Re: [newbie] problems with 2.6.0

2004-01-12 Thread tuija t.
Tom Brinkman kirjoitti viestissään (lähetysaika Sunnuntai 11. Tammikuuta 2004 
17:20):
 On Sunday 11 January 2004 12:33 am, tuija t. wrote:
   I believe from reading cooker, he's got many of 2.6's
   module problems solved, proper ReiserFS support, and the
   latest supermount and other Mandrake specifics.  Be aware tho
   you'll still have other 2.6 kernel hardware/software/config
   issues (see lkml and cooker archives). So report back and
   tells us how it goes.
 
  Thank you, I must try that Thomas' kernel-source too, it seems
  to have lot of intresting patches.

 Well , you should read cooker or at least the archive. There
 where a lot of problems that some testers experienced with his
 first 2.6.0, and now 2.6.1 efforts (latest is 2.6.1-3tmb). They
 are considered better than the regular cooker 2.6.x's tho.  I'm
 also reading on both cooker and lkml that 2.6.1 is unstable, even
 more so than 2.6.0.  Look here for the changelog
 http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.1

 IMO, 2.6.x won't be ready for (newbie) user or enterprise
 service until late next summer.  I hope I'm wrong, but I suspect
 I'm not.  I believe Mandrake should wait til 10.1, not 10.0 to
 introduce this drastically new and differnent kernel. Still, in
 the meantime, sombody needs to be testin it. Guess that's us ;)

That could be so, at least for production servers or other complex
hardware.
To my simple laptop (fujitsu c1020) I have tested 2.6.0 since test2
and after 2.6.0-test5 I haven't experience any real problems.
I also have used 2.6.0 with several distros (debian, slackware, fedora)
and now mdk9.2.
I have always used vanilla source without initrd, only in this mdk's case
I merged that supermount-patch to vanilla.
I know that 2.6 has bring up some buggy code in many apps for example
old kmail when used on reiserfs that andrew mentioned in vger and some chipset 
has turned out to be buggy too.
I have never tried SUSE but heard some rumors they ship their 9.0 with
2.6?

-- 
best rgds
~tt


Want to buy your Pack or Services from MandrakeSoft? 
Go to http://www.mandrakestore.com


Re: [newbie] problems with 2.6.0

2004-01-11 Thread Tom Brinkman
On Sunday 11 January 2004 12:33 am, tuija t. wrote:
      I believe from reading cooker, he's got many of 2.6's
  module problems solved, proper ReiserFS support, and the
  latest supermount and other Mandrake specifics.  Be aware tho
  you'll still have other 2.6 kernel hardware/software/config
  issues (see lkml and cooker archives). So report back and
  tells us how it goes.

 Thank you, I must try that Thomas' kernel-source too, it seems
 to have lot of intresting patches.

Well , you should read cooker or at least the archive. There 
where a lot of problems that some testers experienced with his 
first 2.6.0, and now 2.6.1 efforts (latest is 2.6.1-3tmb). They 
are considered better than the regular cooker 2.6.x's tho.  I'm 
also reading on both cooker and lkml that 2.6.1 is unstable, even 
more so than 2.6.0.  Look here for the changelog
http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.1

IMO, 2.6.x won't be ready for (newbie) user or enterprise 
service until late next summer.  I hope I'm wrong, but I suspect 
I'm not.  I believe Mandrake should wait til 10.1, not 10.0 to 
introduce this drastically new and differnent kernel. Still, in 
the meantime, sombody needs to be testin it. Guess that's us ;)
-- 
  Tom Brinkman Corpus Christi, Texas

Want to buy your Pack or Services from MandrakeSoft? 
Go to http://www.mandrakestore.com


Re: [newbie] problems with 2.6.0

2004-01-10 Thread tuija t.
Tom Brinkman kirjoitti viestissään (lähetysaika Perjantai 9. Tammikuuta 2004 
17:04):
 On Thursday 08 January 2004 01:12 pm, tuija t. wrote:
  I answer to myself :)
  I compiled vanilla 2.6.0 and;
  + no errors
  + no reiserfs debug mode
  + no frame pointers
  But
  - no supermount either :(
  Howto merge supermount patch to vanilla?

 I've had a lot of problems with recent 2.6 cooker kernels
 myself, particulary with initrd and Reiser (won't even boot).
 You might try lookin in contribs (RPMS2/) on cooker mirrors for
 Thomas Backlund's version.
 kernel-tmb-2.6.0-1.tmb.1mdk-1-1mdk.i586.rpm
   (kernel-source is there too)

 I believe from reading cooker, he's got many of 2.6's module
 problems solved, proper ReiserFS support, and the latest
 supermount and other Mandrake specifics.  Be aware tho you'll
 still have other 2.6 kernel hardware/software/config issues (see
 lkml and cooker archives). So report back and tells us how it
 goes.

Thank you, I must try that Thomas' kernel-source too, it seems to
have lot of intresting patches.

