2.6.9-nitro4 "Less is more"

Message

hds · Post by **hds** » Fri Nov 26, 2004 11:35 pm

doesnt this exist in 2.6.x?

CONFIG_USB_SCANNER

John5788 · Post by **John5788** » Sat Nov 27, 2004 12:44 am

hds wrote:thx! had have to disable 4kstacks anyway for nvidia, if i recall correctly.
yes, riser4 option is there now, thx again.

@seppe.. maybe you could adjust some default settings of the kernel? maybe it saves others some headache as well.
otherwise: nice job sofar!
will report back if i get in troubles.

the new nvidia drivers work with 4kb stacks

cuchumino · Post by **cuchumino** » Sat Nov 27, 2004 6:34 am

which patch is better for desktop performance, cko or nitro in speed wise desktop performance?

John5788 · Post by **John5788** » Sat Nov 27, 2004 6:49 am

nitro is supposed to be faster

thecrazyperson_ws · Post by **thecrazyperson_ws** » Sat Nov 27, 2004 9:43 am

Thank you Seppe, for this wonderful nitro. It is the fastest, most responsive kernel that I have found, and I have been pushing the load average well above 10 just to see if it'll hang or break, and so far... nothing.

warren@Tidus ~ $ uptime
03:42:25 up 14:13, 10 users, load average: 6.29, 6.04, 5.34

I've had it that way for the entire 14 hours it's been up. I believe that you have a lifetime nitro user, Seppe.

Mmmm, nitro.

EDIT: Oh yes, I forgot to say that it's stable enough on this computer that it's the only kernel in my bootmanager. Absolutely zero problems with my setup.

jzono1 · Post by **jzono1** » Sat Nov 27, 2004 12:22 pm

Any way of disabling the staircase scheduler?
Optimally, i would like a way of disabling it runtime...
will pluggable cpu schedulers be in nitro soon, as staircase makes sound lag in warcraft, but increase doom3 performance?

cuchumino · Post by **cuchumino** » Sat Nov 27, 2004 4:42 pm

John5788 wrote:nitro is supposed to be faster

thankyou for answering that question

Deranger · Post by **Deranger** » Sat Nov 27, 2004 4:45 pm

Well, nitro-sources aren't faster than ck- or cko-sources. Nitro doesn't have any patches that really improve performance, and all those patchsets uses ck-sources as a base.

hds · Post by **hds** » Sat Nov 27, 2004 4:58 pm

thecrazyperson_ws wrote: I've had it that way for the entire 14 hours it's been up. I believe that you have a lifetime nitro user, Seppe.

if an uptime of 14 hours gets you excited, i really wonder what kernels you tried sofar

vrln · Post by **vrln** » Sat Nov 27, 2004 6:01 pm

John5788 wrote:nitro is supposed to be faster

Why? Nitro is almost the same as cko, and contains no additional speed related patches compared to cko. Both are good patchsets.

As Oktane already said, almost all desktop patchsets are based on Con Kolivas -ck patchset (as in they are just ck + some other things like fbsplash/reiser4 etc). That's where all the speed/interactivity patches come from. All praise for speed improvements should be directed there too.

monkey89 · Post by **monkey89** » Sun Nov 28, 2004 6:44 am

jzono1 wrote:Any way of disabling the staircase scheduler?
Optimally, i would like a way of disabling it runtime...
will pluggable cpu schedulers be in nitro soon, as staircase makes sound lag in warcraft, but increase doom3 performance?

