Re: OpenBSD 7.4 in virtualize env

[jrmu]

> Sometimes, rarely, across multiple version ( did not see it in 7.5 so far )
> the log `scsi_xfer pool exhausted` just get spammed forever,
> 
> It doesn't crash, the device just spam the message , so it s active
> 
> I do not have a way to create the problem , but,
> i wonder if the code could be modified so the device just drop to DDB
>Did you run out of memory / swap perhaps?

I have noticed that occurring when my system runs out of swap space.

-- 
jrmu
IRCNow (https://ircnow.org)


signature.asc
Description: PGP signature

Re: OpenBSD 7.4 in virtualize env

[Stuart Henderson]

On 2024-05-24, Sven F.  wrote:
> --c4123906193364e5
> Content-Type: text/plain; charset="UTF-8"
>
> Hello,
>
> Sometimes, rarely, across multiple version ( did not see it in 7.5 so far )
> the log `scsi_xfer pool exhausted` just get spammed forever,
>
> It doesn't crash, the device just spam the message , so it s active
>
> I do not have a way to create the problem , but,
> i wonder if the code could be modified so the device just drop to DDB

It can, just change the printf to panic.

/sys/scsi/scsi_base.c r1.283 fixed the main thing triggering that
problem, but it was already committed before 7.4

OpenBSD 7.4 in virtualize env

[Sven F.]

Hello,

Sometimes, rarely, across multiple version ( did not see it in 7.5 so far )
the log `scsi_xfer pool exhausted` just get spammed forever,

It doesn't crash, the device just spam the message , so it s active

I do not have a way to create the problem , but,
i wonder if the code could be modified so the device just drop to DDB

something like if this pool is exhausted for "longtime" just crash
(or reboot if sysctl is configured that way )

```
bios0 at mainbus0: SMBIOS rev. 2.8 @ 0xf68b0 (9 entries)
bios0: vendor SeaBIOS version "2:1.10.2-58953eb7" date 04/01/2014
bios0: OpenStack Foundation OpenStack Nova
...
vioscsi0 at virtio1: qsize 128
scsibus1 at vioscsi0: 255 targets
sd0 at scsibus1 targ 0 lun 0: 
```

I will upgrade to 7.5 soon anyway

Best.