Re: [GIT PULL] libfuse: run fuse servers as a contained service

["Darrick J. Wong" <djwong@kernel.org>]

On Fri, Apr 10, 2026 at 02:24:28PM +0200, Bernd Schubert wrote:
> Hi Darrick,
> 
> On 4/10/26 00:24, Darrick J. Wong wrote:
> > Hi Bernd,
> > 
> > Please pull this branch with changes for libfuse.
> > 
> > As usual, I did a test-merge with the main upstream branch as of a few
> > minutes ago, and didn't see any conflicts.  Please let me know if you
> > encounter any problems.
> 
> I had already merged another patch beginning of this week that
> introduced a new merge conflict in
> .github/workflows/install-ubuntu-dependencies.sh
> 
> I resolved it myself now.
> 
> > 
> > --D
> > 
> > The following changes since commit 3bb9cd99b3688df1bb7536de158ed9e532b277fb:
> > 
> > Add the codechecker/cppcheck github action (2026-04-07 10:08:49 +0200)
> > 
> > are available in the Git repository at:
> > 
> > https://git.kernel.org/pub/scm/linux/kernel/git/djwong/libfuse.git tags/fuse-service-container_2026-04-09
> > 
> > for you to fetch changes up to 0db39a8f8844ff88496ed36da417554cb4f3751d:
> > 
> > nullfs: support fuse systemd service mode (2026-04-09 12:31:08 -0700)
> 
> I had quite a busy last weekend, because I got libfuse security reports.
> In ordert to avoid lots of reports, I just added support for CodeChecker
> + cppcheck and CodeChecker+gcc checker. Clang CTU is is prepared, but I
> need to let AI work over the weekend to fixes for these (obviously will
> need manual editing, as last weekend).
> In the end not too bad and I don't think and urgent need for a new
> release. Although I should probably make a new 3.18.3 version soon.
> 
> That also made me to add new checks. I actually already that github
> codeql checker in place, but only realized last weekend that it didn't
> generate any reports. On your PR that gives
> 
> New alerts in code changed by this pull request
> 
> Security Alerts:
> 
>     4 critical
>     3 high
> 
> Other Alerts:
> 
>     4 warnings
>     10 notes
> 
> 
> Some of these, like "This switch statement should either handle more
> cases, or be rewritten as an if statement." can be suppressed. With some
> I probably agree
> 
> 
> 	cmd = malloc(iov.iov_len);
> 	if (!cmd) {
> 		int error = errno;
> Note
> Declaration hides parameter
> Local variable 'error' hides a parameter of the same name.
> CodeQL

Oh yikes yes.  I probably ought to have added the same "turn the gcc
warnings up real high" patch to my branch that I use for xfs stuff.

(Huh.  No warnings.  Well, that's good at least.)

> The new CodeChecker (gcc,cppchechk) actions also complain, the GUI
> should give the output, but I don't see it yet.

<nod> Any reports you get about my branch, please feel free to lob it at
me and tell me to fix it :)

I can run cppcheck locally, so I'll try to fix all the warnings that
your copy is throwing out, and send you an updated pr.

> My plan for the weekend is to address your comments in my series and
> then to merge it. If still some time left, I would like to fix all the
> reports in your series and merge that as well.
> Then maybe next week to combine some mount code dup that comes from both
> series.
> 
> Ok with you?

Yes, that sounds good.  Let me know if you want any help. :)

--D

> 
> Thanks,
> Bernd