Please track all bugs related to the beta livecd here. This includes kernel configs, spec files, etc.
Apologies for opening a new bug, I misunderstood and thought this bug would be used as a "tracker" (e.g. like bug 49494) Here is my problem again: I'm using install-x86-20040518beta-2004.2.iso My main desktop system will not boot from this CD. It goes through bootup as normal.. looks for a CDROM to boot off: Boot from ATAPI CDROM... (at this point, you can hear the cd spinning up) Boot from ATAPI CDROM... (it now goes onto the next boot devices) Boot from Serial ATA... And then my own grub menu comes up and it starts booting from the hard disk. I run an nforce2 motherboard, and have tried both a DVD reader, and a CD writer (both IDE) and it goes through as if the disc in the drive is not bootable. The 2004.1 and 2004.0 livecd's boot up fine in the same configuration. MD5sum of downloaded .iso file matches the .iso.md5 file, running "md5sum /dev/dvd" also produces the correct md5. Strangely enough, the CD does boot fine on my laptop.
Created attachment 31664 [details] Work in-progress kernel config
SCSI Disk/CDROM for USB/PCMCIA CDROM? Let's find out and bug genkernel@ if we need it...
I have merged my config with wolf31o2's gamecd config. We agreed to leave out preempt since that is only (if at all) needed on gamecds and can be easily added to a gamecd config. devfs mount at boot was included to get rid of some dev mounting error at boot time. merged-20040518-gentoo-2.6.5-r1-config
Created attachment 31675 [details] merged-20040518-gentoo-2.6.5-r1-config
Created attachment 31676 [details] 20040518-x86-gentoo-2.6.5-r1-smp.config
So like... roger and I posted at the same time... they're the same kernel... mine's just named better..... ;]
Comment on attachment 31675 [details] merged-20040518-gentoo-2.6.5-r1-config obsoleted by chris's config (i know they are the same, I am just keeping things tidy ;)
Okay, this one boots. That is an improvement - check my devspace for the updated iso for testing. (http://dev.gentoo.org/~zhen/2004.2) I tried the livecd on some different boxen and came up with some different bugs. Looks like the problems lie in the gentoo (2.4) kernel. I did not test the smp kernel since the only boxen that I have are uniproc. Box: P4 1.80 82845G/GL Chipset w/ integrated vid (same num on the vid) -scrolling non-fatal error at boot; error is: hw_random i810_rng: RNG not detected -right after the boot prompt, vesafb asks for a valid mode number, the one that we are using does not work. Could be due to the fact that it is an LCD monitor, but I am not sure that matters. Boxen: Same P4 as above, AMD XP KT400 chipset -reboot hangs on "Stopping USB Hotplugging". Both boxen have usb 2.0, and the non-usb-2.0 boxen did not have this problem. If I hit enter, I am dumped back to a shell and can issue a "reboot" again, but this time successfully. The attachment below is the 2.4 config that I have modded to include auto-mounting of devfs so that the initrd does not complain. Livewire - any idea on the bugs?
Created attachment 31693 [details] updated 2.4.25 config updated 2.4.25 config, added automount of devfs
Take the random number generators out of the kernel totally, or make them a module (i810_rng error). the vesa error is because of the lcd, tried passing another vga= argument? Any reason your not using vga=791 now? issue a "/etc/init.d/hotplug stop" to see if you can debug it better
Ok, my experiences with install-x86-20040518beta-rev1-2004.2.iso : ---- When booting up into the "gentoo" kernel, I get errors loading modules. The first errors relate to loading "forcedeth". A manual modprobe of "forcedeth" shows errors relating to "device not present". This module fails to load because the "nvnet" driver has already been loaded. In summary, nvnet and forcedeth both support the same devices (nforce ethernet). nvnet is nvidia's buggy driver, forcedeth is a reverse-engineered driver included in the kernel. I vote to drop nvnet. (The "smp" kernel does not get affected by this issue, it tries to load forcedeth and not nvnet, and succeeds). ---- The second module load error relates to Serial ATA. The broken siimage driver is claiming my SATA device, and then the working sata_sil module tries to be loaded. Manually modprobing shows similar errors as the forcedeth module did - device is already claimed. I have proposed a solution to this in bug 51464, this involves patching gentoo-sources and gentoo-dev-sources. Another bit of misbehaviour I observed: When booting up "gentoo", the siimage driver creates /dev/hde[1-7] with my partitions. When booting up "smp", the siimage driver claims in dmesg to create /dev/hde, but this device and the partitions do not appear in /dev. ---- When booting up into "gentoo", the LiveCD correctly detects presence of firewire hardware, and loads the relevant modules, including "eth1394". eth1394 creates eth0. Later on in bootup, the nvnet module is loaded for my NIC adapter, which is created as eth1. Gentoo then prints out the "Network device eth0 detected, dhcp broadcasting for IP..." and quickly continues. The issue here is that dhcpcd can not run on eth1394 style connections (check the system log), so this seems pretty pointless if the firewire modules will always be loaded first. Better behaviour would be to run dhcpcd for the first ethernet network devices found. ---- I also noticed that the eth1394 module was present under the "gentoo" kernel but not "smp". Are we aiming for consistency here? If so, the module should probably be compiled for both kernels. ---- On the "smp" kernel, when plugging in my usb storage device, the dmesg log got completely filled up. Is there any reason why we have USB verbose debug enabled? The "gentoo" kernel does not exhibit this verbose behaviour. ---- I noticed two big logfiles in /var/log: emerge.log and genkernel.log Do these serve any purpose for being there?
