2.6.13.2-nitro1 "Down with latency!'

Message

Cinder6 · Post by **Cinder6** » Wed Sep 28, 2005 5:14 am

Ah, okay, that makes more sense.

Phlogiston · Post by **Phlogiston** » Wed Sep 28, 2005 5:53 am

hey guys. Do you can use USB 2.0 at full speed.

Here, my card reader is too slow. Somebody mentioned this also in a previous version of nitro.

Who can help me?

dmesg:

Code: Select all

ehci_hcd 0000:00:1d.7: GetStatus port 7 status 001002 POWER sig=se0 CSC
hub 1-0:1.0: port 7, status 0100, change 0001, 12 Mb/s
usb 1-7: USB disconnect, address 2
usb 1-7: usb_disable_device nuking all URBs
usb 1-7: unregistering interface 1-7:1.0
usb 1-7:1.0: hotplug
usb 1-7: unregistering device
usb 1-7: hotplug
hub 1-0:1.0: debounce: port 7: total 100ms stable 100ms status 0x100
hub 1-0:1.0: state 5 ports 8 chg 0000 evt 0040
ehci_hcd 0000:00:1d.7: GetStatus port 6 status 001803 POWER sig=j CSC CONNECT
hub 1-0:1.0: port 6, status 0501, change 0001, 480 Mb/s
hub 1-0:1.0: debounce: port 6: total 100ms stable 100ms status 0x501
ehci_hcd 0000:00:1d.7: port 6 high speed
ehci_hcd 0000:00:1d.7: GetStatus port 6 status 001005 POWER sig=se0 PE CONNECT
usb 1-6: new high speed USB device using ehci_hcd and address 3
ehci_hcd 0000:00:1d.7: port 6 high speed
ehci_hcd 0000:00:1d.7: GetStatus port 6 status 001005 POWER sig=se0 PE CONNECT
usb 1-6: default language 0x0409
usb 1-6: new device strings: Mfr=0, Product=3, SerialNumber=4
usb 1-6: Product: USB2.0 8-in-1 Card Reader
usb 1-6: SerialNumber: 73794
usb 1-6: hotplug
usb 1-6: adding 1-6:1.0 (config #1, interface 0)
usb 1-6:1.0: hotplug
Initializing USB Mass Storage driver...
usb-storage 1-6:1.0: usb_probe_interface
usb-storage 1-6:1.0: usb_probe_interface - got id
scsi2 : SCSI emulation for USB Mass Storage devices
usb-storage: device found at 3
usb-storage: waiting for device to settle before scanning
usbcore: registered new driver usb-storage
USB Mass Storage support registered.
  Vendor: Generic   Model: STORAGE DEVICE    Rev: 0128
  Type:   Direct-Access                      ANSI SCSI revision: 00
SCSI device sdb: 2041200 512-byte hdwr sectors (1045 MB)
sdb: Write Protect is off
sdb: Mode Sense: 02 00 00 00
sdb: assuming drive cache: write through
SCSI device sdb: 2041200 512-byte hdwr sectors (1045 MB)
sdb: Write Protect is off
sdb: Mode Sense: 02 00 00 00
sdb: assuming drive cache: write through
 sdb: sdb1
Attached scsi removable disk sdb at scsi2, channel 0, id 0, lun 0
Attached scsi generic sg1 at scsi2, channel 0, id 0, lun 0,  type 0
usb-storage: device scan complete

