Opened 5 years ago
Closed 2 years ago
#15731 closed bug (invalid)
Unexpected exception “Divide Error Exception” occurred in kernel mode! Error code: 0x0
| Reported by: | Jose64141 | Owned by: | korli |
|---|---|---|---|
| Priority: | normal | Milestone: | Unscheduled |
| Component: | Drivers/Audio/HDA | Version: | R1/Development |
| Keywords: | Crash, boot | Cc: | |
| Blocked By: | Blocking: | ||
| Platform: | All |
Description
I was trying to get some information of other problem I have with GPU, and suddenly I got this. I booted from the legacy loader with CSM in EFI, and when I first tried to run from the USB it worked just fine, but I don’t know why this time it crashed
Attachments (6)
Change History (15)
by , 5 years ago
| Attachment: | 9FD8BC71-E482-4606-9514-74C4909157A2.jpeg added |
|---|
comment:1 by , 5 years ago
comment:3 by , 5 years ago
From what I can tell from the error it seems your HDA controller is showing a zero for CORB size, which really shouldn't happen, but sometimes there are bugs in the hardware.
Please attached a syslog (make sure it has lines with hda: at the beginning) and maybe the output of listdev so we know which HDA controller you have.
Once we know the controller we could check Linux to see if they do something with the CORB value.
comment:4 by , 5 years ago
Now the problem has gone, but I can get a new syslog and a listdev so some research can be done ;p
by , 5 years ago
This is the syslog in the first succesful boot after the crash. Booted the legacy loader from EFI CSM, using UHD and Nvidia GPU.
comment:6 by , 5 years ago
It is weird that this happens inconsistently. In the syslog screen shot the HDA hardware definitely does not look healthy (no codec, no audio group, can't start stream, etc.)
Unfortunately I can't see the vendor and device ID in the part of the syslog you have photographed.
It would be helpful if next time you have a successful boot you open Terminal and run listdev > listdev.txt to get the listdev output into a file, and also cp /var/log/syslog syslog.txt to get the syslog as a text file. Please upload both of those here, if you can. Having the full syslog is useful to see what happens when the boot succeeds, and it will probably also capture the crashing boots (though the Haiku syslog rolls over fairly fast.)
As I noted in my previous comment I am pretty sure what is happening to cause the crash, but it should not happen with normal hardware, so obviously something is buggy in your hardware. If we can get the vendor and device ID we can see if other operating systems do some mitigation.
Even if the hardware is buggy, our driver should not crash the system.
comment:7 by , 4 years ago
indeed the hda driver should handle a zero corb size, and bail out cleanly.
Please write Haiku version.