Con's working on it for ck-sources; check his mailing list. I believe that he had a patch released for plugsched against 2.6.10.

http://bhhdoa.org.au/pipermail/ck/

nxsty · Post by **nxsty** » Sun Nov 28, 2004 10:55 am

The plugsched patches can be found here:
http://ck.kolivas.org/patches/plugsched/

But they are only against -mm ATM.

jzono1 · Post by **jzono1** » Sun Nov 28, 2004 11:52 am

gonna hack em, in a way so they apply against nitro4

staircase does wonders for doom3 timedemos, i saw a 50%increase with sc&latest nvidia drivers

plugsched should be included in the next nitro

kostian · Post by **kostian** » Sun Nov 28, 2004 6:10 pm

Hi, i cant sync my palm over usb with nitro4

Code: Select all

Unable to handle kernel NULL pointer dereference at virtual address 00000008
 printing eip:
*pde = 00000000
Oops: 0002 [#1]
PREEMPT
Modules linked in: visor usbserial snd_pcm_oss snd_mixer_oss uhci_hcd nvidia_agp ohci_hcd ehci_hcd forcedeth 
snd_intel8x0 snd_ac97_codec snd_pcm snd_timer snd_page_alloc gameport snd_mpu401_uart snd_rawmidi snd_seq_device snd 
soundcore nls_iso8859_1 nls_cp437 vfat fat 8250 serial_core floppy pcspkr nvidia evdev psmouse w83627hf eeprom 
i2c_sensor i2c_isa i2c_nforce2 i2c_core ipv6 8139too mii crc32 agpgart reiserfs
CPU:    0
EIP:    0060:[<d0df4037>]    Tainted: P   VLI
EFLAGS: 00010002   (2.6.9-nitro4)
EIP is at visor_open+0x37/0x1e0 [visor]
eax: 00000286   ebx: 00000000   ecx: cd5f7660   edx: cd236000
esi: ce441660   edi: c126d200   ebp: 00000000   esp: cd237e9c
ds: 007b   es: 007b   ss: 0068
Process pilot-xfer (pid: 3766, threadinfo=cd236000 task=cfb23aa0)
Stack: d0def61b 00000001 00000000 00000000 ce441660 00000001 ce441660 c126d200
       d0dec3ec c126d200 cf7846a0 00000000 cffc8400 cf7846a0 00000000 0bc00001
       c01b5da3 cd6db000 cf7846a0 cd237ef0 08028404 cd6db000 00000001 cf7846a0
Call Trace:
 [<d0dec3ec>] serial_open+0xbc/0x160 [usbserial]
 [<c01b5da3>] tty_open+0x233/0x2b0
 [<c01b5b70>] tty_open+0x0/0x2b0
 [<c0160834>] chrdev_open+0xe4/0x1d0
 [<c015660a>] dentry_open+0x14a/0x230
 [<c01564b2>] filp_open+0x62/0x70
 [<c0156729>] get_unused_fd+0x39/0xe0
 [<c0156899>] sys_open+0x49/0x90
 [<c010722b>] syscall_call+0x7/0xb
Code: 7c 24 24 85 c0 8b 37 8b 9f e8 00 00 00 0f 85 8a 01 00 00 8b 4f 1c 85 c9 0f 84 4a 01 00 00 9c 58 fa ba 00 e0 ff ff 
