Re: wm(4): failed to aquire semaphore

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Thu, Oct 22, 2020 at 08:00:23PM +0900, SAITOH Masanobu wrote:
> On 2020/10/22 19:48, Manuel Bouyer wrote:
> > Hello,
> > I have an older desktop system with:
> > wm0 at pci0 dev 25 function 0: 82801I (AMT) LAN Controller (rev. 0x02)
> > wm0: interrupting at msi0 vec 0
> > wm0: PCI-Express bus
> > wm0: 2048 words FLASH, version 1.1.1
> > wm0: Ethernet address 00:1a:a0:d5:a0:0d
> > wm0: 0x2a4480<FLASH,IOH_VALID,PCIE,ASF_FIRM,AMT,WOL>
> > igphy0 at wm0 phy 1: i82566 10/100/1000 media interface, rev. 0
> > igphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto
> > 
> > running netbsd 9.1.
> > Every 5mn, I see in dmesg:
> > [ 78053.063361] wm0: failed to acquire semaphore
> > [ 78053.073383] wm0: wm_gmii_i82544_readreg: failed to get semaphore
> > [ 78353.800448] wm0: failed to acquire semaphore
> > [ 78353.800448] wm0: wm_gmii_i82544_readreg: failed to get semaphore
> > [ 78654.537999] wm0: failed to acquire semaphore
> > [ 78654.537999] wm0: wm_gmii_i82544_readreg: failed to get semaphore
> > [ 78955.275254] wm0: failed to acquire semaphore
> > [ 78955.275254] wm0: wm_gmii_i82544_readreg: failed to get semaphore
> > [ 79256.013287] wm0: failed to acquire semaphore
> > [ 79256.013287] wm0: wm_gmii_i82544_readreg: failed to get semaphore
> > 
> > and network traffic stops for a few seconds.
> > 
> > Does it ring a bell to someone ? Is it a hardware failure or software issue ?
> 
> Does it work without any problem on 9.0?
> If so, could you bisect wm's changes between 9.0 and 9.1?

Before upgrading to 9.1 is was running a 9.0_STABLE kernel from early may.
It had the same issue.

The full story is:
I set up NetBSD on a Dell optiplex 755 early may, it went witout problems
and did the job, running the same kernel since install.
last week-end the hardware died (probably a power supply issue), so I took
the hard drive and plugged it in another optiplex 755, which may not be
strictly identical (different manuf. dates). This problem showed up
immediately. The upgrade to 9.1 was becase of it, but it didn't
change anything.

Next I can try another optiplex 755, I have several of them

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--