The only way I'm aware of to get Mandrake kernel patches like
 supermount is to extract them from a Mandrake kernel src.rpm, and
 I don't know how compatible they'd be with vanilla kernel.org
 source.

I managed to merge this supermount patch
supermount-2.0.3a-2.6.0.patch.gz
to vanilla-kernel-2.6.1 with following steps:
Extract the kernel vanilla in /usr/src
then copy this patch to /usr/src/linux-2.6.1/
extract the patch
cd in the directory /usr/src/linux-2.6.1
and there
# patch -p1  supermount-2.0.3a-2.6.0
# make mrproper
# make
# make modules_install
# make install
Everything else went well but system failed to create
initrd.img
But I personally don't need it cause I have choosen
y not m in ide/ata and /fs  which is reiserfs
Kernel seems to work well and there is no errors in dmesg
-- 
best rgds
~tt


Want to buy your Pack or Services from MandrakeSoft? 
Go to http://www.mandrakestore.com


[newbie] problems with 2.6.0

2004-01-08 Thread tuija t.
Hello,

I installed:
Linux version 2.6.0-1mdk ([EMAIL PROTECTED]) (gcc version 3.3.2 
(Mandrake Linux 10.0 3.3.2-3mdk)) #1 Wed Dec 31 16:54:15 CET 2003

From cooker contrib, and also module-init-tools,
to my Mdk9.2 which is in my laptop Fujitsu C1020
and I got in dmesg following symbols:

Mounted devfs on /dev
SCSI subsystem initialized
sbp2: Unknown symbol hpsb_set_hostinfo_key
sbp2: Unknown symbol hpsb_get_hostinfo
sbp2: Unknown symbol hpsb_node_write
sbp2: Unknown symbol hpsb_register_addrspace
sbp2: Unknown symbol hpsb_node_fill_packet
sbp2: Unknown symbol hpsb_register_protocol
sbp2: Unknown symbol hpsb_make_writepacket
sbp2: Unknown symbol hpsb_speedto_str
sbp2: Unknown symbol hpsb_send_packet
sbp2: Unknown symbol hpsb_unregister_protocol
sbp2: Unknown symbol hpsb_get_host_bykey
sbp2: Unknown symbol ieee1394_bus_type
sbp2: Unknown symbol hpsb_destroy_hostinfo
sbp2: Unknown symbol hpsb_create_hostinfo
sbp2: Unknown symbol free_hpsb_packet
sbp2: Unknown symbol hpsb_register_highlevel
sbp2: Unknown symbol hpsb_free_tlabel
sbp2: Unknown symbol hpsb_unregister_highlevel
sbp2: Unknown symbol hpsb_set_packet_complete_task
sbp2: Unknown symbol hpsb_get_hostinfo_bykey

Should I remove devfs=mount from menu.lst kernel line?

found reiserfs format 3.6 with standard journal
reiserfs:warning: CONFIG_REISERFS_CHECK is set ON
reiserfs:warning: - it is slow mode for debugging.

Why they set this debug on, ain't 2.6.0 should be stable?
And then no need to debug?

Something more strange:
Call Trace:
 [c011f0fa] __might_sleep+0x9a/0xc0
 [c0143a66] kmem_cache_alloc+0x46/0x60
 [c011f4e1] dup_task_struct+0x21/0xa0
 [c0120161] copy_process+0xb1/0xc00
 [c011cb3c] recalc_task_prio+0x9c/0x200
 [c0120cec] do_fork+0x3c/0x170
 [e08c3130] supermount_show_sbi+0x1d0/0x240 [supermount]
 [e08cf14d] ohci_init_config_rom+0x34d/0x460 [ohci1394]
 [c01092e4] kernel_thread+0x64/0x70
 [e09166d0] nodemgr_host_thread+0x0/0x180 [ieee1394]
 [c0109270] kernel_thread_helper+0x0/0x10
 [e0916a8c] nodemgr_add_host+0xbc/0x110 [ieee1394]
 [e09166d0] nodemgr_host_thread+0x0/0x180 [ieee1394]
 [e091297a] highlevel_add_host+0x4a/0x80 [ieee1394]
 [e0911f5e] hpsb_add_host+0x5e/0x80 [ieee1394]
 [e08cf61d] ohci1394_pci_probe+0x33d/0x480 [ohci1394]
 [c01b3197] pci_device_probe_static+0x37/0x60
 [c01b31e9] __pci_device_probe+0x29/0x40
 [c01b321e] pci_device_probe+0x1e/0x40
 [c01ee77f] bus_match+0x2f/0x70
 [c01ee8ad] driver_attach+0x4d/0x90
 [c01eeb3d] bus_add_driver+0x7d/0xa0
 [c01eef2f] driver_register+0x2f/0x40
 [c01b33d0] pci_register_driver+0x50/0x80
 [e076e00e] ohci1394_init+0xe/0x3b [ohci1394]
 [c01377e0] sys_init_module+0x110/0x280
 [c010b277] syscall_call+0x7/0xb

Is it better choice to pick up vanilla source from kernel.org
and compile own custom version?
But what about supermount then, is it included in vanilla
or does it have to merge from some patch?
I like to use supermount myself.