21 e2 ff 42 14 <c7> 43 08 00 00 00 00 c7 03 00 00 00 00 c7 43 04 00 00 00 00 50
 <6>note: pilot-xfer[3766] exited with preempt_count 1
bad: scheduling while atomic!
 [<c02ada5d>] schedule+0x41d/0x430
 [<c014747b>] unmap_page_range+0x4b/0x80
 [<c011bf07>] cond_resched_lock+0x37/0x60
 [<c01475fe>] unmap_vmas+0x14e/0x1b0
 [<c014bb43>] exit_mmap+0x83/0x160
ds: 007b   es: 007b   ss: 0068
Process pilot-xfer (pid: 3766, threadinfo=cd236000 task=cfb23aa0)
Stack: d0def61b 00000001 00000000 00000000 ce441660 00000001 ce441660 c126d200
       d0dec3ec c126d200 cf7846a0 00000000 cffc8400 cf7846a0 00000000 0bc00001
       c01b5da3 cd6db000 cf7846a0 cd237ef0 08028404 cd6db000 00000001 cf7846a0
Call Trace:
 [<d0dec3ec>] serial_open+0xbc/0x160 [usbserial]
 [<c01b5da3>] tty_open+0x233/0x2b0
 [<c01b5b70>] tty_open+0x0/0x2b0
 [<c0160834>] chrdev_open+0xe4/0x1d0
 [<c015660a>] dentry_open+0x14a/0x230
 [<c01564b2>] filp_open+0x62/0x70
 [<c0156729>] get_unused_fd+0x39/0xe0
 [<c0156899>] sys_open+0x49/0x90
 [<c010722b>] syscall_call+0x7/0xb
Code: 7c 24 24 85 c0 8b 37 8b 9f e8 00 00 00 0f 85 8a 01 00 00 8b 4f 1c 85 c9 0f 84 4a 01 00 00 9c 58 fa ba 00 e0 ff ff 
21 e2 ff 42 14 <c7> 43 08 00 00 00 00 c7 03 00 00 00 00 c7 43 04 00 00 00 00 50
 <6>note: pilot-xfer[3766] exited with preempt_count 1
bad: scheduling while atomic!
 [<c02ada5d>] schedule+0x41d/0x430
 [<c014747b>] unmap_page_range+0x4b/0x80
 [<c011bf07>] cond_resched_lock+0x37/0x60
 [<c01475fe>] unmap_vmas+0x14e/0x1b0
 [<c014bb43>] exit_mmap+0x83/0x160
 [<c011ca94>] mmput+0x64/0xb0
 [<c0120f06>] do_exit+0x156/0x430
 [<c01196b0>] do_page_fault+0x0/0x599
 [<c01084a8>] die+0x188/0x190
 [<c01196b0>] do_page_fault+0x0/0x599
 [<c011ef47>] printk+0x17/0x20
 [<c0119989>] do_page_fault+0x2d9/0x599
 [<c013dc55>] buffered_rmqueue+0xf5/0x1c0
 [<c013dc55>] buffered_rmqueue+0xf5/0x1c0
 [<c013df60>] __alloc_pages+0x240/0x400
 [<c01196b0>] do_page_fault+0x0/0x599
 [<c0107c35>] error_code+0x2d/0x38
 [<d0df4037>] visor_open+0x37/0x1e0 [visor]
 [<d0dec3ec>] serial_open+0xbc/0x160 [usbserial]
 [<c01b5da3>] tty_open+0x233/0x2b0
 [<c01b5b70>] tty_open+0x0/0x2b0
 [<c0160834>] chrdev_open+0xe4/0x1d0
 [<c015660a>] dentry_open+0x14a/0x230
 [<c01564b2>] filp_open+0x62/0x70
 [<c0156729>] get_unused_fd+0x39/0xe0
 [<c0156899>] sys_open+0x49/0x90
 [<c010722b>] syscall_call+0x7/0xb
Segmentation fault

