Re: 8.99.41 panic in nvm_poll

2019-06-10 Thread Thomas Klausner 
On Mon, Jun 10, 2019 at 05:38:56PM +0900, Masanobu SAITOH wrote:
> On 2019/06/10 17:17, Thomas Klausner wrote:
> > I tried some more stuff.
> > 
> > Enabling the NVME_QUIRK_DELAY_B4_CHK_RDY quirk didn't help, it paniced
> > during boot. However, forcing nvme_pci_force_intx to 1 makes it boot
> > successfully!
> 
> Please show the following information:
> 
>   - "FULL" dmesg
> 
>   - output of cpuctl list
> 
>   - output of intrctl list
> 
>   - output of pcictl pci0 dump -b X -d Y of the device.

Attached. Thanks for looking at this!
 Thomas

> 
> > Should this variable default to 1, or how do we improve the situation
> > in general?
> > 
> > Cheers,
> >  Thomas
> > 
> 
> 
> -- 
> ---
> SAITOH Masanobu (msai...@execsw.org
>  msai...@netbsd.org)
> 


dmesg.gz
Description: application/gunzip
Num  HwId Unbound LWPs Interrupts Last change  #Intr
   --  -
00online   intr   Mon Jun 10 15:29:23 2019 12   
12online   intr   Mon Jun 10 15:29:23 2019 1
24online   intr   Mon Jun 10 15:29:23 2019 1
36online   intr   Mon Jun 10 15:29:23 2019 0
48online   intr   Mon Jun 10 15:29:23 2019 0
5aonline   intr   Mon Jun 10 15:29:23 2019 0
6conline   intr   Mon Jun 10 15:29:23 2019 0
7eonline   intr   Mon Jun 10 15:29:23 2019 0
810   online   intr   Mon Jun 10 15:29:23 2019 0
912   online   intr   Mon Jun 10 15:29:23 2019 0
10   14   online   intr   Mon Jun 10 15:29:23 2019 0
11   16   online   intr   Mon Jun 10 15:29:23 2019 0
12   18   online   intr   Mon Jun 10 15:29:23 2019 0
13   1a   online   intr   Mon Jun 10 15:29:23 2019 0
14   1c   online   intr   Mon Jun 10 15:29:23 2019 0
15   1e   online   intr   Mon Jun 10 15:29:23 2019 0
16   1online   intr   Mon Jun 10 15:29:23 2019 0
17   3online   intr   Mon Jun 10 15:29:23 2019 0
18   5online   intr   Mon Jun 10 15:29:23 2019 0
19   7online   intr   Mon Jun 10 15:29:23 2019 0
20   9online   intr   Mon Jun 10 15:29:23 2019 0
21   bonline   intr   Mon Jun 10 15:29:23 2019 0
22   donline   intr   Mon Jun 10 15:29:23 2019 0
23   fonline   intr   Mon Jun 10 15:29:23 2019 0
24   11   online   intr   Mon Jun 10 15:29:23 2019 0
25   13   online   intr   Mon Jun 10 15:29:23 2019 0
26   15   online   intr   Mon Jun 10 15:29:23 2019 0
27   17   online   intr   Mon Jun 10 15:29:23 2019 0
28   19   online   intr   Mon Jun 10 15:29:23 2019 0
29   1b   online   intr   Mon Jun 10 15:29:23 2019 0
30   1d   online   intr   Mon Jun 10 15:29:23 2019 0
31   1f   online   intr   Mon Jun 10 15:29:23 2019 0
interrupt id   CPU0  CPU1  CPU2  CPU3  CPU4  CPU5  CPU6  CPU7  CPU8  CPU9  
CPU10  CPU11  CPU12  CPU13  CPU14  CPU15  CPU16  CPU17  CPU18  CPU19  CPU20  
CPU21  CPU22  CPU23  CPU24  CPU25  CPU26  CPU27  CPU28  CPU29  CPU30  CPU31  
device name(s)
ioapic0 pin 9 0*0 0 0 0 0 0 0 0 0  
0  0  0  0  0  0  0  0  0  0  0  0  
0  0  0  0  0  0  0  0  0  0  acpi SCI
msix0 vec 0   0*0 0 0 0 0 0 0 0 0  
0  0  0  0  0  0  0  0  0  0  0  0  
0  0  0  0  0  0  0  0  0  0  xhci0
msi1 vec 0 5395*0 0 0 0 0 0 0 0 0  
0  0  0  0  0  0  0  0  0  0  0  0  
0  0  0  0  0  0  0  0  0  0  ahcisata0
msix2 vec 0   0   125*0 0 0 0 0 0 0 0  
0  0  0  0  0  0  0  0  0  0  0  0  
0  0  0  0  0  0  0  0  0  0  wm0TXRX0
msix2 vec 1   0 089*0 0 0 0 0 0 0  
0  0  0  0  0  0  0  0  0  0  0  0  
0  0  0  0  0  0  0  0  0  0  wm0TXRX1
msix2 vec 2   2*0 0 0 0 0 0 0 0 0  
0  0  0  0  0  0  0  0  0  0  0  0  
0  0  0  0  0  0  0  0  0  0  wm0LINK
ioapic1 pin 12   67*0 0 0 0 0 0 0 0 0  
0  0  0  0  0  0  0  0  0  0  0  0  
0  0  0  0  0

Re: 8.99.41 panic in nvm_poll

2019-06-10 Thread Masanobu SAITOH 
On 2019/06/10 17:17, Thomas Klausner wrote:
> I tried some more stuff.
> 
> Enabling the NVME_QUIRK_DELAY_B4_CHK_RDY quirk didn't help, it paniced
> during boot. However, forcing nvme_pci_force_intx to 1 makes it boot
> successfully!

Please show the following information:

- "FULL" dmesg

- output of cpuctl list

- output of intrctl list

- output of pcictl pci0 dump -b X -d Y of the device.


> Should this variable default to 1, or how do we improve the situation
> in general?
> 
> Cheers,
>  Thomas
> 


-- 
---
SAITOH Masanobu (msai...@execsw.org
 msai...@netbsd.org)

Re: 8.99.41 panic in nvm_poll

2019-06-10 Thread Thomas Klausner 
I tried some more stuff.

Enabling the NVME_QUIRK_DELAY_B4_CHK_RDY quirk didn't help, it paniced
during boot. However, forcing nvme_pci_force_intx to 1 makes it boot
successfully!

Should this variable default to 1, or how do we improve the situation
in general?

Cheers,
 Thomas

Re: 8.99.41 panic in nvm_poll

2019-05-31 Thread Thomas Klausner 
On Wed, May 29, 2019 at 09:04:25AM +1000, Matthew Green wrote:
> i have a change in testing that makes nvme mostly attach after interrupts
> happen.  i need it on arm64 where the interrupt controller isn't setup
> properly yet and attemtping to setup affinity doesn't work.
> 
> i wonder if this would help here, since nvm_poll() would be called later.
> 
> unfortunately, i didn't get interrupts yet so i haven't posted the 
> nvme_attach change for review.  i'll try to do that soon.

I've now managed to get the machine to boot and build kernels by
disabling nvme.

I can test the patch when it's ready.

Thank you!
 Thomas

re: 8.99.41 panic in nvm_poll

2019-05-28 Thread matthew green 
i have a change in testing that makes nvme mostly attach after interrupts
happen.  i need it on arm64 where the interrupt controller isn't setup
properly yet and attemtping to setup affinity doesn't work.

i wonder if this would help here, since nvm_poll() would be called later.

unfortunately, i didn't get interrupts yet so i haven't posted the 
nvme_attach change for review.  i'll try to do that soon.


.mrg.