Re: 2.6 kernel on HyperSPARC
- To: Ludovic Courtès <firstname.lastname@example.org>
- Cc: email@example.com
- Subject: Re: 2.6 kernel on HyperSPARC
- From: Jurij Smakov <firstname.lastname@example.org>
- Date: Wed, 7 Jun 2006 21:16:19 -0700 (PDT)
- Message-id: <Pine.LNX.4.63.0606072109370.13308@bobcat>
- Reply-to: email@example.com
- In-reply-to: <20060607210500.GA2330@cassis>
- References: <Pine.LNX.4.63.0605252230090.3640@bobcat> <20060607210500.GA2330@cassis>
Thanks a lot for all the testing!
On Wed, 7 Jun 2006, Ludovic Courtès wrote:
I just grabbed another SS20 with a pair of ROSS RT626 HyperSPARCs .
Your kernel[*] boots fine but the `sunlance' driver does not work and
soon triggers the following error message:
eth0: Memory error, status 88c3, addr 2d6172
Yes, I've seen that too. I don't think it's the kernel problem though, I
have a feeling that it's some problem with udev (Brent reported  that
udev assignes completely bogus interface names to sunlance devices,
causing various problems). In my case sometimes the interface comes up
just fine and then the machine is rock solid, other times any attempt to
bring up the interface results in identical Memory error messages. I'll
play with udev and try to figure out what's going on.
Note that the same kernel with a SuperSPARC 2 (aka. 390Z55) works fine
(and so does it with a 390Z50 as mentioned earlier in this thread).
With the RT626 and the 2.6.16 kernel that's currently available in the
archive[*], the symptoms are quite different. The `sunlance' driver
seems to work fine at first but starts going demented after a single
esp0: SCSI bus reset interrupt
esp0: Resetting scsi bus
(I can see a strange noise coming from the HDD at the same time...)
Joël Bertrand report similar problems not long ago .
Right, I see that too (and the patch fixes it). So at least it's a step in
the right direction.
Jurij Smakov firstname.lastname@example.org
Key: http://www.wooyd.org/pgpkey/ KeyID: C99E03CC