ahorn · Post by **ahorn** » Wed Sep 28, 2005 8:54 am

Code: Select all

Sep 28 03:23:47 aquapeter ------------[ cut here ]------------
Sep 28 03:23:47 aquapeter kernel BUG at mm/vmscan.c:407!
Sep 28 03:23:47 aquapeter invalid operand: 0000 [#1]
Sep 28 03:23:47 aquapeter Modules linked in: sch_ingress cls_u32 sch_sfq sch_cbq w83627hf i2c_sensor i2c_isa i2c_nforce2 snd_
pcm_oss snd_mixer_oss snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_intel8x0 snd_ac97_codec snd_pcm snd_timer snd
 soundcore snd_page_alloc lirc_atiusb lirc_dev pcspkr
Sep 28 03:23:47 aquapeter CPU:    0
Sep 28 03:23:47 aquapeter EIP:    0060:[<c0145a74>]    Tainted: GF     VLI
Sep 28 03:23:47 aquapeter EFLAGS: 00010202   (2.6.13.2-nitro1)
Sep 28 03:23:47 aquapeter EIP is at shrink_list+0x3e4/0x420
Sep 28 03:23:47 aquapeter eax: 4000084d   ebx: ddf21e78   ecx: c10a4078   edx: c12fe238
Sep 28 03:23:47 aquapeter esi: c10a4060   edi: dfc6dea0   ebp: dfc6df48   esp: dfc6de04
Sep 28 03:23:47 aquapeter ds: 007b   es: 007b   ss: 0068
Sep 28 03:23:47 aquapeter Process kswapd0 (pid: 189, threadinfo=dfc6c000 task=dfc83080)
Sep 28 03:23:47 aquapeter Stack: dfc6de2c 000000d0 00000000 00000001 00000001 ddf21e78 0000001c 00000000
Sep 28 03:23:47 aquapeter dfc6de24 dfc6de24 00000000 00000001 c10c9e20 c127b660 c109ab40 c12c6400
Sep 28 03:23:47 aquapeter c11d11c0 c10ff560 c10db520 c106aac0 c10b8140 c105f4a0 c13f0d40 c1084f80
Sep 28 03:23:47 aquapeter Call Trace:
Sep 28 03:23:47 aquapeter [<c0145c82>] shrink_cache+0x112/0x2a0
Sep 28 03:23:47 aquapeter [<c01462be>] shrink_zone+0xae/0xd0
Sep 28 03:23:47 aquapeter [<c014679c>] balance_pgdat+0x26c/0x430
Sep 28 03:23:47 aquapeter [<c0146a5f>] kswapd+0xff/0x150
Sep 28 03:23:47 aquapeter [<c012f710>] autoremove_wake_function+0x0/0x60
Sep 28 03:23:47 aquapeter [<c012f710>] autoremove_wake_function+0x0/0x60
Sep 28 03:23:47 aquapeter [<c0146960>] kswapd+0x0/0x150
Sep 28 03:23:47 aquapeter [<c0101375>] kernel_thread_helper+0x5/0x10
Sep 28 03:23:47 aquapeter Code: 30 39 42 30 8d 76 00 0f 84 50 fd ff ff e9 46 fd ff ff 8b 5e 0c 89 34 24 e8 da ac 00 00 fb 89
1c 24 e8 d1 b5 00 00 e9 f7 fe ff ff <0f> 0b 97 01 32 71 3f c0 e9 98 fc ff ff bb e0 df 44 c0 e9 4e fd
Sep 28 03:23:49 aquapeter ------------[ cut here ]------------
and so on...

is this a known problem?

Robin79 · Post by **Robin79** » Wed Sep 28, 2005 9:52 am

Robin79 wrote:hmm i tried to compile and gets this anyone?

Code: Select all

CC      init/main.o
init/main.c: I funktion `setup_per_cpu_areas':
init/main.c:329: error: `__per_cpu_end' undeclared (first use in this function)
init/main.c:329: error: (Each undeclared identifier is reported only once
init/main.c:329: error: for each function it appears in.)
init/main.c:329: error: `__per_cpu_start' undeclared (first use in this function)
make[1]: *** [init/main.o] Fel 1
make: *** [init] Fel 2

noone who can help me?

Guinpen · Post by **Guinpen** » Wed Sep 28, 2005 11:50 am

Sorry, I can't help you; I have two USB 2.0 devices (a soundcard and a hard drive) and both work fast simultaneously.

Even with swap prefetch disabled, my computer is uber slow with both the anticipatory (non-genetic) and CFQ shedulers. The moment I "swapoff -a", it flies. WTF? Next I'll try 2.6.13.1-nitro1 and see what happens.

Legoguy · Post by **Legoguy** » Wed Sep 28, 2005 10:38 pm

Godji wrote:Sorry, I can't help you; I have two USB 2.0 devices (a soundcard and a hard drive) and both work fast simultaneously.

Even with swap prefetch disabled, my computer is uber slow with both the anticipatory (non-genetic) and CFQ shedulers. The moment I "swapoff -a", it flies. WTF? Next I'll try 2.6.13.1-nitro1 and see what happens.

Confirmed... even swap prefetch disabled doesn't completely resolve it, in fact at times it almost seems worse...

Robin79 · Post by **Robin79** » Wed Sep 28, 2005 11:05 pm

Okey so i understand that the issue has something to do with smt and hp when i disable that it works when i enable it it dont work..

So Seppe please?

