Received-SPF: pass (google.com: domain of linux-ext4-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Mon, 30 May 2022 21:37:37 -0400
From:   "Theodore Ts'o" <tytso@mit.edu>
To:     "Stephen E. Baker" <baker.stephen.e@gmail.com>
Cc:     adilger.kernel@dilger.ca, hch@lst.de, linux-ext4@vger.kernel.org
Subject: Re: simplify ext4_sb_read_encoding regression
Message-ID: <YpVxYchs1wScNRDw@mit.edu>
References: <CAFDdnB1Rq3vNe_qt_0u+inzOuL4vrGhgbOoQZKBwfBktni=Npw@mail.gmail.com>
 <YpLLTkje/QUYPP9z@mit.edu>
 <YpNk4bQlRKmgDw8E@mit.edu>
 <CAFDdnB0a3mfcoY7rg5N4dO13qMeSsV+PkA2YHeerEOFRv8484Q@mail.gmail.com>
 <YpQixl+ljcC1VHNU@mit.edu>
 <CAFDdnB1WxrqddeLJwjsqqgoij1q_QGa=SBs-i=j31W2QbksJ=Q@mail.gmail.com>
 <YpVeeKRH1bycP7P1@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YpVeeKRH1bycP7P1@mit.edu>
Precedence: bulk

On Mon, May 30, 2022 at 08:16:56PM -0400, Theodore Ts'o wrote:
> On Mon, May 30, 2022 at 06:20:46PM -0400, Stephen E. Baker wrote:
> > The other thought I had was - this device has a strict size limit on kernels,
> > which is why I had to start building my own in the first place. Nothing over
> > 32MB can boot. However I verified my kernel was smaller than that. I
> > minimized my config even further and my Image is now 20MB uncompressed
> > which is smaller than it was before, so that's not it either.
> > 
> > Here is my full config:
> 
> I don't know what to tell you.  I took your config, stripped out all
> of the modules, and enabled CONFIG_HYPERVISOR_GUEST,
> CONFIG_VIRTIO_MENU, and CONFIG_VIRTIO_BLK, and build a 5.16 kernel.

Sorry, this should have been 5.17 kernel.

> This worked just fine using kvm-xfstests booting into a root file
> system that had case folding enabled.  (See my previous e-mail for a
> description for how to test things.)
> 
> So I can't replicate your problem.  Without the boot logs, I'm not
> sure what else we can do here.  Maybe you can figure out a way to
> replicate the problem using kvm-xfstests?  We really need a solid
> reproducer on a system where we can debug things and see the kernel
> logs.
> 
> 						- Ted