okay, building a new test iso after I get the updated 2.6 config - here are the changes that I will make: -dropping nvnet in favor of forcedeth. forcedeth seems to be the more logical choice after discussing with some other devs. --do we drop siimage? Do we need it? IIRC, you said in IRC that it is needed for some IDE controllers. We either get gentoo-sources officially patched, or we drop it entirely. Its not good practice to patch as we go :) --as for eth1394 - this may be a baselayout fix .... aron, any idea how to address this issue? -2.6 kernel peeps can take care of the module consistency and usb verbose stuff -I can have catalyst remove emerge.log and genkernel.log Again, as soon as I hear from the 2.6 kernel people, I will rebuild :) Thanks again for the feedback!
siimage? Actually, this is being patched in the next gentoo-sources (thanks plasmaroo). Hopefully, we will get the same on the gentoo-dev-sources ebuild. The idea is to patch *out* SATA support from siimage if USE=livecd during the ebuild. I will attach a new kernel config below.
Created attachment 31713 [details] 20040518-x86-smp-gentoo-2.6.5-r1.config Here's the newest kernel version minus "USB Mass Storage Verbose", HW_RANDOM, and with eth1394 added back in to match 2.4 config.
BTW: siimage stuff can be followed here: http://bugs.gentoo.org/show_bug.cgi?id=51464
aacraid module not found on LiveCD 2004.1 with smp was booted with 2.6 kernel On x86hardware
This bug is really for our beta CDs, which we have been working on, but to answer your question, if you look into the kernel config attached to this bug, you will see that aacraid is indeed included in the new smp (and up) kernels. Thanks for the input!
Created attachment 31766 [details] 20040520-gentoo-2.4.25-r2-config new 2.4 config that removes HW randomization
Created attachment 31774 [details] 20040518-x86-smp-gentoo-2.6.5-r1.config I forgot the devfs automount...
I found a bug in the live cd beta, if I boot it up it shows the error message 'can't load module 8139cp, missing kernel or user mode driver' this error message is repeated about 6 times and then it boots up normaly, this is particulary strange since I normaly use 8139too but it could be that both drivers would work for my network card.
please give us the exact error here, as well as what hardware you have in that box.
Tested 2004.2-test2 on Dell PowerEdge 700 (a popular new low-end dell model): booted with "smp doscsi" - correctly detected aic7xxx and e1000 hardware - kernel oops during hotplug, apparently when attempting to load aic79xx - hotplug also segfaulted during usb scan - hyperthreaded cpu reported as only one cpu - did manage to get to the command prompt (which is a major improvement over 2004.1) Moved on Dell PowerEdge 1750 (old reliable mid-range box): booted with "smp doscsi" - appeared to lockup completely when "scanning for eata" - unable to test detection of dual xeon cpu's or megaraid/tg3 Finally, took a stab at a a Dell Poweredge 750 (rackmount equiv of Dell 700): - appeared to crash completely when "scanning for eata" - appeared to correctly load aic7xxx - unable to test detection of hyperthreaded cpu or e1000 driver
Do those Dell's work in Linux at all? Have you also tried them with the "gentoo" kernel? I am curious if it is just a 2.6 issue or what.