-- 
best rgds
~tt


Want to buy your Pack or Services from MandrakeSoft? 
Go to http://www.mandrakestore.com


Re: [newbie] problems with 2.6.0

2004-01-08 Thread tuija t.
tuija t. kirjoitti viestissään (lähetysaika Torstai 8. Tammikuuta 2004 19:12):
 Hello,

 I installed:
 Linux version 2.6.0-1mdk ([EMAIL PROTECTED]) (gcc version 3.3.2
 (Mandrake Linux 10.0 3.3.2-3mdk)) #1 Wed Dec 31 16:54:15 CET 2003

 From cooker contrib, and also module-init-tools,
 to my Mdk9.2 which is in my laptop Fujitsu C1020
 and I got in dmesg following symbols:

 Mounted devfs on /dev
 SCSI subsystem initialized
 sbp2: Unknown symbol hpsb_set_hostinfo_key
 sbp2: Unknown symbol hpsb_get_hostinfo
 sbp2: Unknown symbol hpsb_node_write
 sbp2: Unknown symbol hpsb_register_addrspace
 sbp2: Unknown symbol hpsb_node_fill_packet
 sbp2: Unknown symbol hpsb_register_protocol
 sbp2: Unknown symbol hpsb_make_writepacket
 sbp2: Unknown symbol hpsb_speedto_str
 sbp2: Unknown symbol hpsb_send_packet
 sbp2: Unknown symbol hpsb_unregister_protocol
 sbp2: Unknown symbol hpsb_get_host_bykey
 sbp2: Unknown symbol ieee1394_bus_type
 sbp2: Unknown symbol hpsb_destroy_hostinfo
 sbp2: Unknown symbol hpsb_create_hostinfo
 sbp2: Unknown symbol free_hpsb_packet
 sbp2: Unknown symbol hpsb_register_highlevel
 sbp2: Unknown symbol hpsb_free_tlabel
 sbp2: Unknown symbol hpsb_unregister_highlevel
 sbp2: Unknown symbol hpsb_set_packet_complete_task
 sbp2: Unknown symbol hpsb_get_hostinfo_bykey

 Should I remove devfs=mount from menu.lst kernel line?

 found reiserfs format 3.6 with standard journal
 reiserfs:warning: CONFIG_REISERFS_CHECK is set ON
 reiserfs:warning: - it is slow mode for debugging.

 Why they set this debug on, ain't 2.6.0 should be stable?
 And then no need to debug?

 Something more strange:
 Call Trace:
  [c011f0fa] __might_sleep+0x9a/0xc0
  [c0143a66] kmem_cache_alloc+0x46/0x60
  [c011f4e1] dup_task_struct+0x21/0xa0
  [c0120161] copy_process+0xb1/0xc00
  [c011cb3c] recalc_task_prio+0x9c/0x200
  [c0120cec] do_fork+0x3c/0x170
  [e08c3130] supermount_show_sbi+0x1d0/0x240 [supermount]
  [e08cf14d] ohci_init_config_rom+0x34d/0x460 [ohci1394]
  [c01092e4] kernel_thread+0x64/0x70
  [e09166d0] nodemgr_host_thread+0x0/0x180 [ieee1394]
  [c0109270] kernel_thread_helper+0x0/0x10
  [e0916a8c] nodemgr_add_host+0xbc/0x110 [ieee1394]
  [e09166d0] nodemgr_host_thread+0x0/0x180 [ieee1394]
  [e091297a] highlevel_add_host+0x4a/0x80 [ieee1394]
  [e0911f5e] hpsb_add_host+0x5e/0x80 [ieee1394]
  [e08cf61d] ohci1394_pci_probe+0x33d/0x480 [ohci1394]
  [c01b3197] pci_device_probe_static+0x37/0x60
  [c01b31e9] __pci_device_probe+0x29/0x40
  [c01b321e] pci_device_probe+0x1e/0x40
  [c01ee77f] bus_match+0x2f/0x70
  [c01ee8ad] driver_attach+0x4d/0x90
  [c01eeb3d] bus_add_driver+0x7d/0xa0
  [c01eef2f] driver_register+0x2f/0x40
  [c01b33d0] pci_register_driver+0x50/0x80
  [e076e00e] ohci1394_init+0xe/0x3b [ohci1394]
  [c01377e0] sys_init_module+0x110/0x280
  [c010b277] syscall_call+0x7/0xb

 Is it better choice to pick up vanilla source from kernel.org
 and compile own custom version?
 But what about supermount then, is it included in vanilla
 or does it have to merge from some patch?
 I like to use supermount myself.

I answer to myself :)
I compiled vanilla 2.6.0 and;
+ no errors
+ no reiserfs debug mode
+ no frame pointers
But
- no supermount either :(
Howto merge supermount patch to vanilla?

-- 
best rgds
~tt


Want to buy your Pack or Services from MandrakeSoft? 
Go to http://www.mandrakestore.com