Ive recompiled without PREEMT support but still no go:

Code: Select all

Unable to handle kernel NULL pointer dereference at virtual address 00000008
 printing eip:
*pde = 00000000
Oops: 0002 [#1]
Modules linked in: visor usbserial snd_pcm_oss snd_mixer_oss uhci_hcd nvidia_agp ohci_hcd ehci_hcd forcedeth 
snd_intel8x0 snd_ac97_codec snd_pcm snd_timer snd_page_alloc gameport snd_mpu401_uart snd_rawmidi snd_seq_device snd 
soundcore nls_iso8859_1 nls_cp437 vfat fat 8250 serial_core evdev floppy pcspkr psmouse w83627hf eeprom i2c_sensor 
i2c_isa i2c_nforce2 i2c_core ipv6 8139too mii crc32 agpgart reiserfs
CPU:    0
EIP:    0060:[<d0def02d>]    Not tainted VLI
EFLAGS: 00010086   (2.6.9-nitro4)
EIP is at visor_open+0x2d/0x1c0 [visor]
eax: 00000286   ebx: 00000000   ecx: ce675160   edx: d0df2dc0
esi: ceafefa0   edi: cffc8400   ebp: 00000000   esp: cd863ea0
ds: 007b   es: 007b   ss: 0068
Process pilot-xfer (pid: 3627, threadinfo=cd862000 task=cf09e000)
Stack: d0db639b 00000001 00000000 00000000 ceafefa0 cffc8400 ceafefa0 cd8a7000
       d0db33bb cffc8400 cf06b6a0 00000000 c126a600 cf06b6a0 00000000 0bc00001
       c01a96d7 cd8a7000 cf06b6a0 cd863ef4 0802a604 cd8a7000 00000001 c126a604
Call Trace:
 [<d0db33bb>] serial_open+0x8b/0x100 [usbserial]
 [<c01a96d7>] tty_open+0x217/0x290
 [<c01a94c0>] tty_open+0x0/0x290
 [<c0158cf8>] chrdev_open+0xa8/0x140
 [<c014f755>] dentry_open+0x115/0x1c0
 [<c014f632>] filp_open+0x62/0x70
 [<c014f833>] get_unused_fd+0x33/0xc0
 [<c014f959>] sys_open+0x49/0x90
 [<c010605f>] syscall_call+0x7/0xb
Code: 53 83 ec 14 a1 80 32 df d0 8b 7c 24 24 85 c0 8b 37 8b 9f e8 00 00 00 0f 85 6b 01 00 00 8b 4f 1c 85 c9 0f 84 2b 01 
00 00 9c 58 fa <c7> 43 08 00 00 00 00 c7 43 04 00 00 00 00 c7 03 00 00 00 00 50
 Segmentation fault

nitro3 works with the same config.