All the dells do work under linux fine. I've typically managed to get them running gentoo by booting an old livecd w/ "gentoo nohotplug nousb" and manually loading the necessary drivers. After install and using a good kernel, they run great. 2004.1 didn't work very well with them. 1.4 has no issues with the 1750's. The testserver I've lent to beejay for releng testing is a Dell 700. We really should be able to get these guys going w/ smp and apic support during the install process though. It sucks big time to not be able to use SMP/etc during installs.
<tigger^> its the LSI / Symbios 53c1010 <tigger^> And it uses the sym53c8xx module Apparently, this is causing problems with doscsi under the "gentoo" kernel.
Testing out 2004.2 Test2, only had one problem. When booting smp doataraid doscsi, the box would freeze when attempting to detect megaraid devices. I have an LSI Logic MegaRAID SATA150-6 Controller in a Penguin Computing 230S Box. Works fine though if I don't use doscsi or doataraid. - anti
test1 boot disk with my system (kt400 via 8v no nforce onboard bcm4401) requires a noapic with gentoo-nofb. smp kernel (2.6, even though the syslinux bootup says nothing about this fact, would be nice to note :)) actually doesn't start up without fb when started with smp-nofb. and the version of the smp kernel appears to be such that my fb doesn't work. nor does it work with 'gentoo' incidently. a test version provided by roger55 (thanks!) now displays working framebuffer yay. and starts up well. unfortunately both the 2.4.25 kernels and the 2.6.5 kernels have buggy drivers for my ethernet card (module == b44). luckily this is fixed in later kernels as per a patch which jgarzik has now applied in bk. http://seclists.org/lists/linux-kernel/2004/May/3223.html hope this helps someone. Alex
the first paragraph of Comment #28 From Alexander Kellett refers to 2004.1 cd. and the 'provided by roger55' part refers to wolf's test2 livecd - I provided the link only. (yay me!) http://dev.gentoo.org/~wolf31o2/releng/gentoo-2004.2-test2.iso <--- this one just for clarification. roger55
Thanks... I'm going to be posting up some updates very shortly... I have made a few changes to both catalyst and to the kernel config and spec files which need adjustment for a good build. I'll also be posting up a -test3 very shortly. I am hoping to get the squashfs problems with 2.4 resolved before making a -test3 release. If I switch back to zisofs, it works perfectly, though.
Created attachment 32447 [details] livecd/runscript-support/livecdfs-update.sh This fixes a problem with the sed for the bootsplash theme...
Created attachment 32448 [details] livecd/runscript-support/kmerge.sh This has a modified genkernel line to use the new --bootsplash option from genkernel 3.0.2b
Created attachment 32449 [details] livecd/runscript/x86-archscript.sh This is the new x86-archscript to take advantage of the tarball naming change in genkernel 3.0.2b and the changes would need to be applied to any other arch that uses genkernel for kernel building.
Created attachment 32450 [details] 20040518-x86-gentoo-2.4.25-r3.config Kernel config for 2.4 kernel with siimage as module and with squashfs enabled
Created attachment 32451 [details] 20040518-x86-smp-gentoo-2.6.5-r1.config Kernel config for 2.6 with siimage as a module
Created attachment 32452 [details] x86-livecd-stage1.spec Here is my modified stage1 spec that I am using to build my -test3 CD
Created attachment 32453 [details] x86-livecd-stage2.spec Here is my modified stage2 spec that I am using to build my -test3 CD
chris, please post diffs for those modified catalyst files :)
Created attachment 32458 [details, diff] livecd/runscript/x86-archscript.sh.diff Patch version
Created attachment 32459 [details, diff] livecd/runscript-support/kmerge.sh.diff Patch version
Created attachment 32460 [details, diff] livecd/runscript-support/livecdfs-update.sh.diff Patch version
tested 2004.2-test3 on a dell 700 w/ aic7xxx scsi and e1000 nic. smp dscsi - system locked up during megaraid module loading (soft reboot worked) smp doscsi no hotplug nousb - same behavior as above smp - oops at hotplug aic79xx, segfault at modprobe ohci-hcd, reboot hangs during termination of usb, had to physically cycle power smp nohoutplug nousb - loads up fine, no probs, detects everything - except that cpu shows up as non hyperthreaded, soft reboot OK.