Guinpen · Post by **Guinpen** » Wed Sep 28, 2005 11:13 pm

Legoguy wrote:
Godji wrote:Sorry, I can't help you; I have two USB 2.0 devices (a soundcard and a hard drive) and both work fast simultaneously.

Even with swap prefetch disabled, my computer is uber slow with both the anticipatory (non-genetic) and CFQ shedulers. The moment I "swapoff -a", it flies. WTF? Next I'll try 2.6.13.1-nitro1 and see what happens.
Confirmed... even swap prefetch disabled doesn't completely resolve it, in fact at times it almost seems worse...

I tried 2.6.13.1-nitro1. Much better, though not quite as good as 2.6.12-morph7 on my machine.
So, what are the differences between the two releases of nitro?

Robin79 · Post by **Robin79** » Wed Sep 28, 2005 11:21 pm

2.6.13.1-nitro1 "The Legend Continues" [STATUS: TESTING]

that one compiles good for me so i stick with that one just w8ing so i can reboot

bollucks · Post by **bollucks** » Thu Sep 29, 2005 1:13 am

Legoguy wrote:
Godji wrote:Sorry, I can't help you; I have two USB 2.0 devices (a soundcard and a hard drive) and both work fast simultaneously.

Even with swap prefetch disabled, my computer is uber slow with both the anticipatory (non-genetic) and CFQ shedulers. The moment I "swapoff -a", it flies. WTF? Next I'll try 2.6.13.1-nitro1 and see what happens.
Confirmed... even swap prefetch disabled doesn't completely resolve it, in fact at times it almost seems worse...

You're probably seeing the bug with the write throttling patch which was removed from ck6 but is still on this aging

kernel.

Fleta · Post by **Fleta** » Thu Sep 29, 2005 6:48 am

Now this is weird, my openoffice-ximian starts to segment fault with this kernel. It doesn't "segment fault completely" in the sense that once I restart X, everything will be fine, but this is really annoying. Anybody has experienced same problem?

bollucks · Post by **bollucks** » Thu Sep 29, 2005 7:26 am

Fleta wrote:Now this is weird, my openoffice-ximian starts to segment fault with this kernel. It doesn't "segment fault completely" in the sense that once I restart X, everything will be fine, but this is really annoying. Anybody has experienced same problem?

You see that with DRI permission problems because oo.o uses DRI.

Fleta · Post by **Fleta** » Thu Sep 29, 2005 7:34 am

bollucks wrote:
Fleta wrote:Now this is weird, my openoffice-ximian starts to segment fault with this kernel. It doesn't "segment fault completely" in the sense that once I restart X, everything will be fine, but this is really annoying. Anybody has experienced same problem?
You see that with DRI permission problems because oo.o uses DRI.

I'm not sure what you mean, the only error message I see (when start it from a terminal-emulator) is just "segmentation fault".

seppe · Post by **seppe** » Thu Sep 29, 2005 10:57 am

Cinder6 wrote: nitro2, seppe?

Not yet, sorry. I'm pretty busy these days for my work. But there is a very little chance that I make a 2.6.13.2-nitro2 in my oh so precious weekend. But next week I certainly don't have the time for it. I'll see what I can do this weekend, but I don't promise anything.

bollucks · Post by **bollucks** » Thu Sep 29, 2005 11:03 am

Fleta wrote:
bollucks wrote:
Fleta wrote:Now this is weird, my openoffice-ximian starts to segment fault with this kernel. It doesn't "segment fault completely" in the sense that once I restart X, everything will be fine, but this is really annoying. Anybody has experienced same problem?
You see that with DRI permission problems because oo.o uses DRI.
I'm not sure what you mean, the only error message I see (when start it from a terminal-emulator) is just "segmentation fault".

Yes that's right you get a meaningless error that occurs as a result of DRI permissions.

bollucks · Post by **bollucks** » Thu Sep 29, 2005 11:04 am

Why not try the ck5 to ck6 patch from the ck web site? It probably applies cleanly to nitro apart from the makefile which is a harmless reject.

iphitus · Post by **iphitus** » Fri Sep 30, 2005 1:07 am

For those who want to keep right up to date, I released archck6 last week, based on ck6. I have just updated it to the latest prefetch which should fix some of the ooms that people are having. ck6 also saw the withdrawl of an experimental patch, which saw some hangs and pauses.

Ebuilds, here but I am not sure if they are up to date. http://bugs.gentoo.org/show_bug.cgi?id=103354

edit: sorry to those who got a bootsplash patch, I accidentally uploaded the wrong one

fbsplash is up now.