LittleGreenMan · Post by **LittleGreenMan** » Sun Nov 28, 2004 7:07 pm

kostian wrote:Hi, i cant sync my palm over usb with nitro4

Code: Select all

Unable to handle kernel NULL pointer dereference at virtual address 00000008
 printing eip:
*pde = 00000000
Oops: 0002 [#1]
PREEMPT
Modules linked in: visor usbserial snd_pcm_oss snd_mixer_oss uhci_hcd nvidia_agp ohci_hcd ehci_hcd forcedeth 
snd_intel8x0 snd_ac97_codec snd_pcm snd_timer snd_page_alloc gameport snd_mpu401_uart snd_rawmidi snd_seq_device snd 
soundcore nls_iso8859_1 nls_cp437 vfat fat 8250 serial_core floppy pcspkr nvidia evdev psmouse w83627hf eeprom 
i2c_sensor i2c_isa i2c_nforce2 i2c_core ipv6 8139too mii crc32 agpgart reiserfs
CPU:    0
EIP:    0060:[<d0df4037>]    Tainted: P   VLI
EFLAGS: 00010002   (2.6.9-nitro4)
EIP is at visor_open+0x37/0x1e0 [visor]
eax: 00000286   ebx: 00000000   ecx: cd5f7660   edx: cd236000
esi: ce441660   edi: c126d200   ebp: 00000000   esp: cd237e9c
ds: 007b   es: 007b   ss: 0068
Process pilot-xfer (pid: 3766, threadinfo=cd236000 task=cfb23aa0)
Stack: d0def61b 00000001 00000000 00000000 ce441660 00000001 ce441660 c126d200
       d0dec3ec c126d200 cf7846a0 00000000 cffc8400 cf7846a0 00000000 0bc00001
       c01b5da3 cd6db000 cf7846a0 cd237ef0 08028404 cd6db000 00000001 cf7846a0
Call Trace:
 [<d0dec3ec>] serial_open+0xbc/0x160 [usbserial]
 [<c01b5da3>] tty_open+0x233/0x2b0
 [<c01b5b70>] tty_open+0x0/0x2b0
 [<c0160834>] chrdev_open+0xe4/0x1d0
 [<c015660a>] dentry_open+0x14a/0x230
 [<c01564b2>] filp_open+0x62/0x70
 [<c0156729>] get_unused_fd+0x39/0xe0
 [<c0156899>] sys_open+0x49/0x90
 [<c010722b>] syscall_call+0x7/0xb
Code: 7c 24 24 85 c0 8b 37 8b 9f e8 00 00 00 0f 85 8a 01 00 00 8b 4f 1c 85 c9 0f 84 4a 01 00 00 9c 58 fa ba 00 e0 ff ff 
21 e2 ff 42 14 <c7> 43 08 00 00 00 00 c7 03 00 00 00 00 c7 43 04 00 00 00 00 50
 <6>note: pilot-xfer[3766] exited with preempt_count 1
bad: scheduling while atomic!
 [<c02ada5d>] schedule+0x41d/0x430
 [<c014747b>] unmap_page_range+0x4b/0x80
 [<c011bf07>] cond_resched_lock+0x37/0x60
 [<c01475fe>] unmap_vmas+0x14e/0x1b0
 [<c014bb43>] exit_mmap+0x83/0x160
ds: 007b   es: 007b   ss: 0068
Process pilot-xfer (pid: 3766, threadinfo=cd236000 task=cfb23aa0)
Stack: d0def61b 00000001 00000000 00000000 ce441660 00000001 ce441660 c126d200
       d0dec3ec c126d200 cf7846a0 00000000 cffc8400 cf7846a0 00000000 0bc00001
       c01b5da3 cd6db000 cf7846a0 cd237ef0 08028404 cd6db000 00000001 cf7846a0
Call Trace:
 [<d0dec3ec>] serial_open+0xbc/0x160 [usbserial]
 [<c01b5da3>] tty_open+0x233/0x2b0
 [<c01b5b70>] tty_open+0x0/0x2b0
 [<c0160834>] chrdev_open+0xe4/0x1d0
 [<c015660a>] dentry_open+0x14a/0x230
 [<c01564b2>] filp_open+0x62/0x70
 [<c0156729>] get_unused_fd+0x39/0xe0
 [<c0156899>] sys_open+0x49/0x90
 [<c010722b>] syscall_call+0x7/0xb
Code: 7c 24 24 85 c0 8b 37 8b 9f e8 00 00 00 0f 85 8a 01 00 00 8b 4f 1c 85 c9 0f 84 4a 01 00 00 9c 58 fa ba 00 e0 ff ff 
21 e2 ff 42 14 <c7> 43 08 00 00 00 00 c7 03 00 00 00 00 c7 43 04 00 00 00 00 50
 <6>note: pilot-xfer[3766] exited with preempt_count 1
bad: scheduling while atomic!
 [<c02ada5d>] schedule+0x41d/0x430
 [<c014747b>] unmap_page_range+0x4b/0x80
 [<c011bf07>] cond_resched_lock+0x37/0x60
 [<c01475fe>] unmap_vmas+0x14e/0x1b0
 [<c014bb43>] exit_mmap+0x83/0x160
 [<c011ca94>] mmput+0x64/0xb0
 [<c0120f06>] do_exit+0x156/0x430
 [<c01196b0>] do_page_fault+0x0/0x599
 [<c01084a8>] die+0x188/0x190
 [<c01196b0>] do_page_fault+0x0/0x599
 [<c011ef47>] printk+0x17/0x20
 [<c0119989>] do_page_fault+0x2d9/0x599
 [<c013dc55>] buffered_rmqueue+0xf5/0x1c0
 [<c013dc55>] buffered_rmqueue+0xf5/0x1c0
 [<c013df60>] __alloc_pages+0x240/0x400
 [<c01196b0>] do_page_fault+0x0/0x599
 [<c0107c35>] error_code+0x2d/0x38
 [<d0df4037>] visor_open+0x37/0x1e0 [visor]
 [<d0dec3ec>] serial_open+0xbc/0x160 [usbserial]
 [<c01b5da3>] tty_open+0x233/0x2b0
 [<c01b5b70>] tty_open+0x0/0x2b0
 [<c0160834>] chrdev_open+0xe4/0x1d0
 [<c015660a>] dentry_open+0x14a/0x230
 [<c01564b2>] filp_open+0x62/0x70
 [<c0156729>] get_unused_fd+0x39/0xe0
 [<c0156899>] sys_open+0x49/0x90
 [<c010722b>] syscall_call+0x7/0xb
Segmentation fault

