Re: backporting 2.6.12.3 fixes to 2.6.8
On Mon, Aug 01, 2005 at 10:36:04PM -0600, dann frazier wrote:
> Horms said he had merged in up to 2.6.12.2, so I went through the
> 2.6.12.3 changes. Here's my notes - I've committed the three I marked
> applicable.
>
> [PATCH] If ACPI doesn't find an irq listed, don't accept 0 as a valid PCI irq.
> - Fixes a bug introduced between 2.6.8 & 2.6.12
- Does not seem to be a problem in 2.4.27
> [PATCH] revert nf_reset change
> - Reverts a broken fix that wasn't added till after 2.6.8
- Does not seem to be a problem in 2.4.27
> [PATCH] ppc32: stop misusing ntps time_offset value
- 2.6.8: Applicable (ppc32-time_offset-misuse.dpatch)
- 2.4.27: Applicable (172_ppc32-time_offset-misuse.diff)
Not present upstream, will submit
(though is ppc32 for 2.4 dead?)
> [PATCH] fix Shaper driver lossage in 2.6.12
- 2.6.8: Does not exist
- 2.4.27: Does not exist
> [PATCH] SMP fix for 6pack driver
- 2.6.8: marks 6pack as non broken on SMP; this driver has changed a lot since 2.6.8,
including some locking changes, so maybe its not a good idea to turn this on
- 2.4.27: Not applicable - driver isn't marked as broken
> [PATCH] tpm breaks 8139cp
> - drivers/char/tpm/ doesn't exist in 2.6.8
- drivers/char/tpm/ doesn't exist in 2.4.27
> [PATCH] v4l cx88 hue offset fix
> - Applicable (v4l-cx88-hue-offset-fix.dpatch)
- Does not exist in 2.4.27
> [PATCH] coverity: tty_ldisc_ref return null check
- 2.6.8: Applicable (tty_ldisc_ref-return-null-check.dpatch)
- 2.4.27: Applicable (173_v4l-cx88-hue-offset-fix.diff)
Will send upstream
N.B: Andrew Morton seems to think that the null ld this patch adds
a check for can't occur. but the fix seems harmless.
> [PATCH] uml: fix TT mode by reverting "use fork instead of clone"
> - Revert a patch that came after 2.6.8
- Does not exist in 2.4.27
> [PATCH] fix semaphore handling in __unregister_chrdev_region
> - fixed a locking bug introduced after 2.6.8
- Does not seem to be a problem in 2.4.27
--
Horms
Reply to: