Difference between revisions of "User talk:Thinker"

From ThinkWiki
Jump to: navigation, search
(Security Issue with pam)
 
Line 65: Line 65:
  
 
--[[User:tony2001|tony2001]]
 
--[[User:tony2001|tony2001]]
 +
 +
== Security Issue with pam ==
 +
 +
Hey, I don't know if you ever saw my suggestion [[Talk:How_to_enable_the_fingerprint_reader#Fingerprint_or_password|here]], but I found that it's not very secure and posted a better common-auth file that works. -[[User:Tro|Tro]] 22:48, 13 September 2007 (UTC)

Latest revision as of 23:48, 13 September 2007

Hi,

regarding those issues with R&R from grub: two things seem quite crucial to me after a bit of experimenting: one must not touch the mbr and one must not touch the partition type for that recovery partition (T43). As long as these two are observed, the rest seems to be fine. I am using this in my grub.conf:

title=IBM rescue and recover rootnoverify (hd0,1) chainloader +1

Partition type is 0x12 (compaq diagnostics in fdisk) and the MBR is T43's "factory" one.

Igor


I had to put GRUB in the MBR, since the default MBR would always boot into the Windows partition, even if I put GRUB in a primary partition and marked (only) that partition bootable.

--Thinker 01:10, 25 January 2006 (CET)


Okey, here is what I did. I hosed the MBR during my bizarre experiments, but IBM has been kind enough to release a floppy fixer for that (somewhere else on the wiki). In no particular order:

  • Restore the original MBR
  • Install grub into the boot partition (sda3). Not the drive (sda), but the boot partition (sda3).
  • Mark sda3 as bootable. This must be the ONLY bootable partition.
  • Make sure that the partition types are correct: NTFS for WinXP (sda1), 0x12 for R&R (sda2), 0x83 for /boot
  • Conjure a suitable grub.conf. Mine is:
 title=Gentoo Linux 2.6.15-r1 (gentoo-sources) + (ibm-acpi)
 root (hd0,2)
 kernel /vmlinuz-2.6.15-gentoo-r1 acpi_sleep=s3_bios pci=noacpi libata.atapi_enabled=1
 title=Windows XP
 rootnoverify (hd0,0)
 chainloader +1

That's it folks :) --Igor


Another thing to try for restauring the MBR is to restore the laptop to the factory state. If other OSes are *NOT* on the first primary partition, a little fiddling with the partition layout in fdisk should restore the partition picture before the restore (factory state restore will restore windows, but it will not touch any partitions except for the first one).

--Igor


Thanks, but I did exactly that, and it never booted into GRUB. No idea why; I triple-checked the bootable partition flags, but the factory MBR just didn't seem to care about them. Eventually I just gave up on R&R and relaimed its 4GB of diskspace (of course, I have backup discs). Is there something particularly exciting that R&R can and IBM's PC Doctor bootable CD can't do?

--Thinker 09:47, 25 January 2006 (CET)


Not that I know of. The only useful scenario is a diagnostic by PC doctor, without any rescue CDs around. BIOS POST -> R&R -> reboot -> Diagnostic. It is just that it is supposed to work. Strange that it did not boot into GRUB.

--Igor

>Regarding your edit to tp_smapi: how does inhbit_charge_minutes fail, and what's does dmesg say when you write to that attribute?

It does not fail, but it doesn't take any effect. dmesg says that inhabit_charge_minutes has been set, but the battery's state is still "idle".

  1. echo 1 > /sys/devices/platform/smapi/BAT0/inhibit_charge_minutes
  2. dmesg

... smapi smapi: set_inhibit_charge_minutes: set to 1 for bat=0

  1. cat /sys/devices/platform/smapi/BAT0/state

idle

--tony2001

Security Issue with pam

Hey, I don't know if you ever saw my suggestion here, but I found that it's not very secure and posted a better common-auth file that works. -Tro 22:48, 13 September 2007 (UTC)