kernel crash with lite-on sata controller Dell XPS 13 9343
Hi all
I have been asked to post here based on the bug I filed at Ubuntu:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1516215
I have the newest kernel running:
cat /proc/version
Linux version 4.5.0-040500rc5-generic (kernel@tangerine) (gcc version 5.2.1 20151010 (Ubuntu 5.2.1-22ubuntu2) ) #201602201730 SMP Sat Feb 20 22:32:16 UTC 2016
Since the beginning (started with 3.19.x kernels on this notebook) I had problems with this sata controller:
03:00.0 SATA controller: Lite-On Technology Corporation Device 0224 (rev 10)
Very often during boot, the boot just failed with error messages like
ata1.00: exception... (I was only able to capture them by taking a picture: https://launchpadlibrarian.net/227876135/IMG_20151128_230313.jpg)
Then I keep trying, sometimes I boot into Windows inbetween, and eventually, the kernel comes up.
Then, when running and putting a lot of load on the disk (e.g. running dump or a simple update via apt-get upgrade), the system looses its disk and I have to hard reset.
The machine runs perfectly fine under Windows, so I don't believe it is a hardware issue.
I hope there is a solution to this issue.
best regards
Carolin
Hello, Carolin.
On Thu, Feb 25, 2016 at 10:45:10PM +0100, Carolin Latze wrote:
...
Quoted 13 lines
Since the beginning (started with 3.19.x kernels on this notebook) I had problems with this sata controller:
03:00.0 SATA controller: Lite-On Technology Corporation Device 0224 (rev 10)
Very often during boot, the boot just failed with error messages like
ata1.00: exception... (I was only able to capture them by taking a picture:
https://launchpadlibrarian.net/227876135/IMG_20151128_230313.jpg)Then I keep trying, sometimes I boot into Windows inbetween, and eventually, the kernel comes up.
Quoted 8 lines
Then, when running and putting a lot of load on the disk (e.g. running dump or a simple update via apt-get upgrade), the system looses its disk and I have to hard reset.
The machine runs perfectly fine under Windows, so I don't believe it is a hardware issue.
I hope there is a solution to this issue.
Can you please boot with libata.force=noncq and if that fails post the kernel messages of the failure?
Thanks.
Signature · 1 line
tejun
Hi Tejun
I actually boot with "libata.force=noncq,1.5G". Sorry, forgot to mention that. Same behavior.
best regards
Carolin
Quoted 30 lines · Tejun Heo
Hello, Carolin.
On Thu, Feb 25, 2016 at 10:45:10PM +0100, Carolin Latze wrote:
...Quoted 13 lines
Since the beginning (started with 3.19.x kernels on this notebook) I had problems with this sata controller:
03:00.0 SATA controller: Lite-On Technology Corporation Device 0224 (rev 10)
Very often during boot, the boot just failed with error messages like
ata1.00: exception... (I was only able to capture them by taking a picture:
https://launchpadlibrarian.net/227876135/IMG_20151128_230313.jpg)Then I keep trying, sometimes I boot into Windows inbetween, and eventually, the kernel comes up.
Quoted 8 lines
Then, when running and putting a lot of load on the disk (e.g. running dump or a simple update via apt-get upgrade), the system looses its disk and I have to hard reset.
The machine runs perfectly fine under Windows, so I don't believe it is a hardware issue.
I hope there is a solution to this issue.
Can you please boot with libata.force=noncq and if that fails post the kernel messages of the failure?
Thanks.
Hello, Carolin.
Quoted 4 lines · Carolin Latze
Hi Tejun
I actually boot with "libata.force=noncq,1.5G". Sorry, forgot to mention that. Same behavior.
Can you please post screen capture from that failure?
Thanks.
Signature · 1 line
tejun
A message that is not in the archive (sent off-list or not yet received): [email protected]
Hello, Carolin.
Quoted 2 lines · Carolin Latze
please see attached screen capture. If there is any better way to capture those logs than taking pictures, please let me know
Heh, so there have been quite a few cases where device locks up after a lot of NCQ commands are issued and/or data transfer is taking place at high rate but device locking up on plain WRITE_DMA_EXT is pretty new. Hmmm.... Does the following patch make any difference?
drivers/ata/ahci.h +2 -2
@@ -366,8 +366,8 @@ extern struct device_attribute *ahci_sdev_attrs[];#define AHCI_SHT (drv_name) \ ATA_NCQ_SHT(drv_name), \ .can_queue = AHCI_MAX_CMDS - 1, \ - .sg_tablesize = AHCI_MAX_SG, \ - .dma_boundary = AHCI_DMA_BOUNDARY, \ + .sg_tablesize = LIBATA_MAX_PRD, \ + .dma_boundary = ATA_DMA_BOUNDARY \ .shost_attrs = ahci_shost_attrs, \ .sdev_attrs = ahci_sdev_attrs
Hi Tejun
just a quick update from my side: I did patch the rc5 kernel yesterday (I am aware that there is a newer one but when you sent the patch, rc5 was recent so I wanted to be on the safe side). There is a little typo on the "dma_boundary" line, the trailing comma is missing. But other than that, the patching went smoothly. Since then I reboot whenever I can. So far 15-20 times and no issue up until now. I will keep rebooting very often for the next days and keep you posted.
best regards
Carolin
Quoted 25 lines · Tejun Heo
Hello, Carolin.
Quoted 2 lines · Carolin Latze
please see attached screen capture. If there is any better way to capture those logs than taking pictures, please let me know
Heh, so there have been quite a few cases where device locks up after a lot of NCQ commands are issued and/or data transfer is taking place at high rate but device locking up on plain
WRITE_DMA_EXTis pretty new. Hmmm.... Does the following patch make any difference?drivers/ata/ahci.h +2
-2@@ -366,8 +366,8 @@ extern struct device_attribute *ahci_sdev_attrs[];#define AHCI_SHT (drv_name) \ ATA_NCQ_SHT(drv_name), \ .can_queue = AHCI_MAX_CMDS - 1, \ - .sg_tablesize = AHCI_MAX_SG, \ - .dma_boundary = AHCI_DMA_BOUNDARY, \ + .sg_tablesize = LIBATA_MAX_PRD, \ + .dma_boundary = ATA_DMA_BOUNDARY \ .shost_attrs = ahci_shost_attrs, \ .sdev_attrs = ahci_sdev_attrs