[OpenIndiana-discuss] Nvidia errors on 151a4

[Bryan N Iotti]

OK, it was just an illusion: despite the new video card, the hard 
freezes are there with Firefox still, but all the console messages about 
NVRM are gone.

I have Firefox 12, the tarball version downloaded from the contrib 
section of the mozilla ftp site.

It appears to crash randomly, I still haven't found a pattern. I used to 
think it was when it would load flash or java as a plugin, maybe a 
permissions issue, but even with text only pages it will grind to a halt.

I installed the NVidia drivers version 295.49 from the NVidia website by 
following the procedure from

http://defect.opensolaris.org/bz/show_bug.cgi?id=12196


That means I have a new BE where the OI drivers have been uninstalled 
and replaced with the NVidia ones (whish they'd at least make a serious 
attempt at packaging them, without us having to fiddle to get stuff 
working, but oh well...)

Now, I have looked at my interrupt table by "pfexec mdb -k" and then 
"::interrupts":

IRQ  Vect IPL Bus    Trg Type   CPU Share APIC/INT# ISR(s)

----------------------------------------------------------------
1    0x41 5   ISA    Edg Fixed  0   1     0x0/0x1   i8042_intr

9    0x80 9   PCI    Lvl Fixed  1   1     0x0/0x9   acpi_wrapper_isr

12   0x42 5   ISA    Edg Fixed  1   1     0x0/0xc   i8042_intr

15   0x43 5   ISA    Edg Fixed  2   1     0x0/0xf   ata_intr

16   0x82 9   PCI    Lvl Fixed  3   4     0x0/0x10  hci1394_isr, uhci_intr, uhci_intr, nv_intr

17   0x60 6   PCI    Lvl Fixed  3   1     0x0/0x11  e1000g_intr

18   0x85 9   PCI    Lvl Fixed  2   3     0x0/0x12  uhci_intr, ata_intr, ata_intr

19   0x84 9   PCI    Lvl Fixed  1   1     0x0/0x13  uhci_intr

23   0x83 9   PCI    Lvl Fixed  0   1     0x0/0x17  ehci_intr

72   0x81 7   PCI    Edg MSI    2   1     -         pcieb_intr_handler

73   0x40 5   PCI    Edg MSI    3   1     -         mpt_intr

74   0x30 4   PCI    Edg MSI    2   1     -         pcieb_intr_handler

75   0x44 5   PCI    Edg MSI    3   1     -         mpt_intr

160  0xa0 0          Edg IPI    all 0     -         poke_cpu

208  0xd0 14         Edg IPI    all 1     -         kcpc_hw_overflow_intr

209  0xd1 14         Edg IPI    all 1     -         cbe_fire

210  0xd3 14         Edg IPI    all 1     -         cbe_fire

240  0xe0 15         Edg IPI    all 1     -         xc_serv

241  0xe1 15         Edg IPI    all 1     -         apic_error_intr


So it would appear that my video card does not share interrupts with the 
e1000g. If you want me to try, I have a 3com Gigabit PCI card I can pop 
in and see if that makes a difference, disabling the internal one. I 
can't remap IRQs on this machine, but I can disable devices.

I find it very frustrating that a single process, as relatively trivial 
as a web browser (our only one, more or less, to make matters worse) can 
bring down the whole system. At least ZFS seems to not care.

I might try to compile it from source using the Solaris Studio compiler, 
maybe that can settle whatever issue it's running into, or at least make 
some noise about why it's unhappy.

Any other ideas? I'm not a good debugger, but send some commands and 
I'll run them and post results.

Thank you for all the answers so far.

Bryan

On 05/ 7/12 08:51 PM, Marion Hakanson wrote:
> ironsides.medvet at gmail.com said:
>> I had to install 295.49 following the procedure from
>>   http://defect.opensolaris.org/bz/show_bug.cgi?id=12196
>>
>> I have a new boot environment with the updated drivers and it would  appear
>> that the hard freezes have stopped.
> I wonder if the hard freezes were like what I have experienced in:
> 	https://www.illumos.org/issues/1625
>
> If you're experiencing the same issue as the above, and if on your machine
> the nVidia card shares interrupts with your ethernet interface, you wouldn't
> be able to login from some other machine to do any diagnosis.
>
>
>> However, it is still  littering my
>> console with these two messages in repetition:
>>
>> May  6 10:50:01 Wraith nvidia: [ID 702911 kern.notice] NOTICE: NVRM: nv_devmap: devmap_devmem_setup() failed (22)
>> May  6 10:50:01 Wraith nvidia: [ID 702911 kern.notice] NOTICE: NVRM:  nv_segmap: ddi_devmap_segmap() failed (6)
>>
>> Can I do anything about that?
> On my systems, this started with nVidia driver version 295.40.  If you
> revert to 295.33 the above messages go away.
>
> Regards,
>
> Marion
>
>
>
> _______________________________________________
> OpenIndiana-discuss mailing list
> OpenIndiana-discuss at openindiana.org
> http://openindiana.org/mailman/listinfo/openindiana-discuss