Love-sources 2.6.0-test9 - Late but great

[michaelarch]

@PrakashKC

I'm a big time noob at patching and very confused as to how to apply your patch. Please help. Thanks for everything! Have a great day!

Michael

[ronmon]

Hey, bssteph,

Nice job on the patches. Here are two things that might be hanging you up on your bootsplash problem.

First, as GoinBald pointed out, I used splash to create the initrd rather than copying it as you did. I used only one '>' for output, though, since I want it to overwrite a possibly preexisting file instead of appending to it:

Code: Select all

/sbin/splash -s -f /etc/bootsplash/default/config/bootsplash-1024x768.cfg > /boot/initrd-1024x768

Also, it looks like there is an '=' equal sign missing in the initrd line in your grub.conf. I use lilo, but according to the How-To it should be the same as mine:

Code: Select all

initrd = /boot/initrd-1024x768

I've been trying to make a custom bootsplash without success so far. The stock setups in /etc/bootsplash/gentoo work fine. I'm using hex values for my vga= settings.

[PrakashP]

@michaelarch

Change to you kernel sources dir (/usr/src/linux hope you symlinked current kernel to it...)

save the patch as file

do a

patch -p1 <"patchfile"

where "patchfile" is the path to your saved file.

[bssteph]

Maybe there would be some sense in starting to push Love-sources patches to Gentoo-dev-sources?

I leave that to your discretion. ;)
You probably have more sway with anyone at Gentoo than I do. I'd thought of that too, in a random musing, but working out the details is up to you.

URLs
test9-mm2
Con's VM swappiness
AMD PM
kexec
bootsplash
Radeon IGP DRM
Orinoco functionality
Nick's sched
exec-shield
vesa refresh rate
Nick's elevator selection
Athlon MCE fix
bonus: this list could also be useful if anyone on gentoo-dev-sources cares!

PrakashKC, based on your older posts, I would really start with removing kexec... rolling that back should go smoothly.

[bssteph]

Hey, bssteph,

Nice job on the patches. Here are two things that might be hanging you up on your bootsplash problem.

First, as GoinBald pointed out, I used splash to create the initrd rather than copying it as you did. I used only one '>' for output, though, since I want it to overwrite a possibly preexisting file instead of appending to it:

Code: Select all

/sbin/splash -s -f /etc/bootsplash/default/config/bootsplash-1024x768.cfg > /boot/initrd-1024x768

Also, it looks like there is an '=' equal sign missing in the initrd line in your grub.conf. I use lilo, but according to the How-To it should be the same as mine:

Code: Select all

initrd = /boot/initrd-1024x768

I've been trying to make a custom bootsplash without success so far. The stock setups in /etc/bootsplash/gentoo work fine. I'm using hex values for my vga= settings.

I've tried to make my own initrd with splash with the same results.
And I also tried both with and without the = ... both seem to be legal for grub (and the kernel _IS_ seeing the initrd, just bootsplash doesn't like the images for some reason)
Stefan emailed me a couple days ago about the release after I expressed interest, and I replied last night with lots of praise, and snippets from my problem. So we'll see how that goes.

[michaelarch]

@PrakashKC

Thanks, i'm doing it now. Have a good one!

Michael

[mojo]

after changing this line :

Code: Select all

/sbin/splash -s -f /etc/bootsplash/default/config/bootsplash-1280x1024.cfg > /boot/initrd-1280x1024

to

Code: Select all

/sbin/splash -s -f /etc/bootsplash/default/config/bootsplash-1280x1024.cfg >> /boot/initrd-1280x1024

the bootsplash error changes to

Code: Select all

bootsplash 3.1.1-2003/11/02: looking for picture.....<6>bootsplash: error while decompressing picture: no EOI (13)      

anybody an idea?

[PrakashP]

@bssteph

Good guess, throwing out kexec machine shuts down again. Gotta make a bug report...

[bssteph]

Another set of interesting patches:
http://www.ussg.iu.edu/hypermail/linux/ ... /0019.html

This adds tunability to cfq.

I've implemented IO nice levels in the CFQ io scheduler. It works as
follows.

A process has an assigned io nice level, anywhere from 0 to 20. Both of
these end values are "special" - 0 means the process is only allowed to
do io if the disk is idle, and 20 means the process io is considered
realtime.

This is half a reminder to myself and half an interest builder.


http://www.ussg.iu.edu/hypermail/linux/ ... /0029.html
Minor scheduler fix from Con - didn't check if this undone by Nick's sched.

[Lovechild]

IO niceness - sounds usefull

[IvoryTiger]

Well I think the re-image of my box was good for me, learned a lot of stuff I'd glossed over before. I now have the 2.4.20-gentoo-r8 kernel and the 2.6.0-test9-love2 working fine on both console and X. I only have to switch my symlink in /usr/src and re-emerge the nvidia-kernel before I switch to the other kernel and it's happy.

As a bit of a bonus, the bootsplash works for both the 2.4.20 and 2.6.0 without any fussing.

This is my grub.conf for anyone that's interested.