Ive recompiled without PREEMT support but still no go:

Code: Select all

Unable to handle kernel NULL pointer dereference at virtual address 00000008
 printing eip:
*pde = 00000000
Oops: 0002 [#1]
Modules linked in: visor usbserial snd_pcm_oss snd_mixer_oss uhci_hcd nvidia_agp ohci_hcd ehci_hcd forcedeth 
snd_intel8x0 snd_ac97_codec snd_pcm snd_timer snd_page_alloc gameport snd_mpu401_uart snd_rawmidi snd_seq_device snd 
soundcore nls_iso8859_1 nls_cp437 vfat fat 8250 serial_core evdev floppy pcspkr psmouse w83627hf eeprom i2c_sensor 
i2c_isa i2c_nforce2 i2c_core ipv6 8139too mii crc32 agpgart reiserfs
CPU:    0
EIP:    0060:[<d0def02d>]    Not tainted VLI
EFLAGS: 00010086   (2.6.9-nitro4)
EIP is at visor_open+0x2d/0x1c0 [visor]
eax: 00000286   ebx: 00000000   ecx: ce675160   edx: d0df2dc0
esi: ceafefa0   edi: cffc8400   ebp: 00000000   esp: cd863ea0
ds: 007b   es: 007b   ss: 0068
Process pilot-xfer (pid: 3627, threadinfo=cd862000 task=cf09e000)
Stack: d0db639b 00000001 00000000 00000000 ceafefa0 cffc8400 ceafefa0 cd8a7000
       d0db33bb cffc8400 cf06b6a0 00000000 c126a600 cf06b6a0 00000000 0bc00001
       c01a96d7 cd8a7000 cf06b6a0 cd863ef4 0802a604 cd8a7000 00000001 c126a604
Call Trace:
 [<d0db33bb>] serial_open+0x8b/0x100 [usbserial]
 [<c01a96d7>] tty_open+0x217/0x290
 [<c01a94c0>] tty_open+0x0/0x290
 [<c0158cf8>] chrdev_open+0xa8/0x140
 [<c014f755>] dentry_open+0x115/0x1c0
 [<c014f632>] filp_open+0x62/0x70
 [<c014f833>] get_unused_fd+0x33/0xc0
 [<c014f959>] sys_open+0x49/0x90
 [<c010605f>] syscall_call+0x7/0xb
Code: 53 83 ec 14 a1 80 32 df d0 8b 7c 24 24 85 c0 8b 37 8b 9f e8 00 00 00 0f 85 6b 01 00 00 8b 4f 1c 85 c9 0f 84 2b 01 
00 00 9c 58 fa <c7> 43 08 00 00 00 00 c7 43 04 00 00 00 00 c7 03 00 00 00 00 50
 Segmentation fault

nitro3 works with the same config.

download file http://ck.kolivas.org/patches/2.6/2.6.9 ... -fix.patch
copy it in /path_to_kernel/ and do

Code: Select all

patch -p1 ./2.6.9-usb-visor-fix.patch -R

then rebulid kernel (make && make modules_install)

thecrazyperson_ws · Post by **thecrazyperson_ws** » Sun Nov 28, 2004 7:41 pm

hds wrote:
thecrazyperson_ws wrote: I've had it that way for the entire 14 hours it's been up. I believe that you have a lifetime nitro user, Seppe.
if an uptime of 14 hours gets you excited, i really wonder what kernels you tried sofar

Well, I've tried both vanilla and gentoo's 2.6.5, 2.6.6, 2.6.8, and 2.6.9, all were either too slow or extremely unstable, and any 2.4 kernel will break an NPTL bootstrap. the only reason it had only been 14 hours when I posted that was because it had rebooted after a power failure. stupid power failures.... I need to invest in a UPS. Normally, I can get a 30 day uptime out of my linux before stuff starts acting screwy / i have a power failure / etc.

