Love-sources 2.6.0-test9 - Late but great

Message

wdreinhart · Post by **wdreinhart** » Tue Nov 11, 2003 2:03 am

bssteph wrote: I haven't had any problems like that... are you using the latest ~x86 nvidia drivers? NvAGP or AGPGART?
Latest ~x86 and AGPGART (in the kernel) here, and I've no problems.

Using AGPGART as a module, and the latest ~x86 nvidia drivers (1.0.4496-r4).

Dinini · Post by **Dinini** » Tue Nov 11, 2003 12:19 pm

Well, I have tested ftp://ftp.suse.com/pub/people/stepan/bo ... test9.diff on 3 different machines using kernels 2.6.0-test9, 2.6.0-test9-mm1, 2.6.0-test9-mm2 with over a dozen configuration files and:

It only works when I use the LILO bootloader!

If I use grub then I get to see tux and then see the "no good signature" found message.

Sample config for lilo and grub shown in http://forums.gentoo.org/viewtopic.php?p=641255 which were as close as I got to having them match and tossing out anything that didn't affect it. 4 days on this is enough, waiting for the next patch before trying again.

sindre · Post by **sindre** » Tue Nov 11, 2003 1:54 pm

With the same config bootsplash works on 2.6.0-test5-love3, but not on 2.6.0-test9-love2. I'm using grub. The bootsplash stuff does eventually appear as the bootsplash init script is run on test9-love2.

Dinini · Post by **Dinini** » Tue Nov 11, 2003 1:59 pm

Just to clarify, the *initial* bootsplash image (verbose or silent) only works for me when using lilo, the background image from the /etc/runlevel/default script will appear on all the test9 flavors I've tried using both grub and lilo.

ed0n · Post by **ed0n** » Tue Nov 11, 2003 2:41 pm

Dinini wrote:Well, I have tested ftp://ftp.suse.com/pub/people/stepan/bo ... test9.diff on 3 different machines using kernels 2.6.0-test9, 2.6.0-test9-mm1, 2.6.0-test9-mm2 with over a dozen configuration files and:

It only works when I use the LILO bootloader!

If I use grub then I get to see tux and then see the "no good signature" found message.

Sample config for lilo and grub shown in http://forums.gentoo.org/viewtopic.php?p=641255 which were as close as I got to having them match and tossing out anything that didn't affect it. 4 days on this is enough, waiting for the next patch before trying again.

I'am trying this too looks neat.

bssteph · Post by **bssteph** » Tue Nov 11, 2003 3:06 pm

ed0n wrote:
Dinini wrote:Well, I have tested ftp://ftp.suse.com/pub/people/stepan/bo ... test9.diff on 3 different machines using kernels 2.6.0-test9, 2.6.0-test9-mm1, 2.6.0-test9-mm2 with over a dozen configuration files and:

It only works when I use the LILO bootloader!

If I use grub then I get to see tux and then see the "no good signature" found message.

Sample config for lilo and grub shown in http://forums.gentoo.org/viewtopic.php?p=641255 which were as close as I got to having them match and tossing out anything that didn't affect it. 4 days on this is enough, waiting for the next patch before trying again. :)
I'am trying this too looks neat.

HMM...
Maybe this is the root of the problem then. I have maybe a half hour of time to try stuff today, and I'll see what LILO does with it.

watersb · Post by **watersb** » Tue Nov 11, 2003 9:34 pm

I think you need root=/dev/ram0 to boot from an init ramdisk on 2.6 kernels.

I posted an example in the Bootsplash support thread:
http://forums.gentoo.org/viewtopic.php?p=642338#642338

Random Task · Post by **Random Task** » Tue Nov 11, 2003 11:20 pm

hey guys,
seeing as I've never patched my kernel before... How do you do it?

I downloaded the patch but I'm not sure how to do it!

Please advise!

nephros · Post by **nephros** » Tue Nov 11, 2003 11:27 pm

It's in this thread, a few posts up.
Uh, and did you notice there is an ebuild, linked to from the first post?

Random Task · Post by **Random Task** » Tue Nov 11, 2003 11:50 pm

Oops, I breezed through the post, searched the forums but I guess i didn't pay enough attention

and I did try to do the ebuild, however figured the patch would be just as easy to install.

My mistake

Gibberx · Post by **Gibberx** » Wed Nov 12, 2003 12:38 am

Are we going to be putting Supermount-ng back in the patchset? Or, will it patch against the current love-sources? I guess I'll give it a shot soon.

Peaceable Frood · Post by **Peaceable Frood** » Wed Nov 12, 2003 12:42 am

Supermount doesn't patch against mm-sources. I've tried.

Gibberx · Post by **Gibberx** » Wed Nov 12, 2003 12:47 am

Oh yeah, and love-sources are based off of mm-sources... bummer. No big deal though, I'm no stranger to the "mount" command. Used it a lot in my Debian days...

Random Task · Post by **Random Task** » Wed Nov 12, 2003 1:42 am

I must say this patchset works like a charm! Thank you guys I have my bootsplash back!!

Lovechild · Post by **Lovechild** » Wed Nov 12, 2003 9:36 am

Phantom wrote:Are we going to be putting Supermount-ng back in the patchset? Or, will it patch against the current love-sources? I guess I'll give it a shot soon.

It's an evil idea anyways, I found this "emerge magicdev" it's less evil but still evil enough for kernel devs to hate it, at least it's userspace based.

asimon · Post by **asimon** » Wed Nov 12, 2003 9:53 am

watersb wrote:I think you need root=/dev/ram0 to boot from an init ramdisk on 2.6 kernels.