Code: Select all

default 0
timeout 30
splashimage=(hd1,0)/grub/splash.xpm.gz

title=2.4.20-gentroo-r8
root (hd1,0)
kernel (hd1,0)/kernel-2.4.20-gentoo-r8 root=/dev/hdg6 idebus=33 video=vesa:ywrap,mtrr vga=791
initrd (hd1,0)/initrd-2.4.20-gentoo-r8

title=2.6.0-test9-love2
root (hd1,0)
kernel (hd1,0)/kernel-2.6.0-test9-love2 root=/dev/hdg6 idebus=33 video=vesa:ywrap,mtrr vga=791
initrd (hd1,0)/initrd-1024x768

My /etc/fstab

Code: Select all

/dev/hdg1          /boot          ext3          noauto,noatime          1 1
/dev/hdg6          /                 ext3          noatime                      0 0
/dev/hde2          none          swap           sw                               0 0
/dev/hdg7          /opt            ext3          noatime                      0 0
/dev/hdg8          /usr             ext3          noatime                      0 0
/dev/hdg9          /var             ext3          noatime                      0 0
/dev/hdg10        /home         ext3          noatime                      0 0
/dev/cdroms/cdrom0 /mnt/cdr iso9660 noauto,ro                     0 0
/dev/cdroms/cdrom1 /mnt/cdr iso9660 noauto,ro                     0 0

none                 /proc            proc         defaults                       0 0
none                 /dev/shm      tmpfs       defaults                       0 0

I have 2 cdrom's on the motherboard ide ports and my two hard drives on an pci ata100 card.

Code: Select all

Drive 1:
1st partition  - running Bootmagic 8.0
2nd partition - linux swap
3rd partition - WinXP

Drive 2:
1st partion - /boot
2nd partition - WinXP swap
3rd partion - Extended parition containing the rest of the linux partitions
etc.

[IvoryTiger]

I've tried to make my own initrd with splash with the same results.
And I also tried both with and without the = ... both seem to be legal for grub (and the kernel _IS_ seeing the initrd, just bootsplash doesn't like the images for some reason)
Stefan emailed me a couple days ago about the release after I expressed interest, and I replied last night with lots of praise, and snippets from my problem. So we'll see how that goes.

Odd question but have you tried lowering the number of color bits from 24/32 down to 16 or even 8 bits?

[bssteph]

I've tried to make my own initrd with splash with the same results.
And I also tried both with and without the = ... both seem to be legal for grub (and the kernel _IS_ seeing the initrd, just bootsplash doesn't like the images for some reason)
Stefan emailed me a couple days ago about the release after I expressed interest, and I replied last night with lots of praise, and snippets from my problem. So we'll see how that goes.

Odd question but have you tried lowering the number of color bits from 24/32 down to 16 or even 8 bits?

I have not, but I was under the assumption that both the gentoo images (those are the ones I've tried) and the prebuilt initrds use a lower depth. But I'll try later.

[Parak]

Hmm, fails to boot for me, hard locks on "Checking all filesystems..."

Used the same config as for test8-love1. Tried to remove checkfs by doing rc-update del checkfs boot, but it still starts for some reason. Not sure how to get the exact part that it hangs on, so if someone could help I'd be very grateful.

Thanks.

[coreutils]

It seems a lot of us get the bootsplash 3.1.1-2003/11/02: looking for picture... no good signature found. message.
I also get this. I can change the background in VTs with splash -s -u 0 /etc/bootsplash/<theme>/config/<config-file>, but I have no luck with getting a splash during boot. If someone knows what the problem is, please post it.

[Seymour]

I'm having a problem with my usb memory device. It works once, but after that, it will not register. Dmesg shows the following after inserting and removing the device after a fresh boot:

