTC  ehci_hcd:usb2
 83:        135        INTC  mmc0
 92:          1        INTC  musb_hdrc
 93:          0        INTC  musb_hdrc
378:          2     twl4030  twl4030_usb
379:          0     twl4030  rtc0
384:          1     twl4030  mmc0          <===== notice the Chip Detect irpt
Err:          0

****** I reinserted the SD card here ****************************

 root@domain.hid:~# [  270.691955] mmc0: new high speed SD card at address 0007
[  270.699981] mmcblk0: mmc0:0007 SD01G 972 MiB
[  270.704742]  mmcblk0: p1 p2
[  271.448608] EXT2-fs (mmcblk0p2): warning: mounting ext3 filesystem as ext2
[  271.455749] EXT2-fs (mmcblk0p2): warning: mounting unchecked fs, running e2fsck is recommended

root@domain.hid:~# cat /proc/interrupts
           CPU0
  7:          4        INTC  TWL4030-PIH
 12:        358        INTC  DMA
 37:       3520        INTC  gp timer
 56:        416        INTC  i2c_omap
 61:          0        INTC  i2c_omap
 74:        625        INTC  serial
 77:      14021        INTC  ehci_hcd:usb2
 83:        247        INTC  mmc0
 92:          1        INTC  musb_hdrc
 93:          0        INTC  musb_hdrc
378:          2     twl4030  twl4030_usb
379:          0     twl4030  rtc0
384:          2     twl4030  mmc0        <===== notice second Chip Detect irpt
Err:          0

2.6.33 with the Adeos patch and Xenomai enabled...
The Angstrom Distribution beagleboard ttyS2

Angstrom 2010.4-test-20100518 beagleboard ttyS2

beagleboard login: root
root@domain.hid:~# cat /proc/interrupts
           CPU0
  7:     161656        INTC  TWL4030-PIH
 12:        176        INTC  DMA
 37:       2642        INTC  gp timer
 56:     585760        INTC  i2c_omap
 61:          0        INTC  i2c_omap
 74:         85        INTC  serial
 77:      11018        INTC  ehci_hcd:usb2
 83:        144        INTC  mmc0
 92:          1        INTC  musb_hdrc
 93:          0        INTC  musb_hdrc
378:          0     twl4030  twl4030_usb
379:          0     twl4030  rtc0
384:          0     twl4030  mmc0
Err:          0
root@domain.hid:~# ls /dev
block               rtp11               tty29
... snipped ...
mem                 rtp30               tty48
mmcblk0             rtp31               tty49
mmcblk0p1           rtp4                tty5
mmcblk0p2           rtp5                tty50
network_latency     rtp6                tty51
... snipped ...
root@domain.hid:~# umount /dev/mmcblk0p1
root@domain.hid:~# umount /dev/mmcblk0p2

****** I removed the SD card here, no removal msg  *****************

root@domain.hid:~# ls /dev
block               rtp11               tty29
... snipped ...
mem                 rtp30               tty48
mmcblk0             rtp31               tty49
mmcblk0p1           rtp4                tty5
mmcblk0p2           rtp5                tty50
network_latency     rtp6                tty51
... snipped ...
root@domain.hid:~# cat /proc/interrupts
           CPU0
  7:    1106318        INTC  TWL4030-PIH
 12:        180        INTC  DMA
 37:      13933        INTC  gp timer
 56:    4029606        INTC  i2c_omap
 61:          0        INTC  i2c_omap
 74:        440        INTC  serial
 77:      12416        INTC  ehci_hcd:usb2
 83:        364        INTC  mmc0
 92:          1        INTC  musb_hdrc
 93:          0        INTC  musb_hdrc
378:          0     twl4030  twl4030_usb <=== also 0, but not sure if its meaningful
379:          0     twl4030  rtc0
384:          0     twl4030  mmc0      <===== no Chip Detect irpt
Err:          0

Reinserting the SD card did not result in a message. The SD card file systems seemed to be in a half mounted state.

I suspect that this bug effects all interrupts received via twl4030 family devices.

Regards,
Bob Feretich

On 7/22/2010 3:14 PM, Gilles Chanteperdrix wrote: 
... snipped ...

As I said, the absence of IRQ 378 in 2.6.33 looks like a regression,
which should be easy to find with git bisect. Then when you know the
commit which caused the regression, tell me, I will revert it on 2.6.33,
so that I can reproduce the bug you are facing on my board, and
hopefully what happens with the Adeos patch.

--------------040401050406050206060908--