Re: [PATCH v2] Squashfs: check block offset is not negative

On Thu, Aug 06, 2026 at 08:54:14PM +0100, Phillip Lougher wrote:
> > > I may be mistaken, but maybe updating the "T:" entry would resolve the issues
> > > you are seeing? Is this the same tree that you described as "purged"/unused?
> > > 
> > 
> > Yes, the git URL (the T: entry) is out of date.
> > 
> > But, until you mentioned it now, I didn't know it existed,
> > because I never added it.
> > 
> > It was added in commit https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=9f273c24ec5f4
> > 
> > author    Fengguang Wu <[email protected]>    2016-01-20 15:03:25 -0800
> > committer    Linus Torvalds <[email protected]>    2016-01-20 17:09:18 -0800
> > commit    9f273c24ec5f4a6f785bb83e931b3808a07b459e (patch)
> > tree    7bc0126edb6ded71bec82e4c057768d76928b2f5
> > parent    b2807f07f4f87362925b8a5b8cbb7b624da10f03 (diff)
> > download    linux-9f273c24ec5f4.tar.gz
> > MAINTAINERS: add/fix git URLs for various subsystems
> > 
> > @@ -10217,6 +10242,7 @@ SQUASHFS FILE SYSTEM
> >   M:    Phillip Lougher <[email protected]>
> >   L:    [email protected] (subscribers-only)
> >   W:    http://squashfs.org.uk
> > +T:    git git://git.kernel.org/pub/scm/linux/kernel/git/pkl/squashfs-next.git
> >   S:    Maintained
> >   F:    Documentation/filesystems/squashfs.txt
> >   F:    fs/squashfs/
> > 
> > I will send a patch to remove it.
> > 
> > Thanks
> > 
> > 
> 
> I have sent a patch to remove that dead URL from the MAINTAINERS file.
> 
> But, the point is, is that enough to make Sashiko default to using Linus
> Torvald's upstream development kernel for Squashfs patches?

It should be. I'm adding a docs entry in the MAINTAINERS_GUIDE for more clarity.
[0]

It will most likely be merged when Roman returns.

> 
> As I said previously I lost access to git.kernel.org in 2015 (if you ask
> why I will tell you privately).  So I can't put a kernel tree on there,
> but I could on Github (the kernel is already on Github and so it will be a
> simple fork).  But it seems a PITA to do that just for Sashiko.

Let's wait and see. Please feel free to continue to forward/cc errors here. If
changing the maintainer's entry doesn't work out, we can add some escape hatch
for squash-fs (like we do for MM) in baseline.rs.

<snip>
        // Check for Linux-MM special handling
        // If the top candidate is akpm/mm or linux-mm, OR the subsystem is MEMORY MANAGEMENT
        let (top_url, _top_branch) = candidates[0].0;
        let is_mm = top_url.contains("akpm/mm")
            || top_url.contains("linux-mm")
            || matched_subsystem_name
                .as_deref()
                .map(|s| s.eq_ignore_ascii_case("MEMORY MANAGEMENT"))
                .unwrap_or(false);

        if is_mm {
            // For linux-mm, we prioritize specific branches: mm-new, mm-unstable, mm-stable
            // We use the discovered URL (likely akpm/mm)
            let mm_url = top_url;
            return vec![
                self.resolve_url(mm_url, Some("mm-new".to_string())),
                self.resolve_url(mm_url, Some("mm-unstable".to_string())),
                self.resolve_url(mm_url, Some("mm-stable".to_string())),
            ];
        }
</snip>

> 
> Phillip
>   >>
> > > Thanks,
> > > 
> > 
> s
> 

[0] https://github.com/sashiko-dev/sashiko/pull/398/changes

Cheers,

-- 
Derek <[email protected]>