[12:15] <BenC> _kylem: Yeah, only patch to sym53c8xx_2/ is the pa6 patch (just checked it)
[12:15] <_kylem> ok.
[12:16] <_kylem> i'll spend some time trying to find the cause on my workstation
[12:16] <_kylem> since it has a front panel lcd to tell me what the erorr is
[12:18] <BenC> I'll try to find out why console isn't working on the ubuntu configs
[12:18] <_kylem> i read through the config you told me to generate but couldn't find anythign that looked suspicious
[12:19] <BenC> obviously it's the config, since a500_defconfig has working console
[12:19] <_kylem> hmm true.
[12:19] <BenC> I don't know enough about the parisc console settings to make since of it though
[12:19] <BenC> s/since/sense/
[12:22] <_kylem> ok
[12:23] <BenC> _kylem: FYI, I was having this problem before applying the pa6 patch
[12:24] <_kylem> there's nothign of consequence in the parisc patchset in the last litle while, iirc
[01:22] <BenC> allocation failed: out of vmalloc space - use vmalloc=<size> to increase size.
[01:22] <BenC> SCSI subsystem initialized
[01:23] <BenC> _kylem: hmm...maybe found something there
[01:23] <_kylem> the vmalloc thing shouldn't be a problem.
[01:23] <_kylem> that's just because a500 have really really big DAC regions
[01:23] <_kylem> (dual address cycle, >4G)
[01:23] <_kylem> hum.
[01:23] <_kylem> is SYM53c8XX using MMIO?
[01:23] <BenC> Linux hippo 2.6.17.3-ubuntu1 #3 SMP Fri Jul 14 23:16:58 UTC 2006 parisc64 GNU/Linux
[01:24] <BenC> interesting...sym2 and sd as modules works
[01:24] <_kylem> er. how'd you manage that? :)
[01:24] <BenC> I think the problem is that sym2 2.2.3 doesn't block during the scsi bus reset
[01:24] <_kylem> ok. i'll look through the patch set. i'm beginning to think it's async scanning.
[01:24] <BenC> and it's not ready when the kernel tries to mount the rootfs
[01:24] <BenC> is that async related?
[01:24] <_kylem> yeah.
[01:24] <_kylem> if async scanning is enabled, you'll see soemthing like this
[01:25] <BenC> makes sense
[01:25] <_kylem> IPv6 over IPv4 tunneling driver
[01:25] <_kylem> NET: Registered protocol family 17
[01:25] <_kylem> scsi: waiting for bus probes to complete ...
[01:25] <_kylem>   Vendor: HP 18.2G  Model: MAN3184MC         Rev: HP04
[01:25] <_kylem>   Type:   Direct-Access                      ANSI SCSI revision: 02
[01:25] <_kylem> [..] 
[01:25] <_kylem> actually. that looks /exactly/ like what's happening.
[01:25] <BenC> doesn't explain the brokeness using the ubuntu config, but atleast I know things should work now
[01:25] <_kylem> sym0: <896> rev 0x7 at pci 0000:00:01.0 irq 19
[01:25] <_kylem> sym0: PA-RISC Firmware, ID 7, Fast-40, LVD, parity checking
[01:25] <_kylem> sym0: SCSI BUS has been reset.
[01:25] <_kylem> scsi0 : sym-2.2.3
[01:25] <_kylem> sym1: <896> rev 0x7 at pci 0000:00:01.1 irq 20
[01:25] <_kylem> sym1: PA-RISC Firmware, ID 7, Fast-40, SE, parity checking
[01:25] <_kylem> sym1: SCSI BUS has been reset.
[01:25] <_kylem> sym1: SCSI BUS mode change from SE to SE.
[01:25] <_kylem> sym1: SCSI BUS has been reset.
[01:25] <_kylem> scsi1 : sym-2.2.3
[01:25] <_kylem> ^-- from a /working/ boot.
[01:26] <_kylem> that's a big hint that something is wrong in drivers/scsi.
[01:27] <BenC> the async problem isn't a big deal for us because our initramfs scripts will wait
[01:28] <BenC> I really need to figure out what's up with the config now
[01:28] <_kylem> righto.
[01:29] <_kylem> gimme a while to grab some food and i'll take a look with you
[01:37] <mdz> BenC: do I understand correctly that linux-source started building linux-kernel-headers just recently?
[01:37] <mdz> BenC: I think the l-k-h produced thereby may be somewhat sub-optimal
[01:37] <mdz> http://librarian.launchpad.net/3419820/buildlog_ubuntu-edgy-i386.totem_1.5.4-0ubuntu1_FAILEDTOBUILD.txt.gz
[01:40] <mdz> BenC: filed bug #53028
[01:44] <_kylem> back.
[03:08] <BenC> mdz: Yeah, that's the issue that I'm trying to get infinity to help me with
[03:08] <BenC> linux-source needs lkh to build, so I need some sort of overrider to use the old lkh so I can produce new ones
[03:09] <mdz> BenC: how about re-uploading the old l-k-h source package with a  higher version number, then overriding it?
[03:09] <mdz> it's saturday for infinity
[03:09] <BenC> hmm, I could try that
[03:09] <BenC> it doesn't use anything in the toolchain to build the package
[03:14] <BenC> mdz: done
[03:15] <mdz> BenC: cool, thanks
[03:15] <mdz> BenC: by the way, the VIA workaround seems to have taken
[03:15] <mdz> I closed the bug accordingly
[03:15] <BenC> good deal, thanks
[03:16] <BenC> I need to get this kernel squared away before the weekend is over so I don't hold up Mith any longer than I already have
[09:47] <crimsun> guh, 34831 strikes again?
[10:56] <BenC> yeah, I saw that
[10:56] <crimsun> man, this one's ugly. Gonna have to look at it in Edgy with the dump mechanism
[10:57] <BenC> next upload In do in edgy will have the kexec/kdump stuff...I hope
[10:58] <crimsun> I'm not entirely convinced it's not a hardware issue