Date: Fri, 6 Oct 2000 08:51:19 -0700 (MST) From: John Reynolds~ <jreynold@sedona.ch.intel.com> To: Warner Losh <imp@village.org> Cc: stable@freebsd.org Subject: Re: breakage with two ed network devices Message-ID: <14813.62711.149661.506143@hip186.ch.intel.com> In-Reply-To: <200010060408.WAA05189@harmony.village.org> References: <14812.58143.609625.133015@hip186.ch.intel.com> <A0E035400B00D4118F9E0008C70D4D77A88A@ITC1> <200010051637.KAA51557@harmony.village.org> <200010060408.WAA05189@harmony.village.org>
next in thread | previous in thread | raw e-mail | index | archive | help
[ On Thursday, October 5, Warner Losh wrote: ]
>
> Wish I knew what to tell you...
>
Yeah ... me too :)
Actually, I sacrificed sleep last night and did binary searches until I
narrowed the kernel problem down to a 6 hour period of commits. What is the
culprit you ask? It HAS to be some ATA commits because those are the only
commits in this "window".
If I CVSup along RELENG_4 and use the date 2000.8.22.06.00.00, rebuild a
kernel, everything works fine. I get ed0 and ed1. But, from there if I change
the CVSup date to 2000.8.22.12.00.00 (i.e. go from 6am sources to noon),
rebuild a kernel, ed0 doesn't show up. These are the only commits to anything
during that time frame on RELENG_4.
Parsing supfile "supfile"
Connecting to cvsup6.FreeBSD.org
Connected to cvsup6.FreeBSD.org
Server software version: REL_16_1
Negotiating file attribute support
Exchanging collection information
Establishing multiplexed-mode data connection
Running
Updating collection src-sys/cvs
Edit src/sys/dev/ata/ata-all.c
Add delta 1.50.2.6 2000.08.22.08.41.28 sos
Edit src/sys/dev/ata/ata-all.h
Add delta 1.26.2.2 2000.08.22.08.41.29 sos
Edit src/sys/dev/ata/ata-disk.c
Add delta 1.60.2.6 2000.08.22.08.41.29 sos
Edit src/sys/dev/ata/ata-dma.c
Add delta 1.35.2.4 2000.08.22.08.41.29 sos
Edit src/sys/dev/ata/atapi-cd.c
Add delta 1.48.2.4 2000.08.22.08.41.29 sos
Edit src/sys/dev/ata/atapi-cd.h
Add delta 1.15.2.3 2000.08.22.08.41.29 sos
Shutting down connection to server
Finished successfully
How could ATA commits screw up the probing of ed0? I wish I had a logical
answer and was hoping to narrow down the "commit window" so that it would show
me something obvious--like in the ed driver itself or in some of the dev/isa
files or something. If I back these commits out, again the kernel probes and
attaches ed0.
I see in the commit logs for these files, there is the comment:
"New improved probe."
Is it possible that this probe is stomping on something on my ed0 card (being
intrusive) and thus the "real" probe for the NIC then doesn't find it? I
looked through LINT and didn't see any "copious debug" option for ATA that I
could enable in the kernel for more "boot -v" messages.
Given THAT information does anybody have any further clues? Again, this
machine is an Abit BP6, two Linksys Ether16 ISA NICs. Kernel config for them
is:
device ed0 at isa? port 0x2c0 irq 15 iomem 0xd8000
device ed1 at isa? port 0x340 irq 9 iomem 0xd8000
Thanks,
-Jr
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
| John Reynolds WCCG, CCE, Higher Levels of Abstraction |
| Intel Corporation MS: CH6-210 Phone: 480-554-9092 pgr: 602-868-6512 |
| jreynold@sedona.ch.intel.com http://www-aec.ch.intel.com/~jreynold/ |
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?14813.62711.149661.506143>