Soul_rebel · Post by **Soul_rebel** » Fri Sep 30, 2005 3:10 pm

Code: Select all

mobile64unix linux-2.6.13.2-nitro1 # make
  CHK     include/linux/version.h
  SPLIT   include/linux/autoconf.h -> include/config/*
make[1]: `arch/x86_64/kernel/asm-offsets.s' is up to date.
  CHK     include/linux/compile.h
  CHK     usr/initramfs_list
  GZIP    kernel/config_data.gz
  IKCFG   kernel/config_data.h
  CC      kernel/configs.o
  LD      kernel/built-in.o
  GEN     .version
  CHK     include/linux/compile.h
  UPD     include/linux/compile.h
  CC      init/version.o
  LD      init/built-in.o
  LD      .tmp_vmlinux1
kernel/built-in.o(.text+0x2b864): In function `do_suspend2_lowlevel':
: undefined reference to `suspend_cpu_counter'
drivers/built-in.o(.data+0x56518): undefined reference to `cfb_fillrect'
drivers/built-in.o(.data+0x56520): undefined reference to `cfb_copyarea'
drivers/built-in.o(.data+0x56528): undefined reference to `cfb_imageblit'
drivers/built-in.o(.data+0x56530): undefined reference to `soft_cursor'
make: *** [.tmp_vmlinux1] Error 1

on amd64, with vesafb (vesafb-tng seems not available on amd64), i can't compile it. On a 32bit pc with vesafb-tng it compiles.

den_RDC · Post by **den_RDC** » Fri Sep 30, 2005 5:32 pm

works fine on amd64 , did apply the amd64 x2 patch to fix the "lost ticks" issue ...
and, to my surprise, suddenly alsa seems to have gained hw mixing on snd_intel8x0 !

iphitus · Post by **iphitus** » Sat Oct 01, 2005 12:01 am

Soul_rebel wrote:
Code: Select all
<compile error>
on amd64, with vesafb (vesafb-tng seems not available on amd64), i can't compile it. On a 32bit pc with vesafb-tng it compiles.

This problem has been seen on amd64 on both nitro, archck and vanilla. It isnt patchset related.

I have sent an email to nigel, the developer of software suspend and we will hopefully see a fix soon.

iphitus

MetX · Post by **MetX** » Sat Oct 01, 2005 12:23 am

den_RDC wrote:works fine on amd64 , did apply the amd64 x2 patch to fix the "lost ticks" issue ...
and, to my surprise, suddenly alsa seems to have gained hw mixing on snd_intel8x0 !

How do you know if an alsa driver has HW mixing?

iphitus · Post by **iphitus** » Sat Oct 01, 2005 12:31 am

MetX wrote:
den_RDC wrote:works fine on amd64 , did apply the amd64 x2 patch to fix the "lost ticks" issue ...
and, to my surprise, suddenly alsa seems to have gained hw mixing on snd_intel8x0 !
How do you know if an alsa driver has HW mixing?

You can use multiple programs at once without an icky sound daemon such as artS or esd.

Chances are, he is just using the newer alsa software dmix. I am pretty sure that the intel integrated audio are not capable of hw mixing.

iphitus

iphitus · Post by **iphitus** » Sat Oct 01, 2005 1:17 am

Uploaded a patch for the software suspend compile problems on AMD64 thanks to Nigel Cunningham, the developer of suspend2

http://iphitus.loudas.com/arch/ck/patch ... pend.patch

den_RDC · Post by **den_RDC** » Sat Oct 01, 2005 7:31 pm

iphitus wrote: How do you know if an alsa driver has HW mixing?

You can use multiple programs at once without an icky sound daemon such as artS or esd.

Chances are, he is just using the newer alsa software dmix. I am pretty sure that the intel integrated audio are not capable of hw mixing.

iphitus[/quote]
True, but i haven't set up dmix in my .asoundrc . So either alsa has "grown" a new functionality that enables it to activate dmix transparantly, or the intel ac97 driver knows how to use the hw mixing capabilities of my mainboard (which is an ac97 thing on an asus a8n-sli deluxe, lspci reports nVidia Corporation CK804 AC'97 Audio Controller , and it's using the Realtek ALC850 chip ). I am really intrested how it works...

iphitus · Post by **iphitus** » Sat Oct 01, 2005 11:36 pm

Latest alsas enable dmix by default, without an asound config.