Author Topic: Need new dkms for 2.6.35.2 kernel, a major attack vector has been plugged  (Read 904 times)

Offline Almost-retired

  • Sr. Member
  • ****
  • Posts: 252
    • What keeps Gene out of the bars
I have built 2.6.35.2, and it boots normally, including dkms installing the nvidia drivers, but something is amiss as x refuses to run.

Is it possible to get a newer dkms, or is is possible to edit one of the config files and make it work again?

Thanks.

--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Obviously I was either onto something, or on something.
             -- Larry Wall on the creation of Perl





Offline Almost-retired

  • Sr. Member
  • ****
  • Posts: 252
    • What keeps Gene out of the bars
Re: Need new dkms for 2.6.35.2 kernel, [SOLVED]
« Reply #1 on: August 22, 2010, 09:30:25 AM »
It turns out I didn't, it was a more than once build, and dkms isn't capable of handling a rebuild of the kernel when the `uname -r` remains the same.

So, I expect that in the long run, I will probably need to add a few more lines to my "./makeit" script that builds and installs a kernel, to duplicate what I do with kernels, which is to rename the existing versions with an appended ".old".  The problem was of course in  /var/lib/dkms, and in this case I just blew away the current, no longer valid `uname -r` stuff and rebooted.  So then it was dutifully rebuilt and reinstalled, and has worked fine for both 2.6.35.2 and 2.6.35.3 now.

pclos needs to get this kernel out as it fixes a long standing X based exploit allowing root access to any user.  There are, it appears, exploits in the wild already.

--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Obviously I was either onto something, or on something.
             -- Larry Wall on the creation of Perl

Offline Texstar

  • Administrator
  • Super Villain
  • *****
  • Posts: 12524
dkms-nvidia drivers build fine against this kernel however there is a bug in the kernel relating to the snd-hda-intel driver.

WARNING: at drivers/pci/pci.c:105 pci_ioremap_bar+0x2c/0x5b()
Hardware name: MS-7253
Modules linked in: snd_hda_intel(+) snd_hda_codec snd_hwdep snd_seq_dummy snd_seq_oss(+) snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_pcm snd_timer snd_mixer_oss snd soundcore via_rhine shpchp sg sr_mod cdrom snd_page_alloc mii i2c_viapro pcspkr serio_raw i2c_core pci_hotplug binfmt_misc cpufreq_ondemand cpufreq_conservative cpufreq_powersave powernow_k8 freq_table mperf ohci1394 sbp2 raw1394 ieee1394 fuse wacom ppdev floppy parport_pc parport k8temp evdev dm_mod thermal button processor ide_generic pata_via ata_generic ide_pci_generic ide_gd_mod ide_core pata_acpi sata_via libata sd_mod scsi_mod crc_t10dif ext4 jbd2 crc16 uhci_hcd ohci_hcd ehci_hcd hid_logitech usbhid hid usbcore ff_memless
Pid: 1174, comm: modprobe Not tainted 2.6.35.3-pclos1 #1
Call Trace:
 [<c01345a4>] warn_slowpath_common+0x6a/0x7f
 [<c0264fcd>] ? pci_ioremap_bar+0x2c/0x5b
 [<c01345cd>] warn_slowpath_null+0x14/0x18
 [<c0264fcd>] pci_ioremap_bar+0x2c/0x5b
 [<f88eef9c>] azx_probe+0x355/0x951 [snd_hda_intel]
 [<c01ff1d9>] ? sysfs_add_one+0x18/0xc6
 [<c02651b8>] local_pci_probe+0x13/0x15
 [<c0265d90>] pci_device_probe+0x48/0x6b
 [<c02da1e4>] driver_probe_device+0x94/0x11c
 [<c02da2ad>] __driver_attach+0x41/0x59
 [<c02d998d>] bus_for_each_dev+0x42/0x6c
 [<c02da098>] driver_attach+0x19/0x1b
 [<c02da26c>] ? __driver_attach+0x0/0x59
 [<c02d9d66>] bus_add_driver+0x94/0x1c6
 [<c0251dc7>] ? kset_find_obj+0x23/0x4f
 [<c02da507>] driver_register+0x7e/0xe5
 [<c039b91f>] ? notifier_call_chain+0x2b/0x4d
 [<c0265f66>] __pci_register_driver+0x3d/0x9a
 [<c014d0c8>] ? __blocking_notifier_call_chain+0x45/0x51
 [<f8931000>] ? alsa_card_azx_init+0x0/0x19 [snd_hda_intel]
 [<f8931017>] alsa_card_azx_init+0x17/0x19 [snd_hda_intel]
 [<c0101154>] do_one_initcall+0x62/0x16b
 [<c015d27f>] sys_init_module+0x7c/0x19d
 [<c01032df>] sysenter_do_call+0x12/0x28
---[ end trace c2f39a6810599191 ]---
hda-intel: ioremap error
HDA Intel 0000:80:01.0: PCI INT A disabled

Thanks to everyone who donates. You keep the servers running.

Offline rubentje1991

  • PCLinuxOS Tester
  • Hero Member
  • *******
  • Posts: 2110
  • Rubenus Parvus MCMXCI
ouch....
http://identi.ca/pclinuxos
Been working on Kernel 2.6.35.3 all day. Got a alsa snd_hda_intel module crashing with hda-intel: ioremap error with via chipset : VT1708/A

---------

ALL DAY?
what's called perseverance.....?  :P

Offline Almost-retired

  • Sr. Member
  • ****
  • Posts: 252
    • What keeps Gene out of the bars
Re: Need new dkms for 2.6.35.2 kernel, a major attack vector has been plugged
« Reply #4 on: September 20, 2010, 11:16:09 AM »
I wrote a bash script that I call from my kernel 'makeit' script, which fixed it for 2.6.35.2,3,4 but now it all
fails with an error 10 for 2.6.35.5-rc1.

After writing that script, I tried to make it work for 2.6.36-rc3, but that wiped it all out till I had rebooted
to 2.6.35.4 and got a level3 terminal and ran mcc from that, otherwise x was being started with a pure white screen.
This is with nvidia-current.  I am on 2.6.35.5-rc1 ATM, but the vesa driver is being used.  No fancy x stuff.