I posted an example in the Bootsplash support thread:
http://forums.gentoo.org/viewtopic.php?p=642338#642338

No, the root kernel parameter should contain your true root filesystem, not /dev/ram0. The initrd code uses /dev/ram which is hardcoded in init/do_mounts_initrd.c. There's no need to specify it in a kernel parameter. To use an initrd you have to tell the boot loader about it, that's all. Here an example for GRUB:

Code: Select all

title=Gentoo GNU/Linux 2.6.0-test9-mm1
        root (hd1,0)
        kernel /bzImage-2.6.0-test9-mm1 root=/dev/main_vg/root_lv hdb=ide-scsi hdd=ide-cd
        initrd /initrd-lvm-2.00.07-r1.gz

Notice that initrd support was broken (only if you used devfs) for several 2.6-test and mm-source releases but is fixed in the current release.

sindre · Post by **sindre** » Wed Nov 12, 2003 5:56 pm

Earlier I noticed that the hdparm results of recent mm-sources and love-sources have been quite a bit lower than on vanilla development-sources. However I ran a bonnie++ test on 2.6.0-test5-love3 (hdparm about 47MB/sec), and 2.6.0-test9-love2 (hdparm about 34MB/sec). The results didn't differ much so I think it's just the hdparm report being wrong, unless of course I interpreted the results of bonnie++ wrongly.

2.6.0-test5-love3: bonnie++ -f

Code: Select all

Version  1.03       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
sindre        1512M           61793  19 13149   3           30575   3 152.0   0
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 24742  93 +++++ +++ 26034 100 25579 100 +++++ +++ 24022 100
sindre,1512M,,,61793,19,13149,3,,,30575,3,152.0,0,16,24742,93,+++++,+++,26034,100,25579,100,+++++,+++,24022,100

2.6.0-test9-love2: bonnie++ -f

Code: Select all

Version  1.03       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
sindre        1512M           65708  20 12123   3           30520   3 202.9   0
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 26685  93 +++++ +++ 25941 100 27809  99 +++++ +++ 23443 100
sindre,1512M,,,65708,20,12123,3,,,30520,3,202.9,0,16,26685,93,+++++,+++,25941,100,27809,99,+++++,+++,23443,100

Sgaduuw · Post by **Sgaduuw** » Wed Nov 12, 2003 8:28 pm

just writing in to confirm that my computer with test9-love 2 with ACPI enabled will shutdown properly after I've unpatched the kexec patch with

Code: Select all

patch -p1 -R < /path/to/kexec.path

the kexec patch in question can be downloaded from here

bssteph · Post by **bssteph** » Wed Nov 12, 2003 9:37 pm

Sorry for the utter lack of communication, it seemed like everyone was answering everyone else well enough that I could clam up. ;)
It's tenth week here, which means it's the week before finals, which means I've been very busy doing all the things I should have done the past nine weeks. I dunno when I'll have time to screw around with the patches again (haven't even touched LILO yet), but yeah...
I'm less busy during exams week than this week, doesn't that rule? ;)

sindre: I want to say that i read somewhere (maybe LKML) that hdparm's been acting strangely on 2.6, but I can't remember where I saw that ATM...

norvillebarnes · Post by **norvillebarnes** » Wed Nov 12, 2003 10:40 pm

I just thought I chime in and say that lilo may indeed be the answer to the bootsplash problems. I've had no problems at all with bootsplash but I use lilo (never tried grub to be honest). With the exception of a couple of the mm patches I've had (vesa) framebuffer working in every 2.6 kernel by simply having vga=791 in my lilo global options. From there it was just a matter of adding initrd=/boot/initrd-1024x768 to lilo.conf, and presto it works great. I wish I could be of more help because it's awesome having bootsplash working with 2.6, but I have no experience with grub.

A problem that I have noticed is that with test9-mm2 or love2 (not sure about mm/love1) the kernel fails to load my agpgart module, leaving me without dri. I have an Intel 845G chipset that uses i830. Agpgart and drm are compiled into my kernel. Vanilla test 9 loads without any errors. Has anyone else with onboard video noticed this?

GoinBald · Post by **GoinBald** » Wed Nov 12, 2003 11:49 pm

Hello Again,
Thank You bssteph, Lovechild and all!
I just got back from vacation and they are piling it on again

soo short n' sweet.
Was one of those seat sales in the middle of the week.
Love is Grand! The best yet.
Anyone tried the cfq prio v2 patch yet?

For removable media, I use http://submount.sf.net
Works great!

Thanks again,
Loyal lover,
Baldy

Random Task · Post by **Random Task** » Thu Nov 13, 2003 2:28 am

I know vesa supports up to 1280x1024, but is there a way to get higher resolutions with the framebuffer, or any other framebuffer for that matter?

My notebook has a default resolution of 1600x1200, yeah its crazy but I love it for the desktop realestate!

Is it possible to have a framebuffer do 1600x1200 or am I just being too picky?

Thanks people!

paulisdead · Post by **paulisdead** » Thu Nov 13, 2003 6:09 am

This may have cropped up since 2.6-test7, since I didn't use test 8 much, and haven't tried burning a DVD+R in vanilla test 9, and also haven't tried the mm-sources for test9, but my DVD+R will only burn at around 1.4-1.5x. Unfortunately, I used up my last DVD+R, and probably won't be able to get more till pay day next week to do more testing with DVD+R burning with the different kernels.

PrakashP · Post by **PrakashP** » Thu Nov 13, 2003 8:27 am

So, impatiently waiting for love3, as mm3 has been released.

antisthenes · Post by **antisthenes** » Thu Nov 13, 2003 8:31 am

-test9-mm3

Looks like more IO fixes from Andrew, but no mention of the problems some people were having with DAO cd burning in -test9-mm2.

Looks easy enogh to -love up.