Tested 2004.2-test3-zisofs with smp-nofb on IBM Netfinity 7000 M10 quad xeon P3/550/1M, everything detected ship-shape: ips (serveraid-3hb), pcnet 10/100 lan, nc3122 lan (dual e100), netelligent dual thunderlan. We only lack ASMA support, but that's only useful after installing ;-)
dell 750's (which are like rackmount versions of the 700's, uses aic7xxx and e1000) - doscsi is just as broken but for a different reason, always hangs on "scanning for eata"...requires hard reboot (power cycle), the smp behavior is the same as the 700's...essentially kernel oops, segfault, failed reboot....works fine w/ smp nohoutplug nousb just like the 700's, except hyperthread borked
results on the final dell model I tested today - the 1750's which are the mainstay general purpose boxes (dual xeon) For the 1750's, behavior was similiar to the 750's w/ doscsi....everything borks on the "eata" module loading....moving to smp, it works perfectly, even w/ hotplug and usb turned on....hyperthread is also seen??? so each of the three models reacted differently, although all agreed that doscsi doesn't work.
testing: 2004.2-test3-zisofs gentoo dokeymap dopcmcia nodhcp (thinkpad 760EL pentium1) dokeymap spits the following error: cat: /lib/keymaps/keymapList: No such file or directory still asks for a number. I gave it 10 (german) since I *know* the number. worked - kindof: umlauts
testing: 2004.2-test3-zisofs gentoo dokeymap dopcmcia nodhcp (thinkpad 760EL pentium1) dokeymap spits the following error: cat: /lib/keymaps/keymapList: No such file or directory still asks for a number. I gave it 10 (german) since I *know* the number. worked - kindof: umlauts öäü don't work and didn't in the 2004.1 release pcmcia worked. -------------------------------- testing: 2004.2-test3-zisofs gentoo dokeymap nodhcp (on celeron 466 mhz) dokeymap see above. several errors about not being able to find 8139cp module (pci hotplugging iirc) bootsplash: no progressbar on startup, verbose mode needs a better quality jpg (artefacts) different verbose splash at the end of booting (old one with larry the cow) on shutdown yet another silent splash, this time with a working progressbar. shutdown hangs at stopping usb hotplugging. reissuing the command then works. testing: 2004.2-test3-zisofs smp nodhcp dokeymap (celeron 466mhz) dokeymap see above *no* errors about 8139cp bootsplash see above shutdown works flawlessly. ------------ Note my isa nic wasn't detected in any case modprobing 3c509 works just fine. It was detected in the 1.4 releases. Maybe Livewire can reveal the secret to this. ;) Will test some more this weekend. roger55
Roger: I'm going to assume that this means that you also have the keymaps on your local machine, correct? Could you let me know what ebuild is providing them?
Ok here come the keymaps used on my system: roger@miliker roger $ grep KEYMAP= /etc/rc.conf KEYMAP="de-latin1-nodeadkeys" and for consolefont I have: CONSOLEFONT="lat9w-16" roger@miliker roger $ qpkg -f /usr/share/consolefonts/lat9w-16.psfu.gz sys-apps/kbd * roger@miliker roger $ qpkg -f /usr/share/keymaps/i386/qwertz/de-latin1-nodeadkeys.map.gz sys-apps/kbd *
do I need to add kdb to the livecd-stage1 template then?
Actually, I think it was a genkernel bug and I'm pretty sure that plasmaroo has already taken care of it in his CVS for the next genkernel. I am just waiting on a new tarball from him to test.
Roger: Could you see what happens if you don't specify the console font? I'd like to see whether this is due to a lack of a console font; or a faulty keymap... Thanks.
The SMP kernel made from install-x86-20040518beta-rev1-2004.2.iso does not see the IDE hard drives (hda and hdb) in my test machine (Brand new Pengiun Computing Relion 1X) but sees the CD-ROM (hdc) just fine. Inserting the ide-disk kernel module makes it work just fine.
This is kinda to the devs, but isn't that the expected behavior if ide-disk is a module? I don't think that hotplug would load it. What about trying to manually mount it? Would that load the module? If not, why not? Isn't that what devfs is all about? If it isn't working, how do we fix it?
why is ide-disk a module to start with?
I don't even remember, but I think it was my fault. I have changed it in my local kernel config to be statically compiled, though.
Just tried the gentoo-2004.2-test3.iso. It doesn't load my IDE disks but loads my IDE cdroms. The SMP kernel doesn't detect my P4 as dual processors. However it doesn't reboot after a couple of minutes like the 2004.1 SMP does.
I think I might have fixed the non-detection of hyperthreading, but I've got to do more testing on it. The -test3 release has ide-disk as a module, so it won't load them. I've fixed this in the upcoming -test4 CD, which should be hitting my desv space sometime in the next week.
irssi doesn't work - it complains about libgmodule-2.0 being not there, which it isn't - 1.2 is the version included.
MSI K8T Neo-FIS2R for AMD64 Realtek TRL8169/8110 Gigabit (Onboard) is 'not detected'. modprobe r8169 fails, complaining of not finding the device.
Jack, which CD are you using?
gentoo-2004.2-test3.iso. Just got it yesterday.
Well, there's already a new .iso (-test4) that you can get from the /experimental/x86/livecd directory on any Gentoo mirror that carries /experimental. I am moving away from using my dev space simply because of the number of people now testing these CDs.
No dice. Same error, modules don't load because it says it cant find any device. gentoo-2004.2-test4.iso 71324 KB 6/17/2004 12:06:00 AM
hey guys... i tried wolf's test4 livecd last night, and had booted into the 2.4 kernel, believing it would detect things better ;) but it ends up i needed the prism wireless pci card driver, which would only work properly with the smp kernel (sorry not to be too specific at this time, i just can't remember what the error was... something at pci hotplugging in 2.4 kernel saying a certain driver could not be found, which i presume was a dependency of the prism_pci2 driver...) anyway, it's not a big deal, since i was able to make everything go smoothly in the smp kernel, which would have been my first choice anyhow due to its speed ;) ...also, i found that some library (libstdc or libstdio or something iirc :-/) seemed to be missing (at least in the smp kernel, don't know about 2.4), which prevented irssi to run, therefore i ssh'ed into my own box to go onlint :-D ...so that's that... i hope this helps a tad... i can reproduce the 2 bugs later on today if you would like, since i'm returning to my friend's to finish up the install... thanks! good day!
Thanks for the report. I would definitely need the exact message with the prism2 driver to solve that one. For the irssi this, it was libgmodule that was missing, and I'm not sure why, since we don't remove it. I will have to investigate further when I get a chance.
okay, I have some PCMCIA oddness here on the test4 livecd :) First of all, the notebook is a dell inspiron 8000 p3 650 MHz w/ a Texas Instruyments PCI4451 Cardbus Controller. The card that I am trying to use is a Xircom Cardbus Ethernet 10/100 (rev 03) (models grepped from lspci). Under our 2.4 kernel (gentoo dopcmcia), cardmgr detects the two slots, tries to enable the card, and then hardlocks. I can boot again w/ dopcmcia, this time the card removed, and upon insertion of the card, it still hardlocks. Under smp (smp dopcmcia), cardmgr says that there are no slots found, but the xircom network card is properly detected and assigned an address. Are these kernel inconsistencies or more 2.4 stupidness?
zhen: I'm not really sure at this point. I am currently working on getting a 2.6-only CD out, but unfortunately, the gentoo-dev-sources guys thought it would be a good idea to remove squashfs from the 2.6.7 versions. I am requesting to have it added back so that I can test further.
Created attachment 34347 [details] 20040622-x86-smp-2.6.7-r6.config
The version of reiserfsprogs on the test-4 livecd is 3.6.11. It seems that version is too old to know how to support badblocks, limiting its' usefulness as a recovery tool for reiserfs users. Should the version on the CD be bumped to 3.6.17, the current version so that checking badblocks is possible on reiser partitions?
All applications are whatever was marked stable at the time of the snapshot. I have added a -test5 CD. Does it exhibit the same problem? (Sorry, not at my dev box, so I can't tell you)
Try my -test5 CD. It resolves the reiserfs-progs problem.
Created attachment 34408 [details] x86-livecd-stage1.spec Here is the livecd-stage1 spec file I used in building my -test5 CD.
Created attachment 34409 [details] x86-livecd-stage2.spec Here is the livecd-stage2 that I used. It uses the 20040622 smp kernel (only) and has speakup support. I'm going to get dmwaters to try out the speakup support, then we'll need to find someone with ISDN. Also, during the build of the kernel, iptables, fcapi, and xfsdump did not build properly.