kostian · Post by **kostian** » Sun Nov 28, 2004 8:30 pm

LittleGreenMan wrote: download file http://ck.kolivas.org/patches/2.6/2.6.9 ... -fix.patch
copy it in /path_to_kernel/ and do
Code: Select all
patch -p1 ./2.6.9-usb-visor-fix.patch -R 
then rebulid kernel (make && make modules_install)

Thank you, compiling it now.

pilx · Post by **pilx** » Mon Nov 29, 2004 5:56 am

skion wrote:Well, i'm using the iteraid patch, but i can surely apply it myself if no one else is using it...

I was using it too, but now I'm trying AC's new ITE8212F driver from this patch with this nitro and everything is working nice so far.

I think there's a newer version on 2.6.9-ac12 but haven't checked it out yet...

Seppe, maybe you can include this new driver in nitro.

ivanova · Post by **ivanova** » Tue Nov 30, 2004 4:12 pm

some problems with k3b here:
http://forums.gentoo.org/viewtopic.php?p=1825497

PredatorX · Post by **PredatorX** » Wed Dec 01, 2004 12:32 pm

Thanks guys!
The patch gave no chunk errors on my 2.6.9-gentoo-r6 kernel.
I put supermount on and am compiling the new kernel.
I will post if this worked.

I noticed somebody mentioning a new automount method.
I guess I missed out on that 1, can anyone post what is meant here?

thanks

miseiler · Post by **miseiler** » Wed Dec 01, 2004 10:31 pm

Just a side note for all P4 w/ HT owners trying to run reiser4 filesystems...

The patch by Pepek, while it may bring nitro4 in line with the cko3 patchset, does NOT work. Enabling P4HT and reiser4 still lock up after it tries to do any work, such as boot.

Oh well.

pilx · Post by **pilx** » Wed Dec 01, 2004 10:51 pm

I'm seeing a 20% performance loss with this kernel on a simple "hdparm -t /dev/md1" compared with my previous 2.6.8.1-love1 kernel... also CPU utilization on long/heavy disk operations is much higher.

Any ideas of why is this happening?
Have anybody seen anything like this?

pilx · Post by **pilx** » Wed Dec 01, 2004 11:01 pm

PredatorX wrote:I noticed somebody mentioning a new automount method.
I guess I missed out on that 1, can anyone post what is meant here?

Maybe you are talking about ivman? It works really good and is totally user-space. Usually I just need to...

Code: Select all

emerge ivman && /etc/init.d/ivman start

PredatorX · Post by **PredatorX** » Thu Dec 02, 2004 8:00 am

Thanks for the tip. I might try it later because I got supermount working. Do you know if this also solves the problem with mounting usb-sticks several times, so it will appear as the same device?

I don't know what you meen with userspace, is that just plain user functionality, like I don't have to fiddle with /etc/fstab?

in that case I might change things back and try it.

thanks for any info!

hds · Post by **hds** » Thu Dec 02, 2004 4:17 pm

someone in this thread mentioned, the nvidia drivers now run fine using 4k stacks. however, if you emerge the alsa-drivers, you might notice they compile fine, but you get a kernel Oops at BootUp.
i was first wondering, but then thought about 4K stacks, disabled this, and voila - alsa drivers are OK now.
just thought i mention this in case someone else is running into the same problems

OTOH, if in KDE and i set my sound to FullDuplex (ES1370 card) from kcontrol, the complete system locks up tight as a drum ;(
i dont know if this is also related to the kernel. i am investigateing this right now. if one of you mates has a workaround this would be welcomed!
system is a test system running nptl/nptlonly, and ithreads in perl. prelinked as well. just if it matters.. playing "bleeding edge", ya know

2.6.9-nitro4 "Less is more"

CONFIG_USB_SCANNER

Performance loss