Re: [PATCH] fbdev/hpfb: Fix an error handling path in hpfb_dio_probe()

From: Christophe JAILLET
Date: Mon Jul 29 2024 - 16:31:08 EST


Le 29/07/2024 à 22:09, Helge Deller a écrit :
On 7/29/24 17:59, Dan Carpenter wrote:
On Mon, Jul 29, 2024 at 10:13:17AM +0200, Helge Deller wrote:
On 7/28/24 20:29, Christophe JAILLET wrote:
If an error occurs after request_mem_region(), a corresponding
release_mem_region() should be called, as already done in the remove
function.

True.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")

I think we can drop this "Fixes" tag, as it gives no real info.

If we're backporting patches then these tags really are useful.  As
I've been doing more and more backporting, I've come to believe this
more firmly.

Sure, "Fixes" tags are useful, but only if they really refer
to another patch which introduced the specific issue.

But the tag 1da177e4c3f4 ("Linux-2.6.12-rc2") isn't useful, as it's
just the initial git commit. It has no relation to why release_mem_region()
might have been initially missed. See:

I agree that the description of this specific tag is not useful by itself.

But at least it means: should it be backported, it can be done up to there. (and sometimes LWN gives some statistics on how long it took to fix an "issue", should it be considered as such)

Without it, it is not easy to guess in which branch the patch is meaningful.

I'll sent a v2 with your suggested minimal change, but I'll keep the Fixes tag.


Up to you to remove it or not, and to add a <stable@xxxxxxxxxx> or a <stable+noautosel@xxxxxxxxxx> or none of them.

Any solution is fine with me.



 commit 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 (tag: v2.6.12-rc2)
Author: Linus Torvalds <torvalds@xxxxxxxxxxxxxxx>
Date:   Sat Apr 16 15:20:36 2005 -0700

    Linux-2.6.12-rc2

    Initial git repository build. I'm not bothering with the full history,
    even though we have it. We can create a separate "historical" git
    archive of that later if we want to, and in the meantime it's about
    3.2GB when imported into git - space that would just make the early
    git days unnecessarily complicated, when we don't have a lot of good
    infrastructure for it.

Helge



On:

> HP300 are old HP machines with an m68k CPU.
> Not sure if someone still has such a machine 🙂

so it really was the one I found on wikipedia, LoL!

So, another way to "fix" the issue is maybe to deprecate the driver or everything related to this old architecture?

No strong opinion about it.

CJ