Code: Select all

hub 3-0:1.0: new USB device on port 2, assigned address 2
Initializing USB Mass Storage driver...
scsi1 : SCSI emulation for USB Mass Storage devices
  Vendor: LEXAR     Model: DIGITAL FILM      Rev: /W1.
  Type:   Direct-Access                      ANSI SCSI revision: 02
SCSI device sda: 125952 512-byte hdwr sectors (64 MB)
sda: Write Protect is off
sda: Mode Sense: 00 10 00 00
sda: assuming drive cache: write through
 /dev/scsi/host1/bus0/target0/lun0: p1
Attached scsi removable disk sda at scsi1, channel 0, id 0, lun 0
Attached scsi generic sg1 at scsi1, channel 0, id 0, lun 0,  type 0
WARNING: USB Mass Storage data integrity not assured
USB Mass Storage device found at 2
drivers/usb/core/usb.c: registered new driver usb-storage
USB Mass Storage support registered.
usb 3-2: USB disconnect, address 2
releasing anticipatory io scheduler
Unable to handle kernel NULL pointer dereference at virtual address 00000001
 printing eip:
00000001
*pde = 00000000
Oops: 0000 [#1]
PREEMPT 
CPU:    0
EIP:    0060:[<00000001>]    Tainted: PF  VLI
EFLAGS: 00010202
EIP is at 0x1
eax: eaedb200   ebx: e88ce400   ecx: 00000001   edx: e88ce400
esi: e88ce401   edi: e88ce410   ebp: eaede224   esp: ef1efe10
ds: 007b   es: 007b   ss: 0068
Process khubd (pid: 2437, threadinfo=ef1ee000 task=ef1f1940)
Stack: c0211f1a e88ce400 eaedb200 c02140d8 e88ce400 e8f88500 c0171991 effe7ba0 
       e8c01c00 c036cff0 c036d020 f18f8d5c e88ce400 e88ce2a0 c036cff0 f18f9deb 
       e8c01c00 00000000 c020eb30 e8c01d80 c01e4da3 e8fab0c0 e8c01da8 e8c01da8 
Call Trace:
 [<c0211f1a>] elevator_exit+0x3a/0x50
 [<c02140d8>] blk_cleanup_queue+0x98/0xf0
 [<c0171991>] destroy_inode+0x51/0x60
 [<f18f8d5c>] scsi_free_sdev+0x12c/0x170 [scsi_mod]
 [<f18f9deb>] scsi_device_dev_release+0x1b/0x30 [scsi_mod]
 [<c020eb30>] device_release+0x20/0x80
 [<c01e4da3>] kobject_del+0x43/0x80
 [<c01e4ee5>] kobject_cleanup+0x85/0xa0
 [<f18fa4a7>] scsi_remove_device+0x87/0xc0 [scsi_mod]
 [<f18f9a49>] scsi_forget_host+0x19/0x30 [scsi_mod]
 [<f18f3f5b>] scsi_remove_host+0x2b/0x80 [scsi_mod]
 [<f1d089d8>] storage_disconnect+0x38/0x48 [usb_storage]
 [<f193e138>] usb_unbind_interface+0x78/0x80 [usbcore]
 [<c020fe56>] device_release_driver+0x66/0x70
 [<c020ff85>] bus_remove_device+0x55/0xa0
 [<c020ef0d>] device_del+0x5d/0xa0
 [<f1944900>] usb_disable_device+0x70/0xb0 [usbcore]
 [<f193ec86>] usb_disconnect+0x96/0xf0 [usbcore]
 [<f19414bf>] hub_port_connect_change+0x32f/0x340 [usbcore]
 [<f1940d9a>] hub_port_status+0x3a/0xb0 [usbcore]
 [<f1941807>] hub_events+0x337/0x3a0 [usbcore]
 [<f194189d>] hub_thread+0x2d/0xf0 [usbcore]
 [<c02e9fc6>] ret_from_fork+0x6/0x14
 [<c011c520>] default_wake_function+0x0/0x30
 [<f1941870>] hub_thread+0x0/0xf0 [usbcore]
 [<c010ab99>] kernel_thread_helper+0x5/0xc

Code:  Bad EIP value.

Reinserting the device shows nothing more in dmesg. The machine must be rebooted for the device to work again.
I tried this with mm2 and it worked as it should have.

[Corp.Nobbs]

@bssteph

Good guess, throwing out kexec machine shuts down again. Gotta make a bug report...

Good stuff.. It fixed it for me too

Now if I could only get that bootsplash working

[sindre]

I have bootsplash working with a somewhat confused progressbar on test5-love3.

[bssteph]

Seymour: is that reproducable?

Update, the kernel gurus decided Con's fix wasn't a fix at all, so that can be ignored.

[Peaceable Frood]

Bootsplash works, but oddly, when it gets to the bootsplash script then it will work, however I can't get silent mode to work.

[Seymour]

Seymour: is that reproducable?

Yup, it happens every time on this machine. I don't have any other setup to test it with though.

[Lovechild]

Nick v18 and CFQ IO Prio #2 released

Nothing in nick for us sadly, its high end improvments.

[nephros]

... along with an update to the AMD power stuff.
http://www.ussg.iu.edu/hypermail/linux/ ... /0209.html
If you have such a system please run the script from that thread and report it to Tony.

[wdreinhart]

Is anyone else having problems with bootsplash and nvidia-kernel?

Just tried to upgrade from -test5-love4 to -test9-love2. If I have a bootsplash image set, the nvidia modules causes a kernel oops, and fails to load. If the nvidia module is loaded, and then I try to set a bootsplash image, the kernel panics and hangs the machine. (I have re-emerged nvidia-kernel with the new /usr/src/linux sylink, so it's not a version issue) Other than bootsplash, test9-love2 seems to be working great. Keep it up bssteph!

[bssteph]

Is anyone else having problems with bootsplash and nvidia-kernel?

Just tried to upgrade from -test5-love4 to -test9-love2. If I have a bootsplash image set, the nvidia modules causes a kernel oops, and fails to load. If the nvidia module is loaded, and then I try to set a bootsplash image, the kernel panics and hangs the machine. (I have re-emerged nvidia-kernel with the new /usr/src/linux sylink, so it's not a version issue) Other than bootsplash, test9-love2 seems to be working great. Keep it up bssteph!

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.