Created attachment 34410 [details] 20040518-x86-gentoo-2.4.26-r1.config Newest 2.4 kernel config (used in -test4)
test-5 works re: reiserfs badblocks.
test6 iso: gentoo dopcmcia nodhcp dokeymaps : works smp dopcmcia nodhcp dokeymaps: doesn't work fix: add/replace in livecdstage1: catalyst_work/tmp/default/livecd-stage1-x86-2004XXXXXX/usr/share/genkernel/x86/modules_load : PCMCIA_MODULES="pcmcia_core i82365 yenta_socket ds ide-cs" for 2.6 kernels please test if that needs to be conditional for 2.6.X / if it breaks 2.4.X this is the only issue I found on my laptop so far. will test some more on my desktop lateron.. roger55
Tested a bit -test7 livecd, here are some notes: - Trying to mount a nfs share results in a long stall, until eventually an RPC timeout occurs. Nonetheless, the share is mounted correctly. Solution: run portmap before trying to mount, that is, add portmap to the default runlevel. - gpm is running after boot, however the mouse does not work out of the box for me. To make it work, I have to 'modprobe psmouse'. So, I suggest to set CONFIG_MOUSE_PS2=y in kernel config. - The options showed by pressing F2 are a little screwed up, the description of 'docache' has suffered a copy/paste error. Moreover the option 'doapm: disables apm module load' is confusing, should it be 'noapm'? - there are no manpages on the livecd, is it meant to be so?
- Trying to mount a nfs share results in a long stall, until eventually an RPC timeout occurs. Nonetheless, the share is mounted correctly. Solution: run portmap before trying to mount, that is, add portmap to the default runlevel. Not going to happen, simply because we do not want *anything* listening by default. If you need to mount NFS volumes, you should /etc/init.d/nfsmount start before doing so. Is this not documented in the handbook? If not, I'll be sure to have them add it before 2004.2, so let me know. - gpm is running after boot, however the mouse does not work out of the box for me. To make it work, I have to 'modprobe psmouse'. So, I suggest to set CONFIG_MOUSE_PS2=y in kernel config. Done. This was already the case with the 2.4 kernel (only on -test6), but not the 2.6 kernel. Thanks for letting me know. - The options showed by pressing F2 are a little screwed up, the description of 'docache' has suffered a copy/paste error. Moreover the option 'doapm: disables apm module load' is confusing, should it be 'noapm'? Fixed. - there are no manpages on the livecd, is it meant to be so? Yes. The man pages are enormous. We remove them on purpose, as it increases the compressed CD size by upwards of 20MB. I am working on an issue at the moment, that once fixed, I will release a new -test8 and -test9 CD set. These will *hopefully* be the last set released before 2004.2, since we're getting ever closer to the freeze.
Thanks for the fixes and clarifications, I'm looking forward for -test8 and -test9.
as per bug #43141, hotplug still doesn't pick up cards that use the sk98lin driver. this is with test7 as well as brand new livecds I've built myself today (based off the specfiles here).
Based on comment #23 From Matthew Marlowe at 2004-05-23 21:27 PST, it seems that he was able to install and operate somewhat normally on a PowerEdge 750. I have not tested the 2004.2 test cd's, but the 2004.1 CD works correctly for a SATA only version of the 750. However, using the kernels from emerge gentoo-sources and emerge gentoo-dev-sources (2.4.26-r5, and 2.6.7-r11), I do not get a system that recognizes the SATA drives (or more specifically even the IDE controller - the 6300ESB). the error message i get on a genkernel version of the above mentioned kernels is (ide1: resource busy). Same goes if set up the config as per seemingly appropriate from make menuconfig. Is there any specific options you have set on the live cd that have changed in regards to the chipset? Setting CONFIG_SCSI_SATA does not result in sd* drives either. Is this related to the bugs being discussed here? Or should I create a new bug report?
Honestly, you need to try a test CD first before going any further. You can get the kernel configuration from the test CD by running "zcat /proc/config.gz" from the CD to see exactly what we did. If you're having a problem with release media, it has no business here. This bug is entirely for the -test series.
Chris: Well, I was under the assumption that the genkernel'd gentoo kernel from an updated portage tree would be the same as the kernel that is on the new test cds (perhaps not bleeding edge but the config would generally be the same). i filed a bug about the 2.6 kernel, its bug id is 57327. again, this may not be the right place to report that, but it may be relevant for the other problems you guys are seeing. the caveat to the poweredge systems at least.
Yeah... because... it works for me... and stuff Guys... nobody use this stuff... it's all wrong... or something
Moving these so we can remove the "Install CD" component from "Gentoo Linux". I apologize to everyone for this spam, but according to the bugzilla developers, this is the only reasonable way to do this.
