bootstd: Scanning for USB bootflow will remove existing SCSI bootflow

[Tony Dinh]

On Mon, Nov 6, 2023 at 11:51 AM Tom Rini <trini at konsulko.com> wrote:
>
> On Mon, Nov 06, 2023 at 11:42:51AM -0800, Tony Dinh wrote:
> > Hi Simon,
> >
> > On Mon, Nov 6, 2023 at 9:25 AM Simon Glass <sjg at chromium.org> wrote:
> > >
> > > Hi Tony,
> > >
> > > On Wed, 4 Oct 2023 at 21:23, Tony Dinh <mibodhi at gmail.com> wrote:
> > > >
> > > > Hi Simon,
> > > >
> > > > On Mon, Oct 2, 2023 at 12:25 PM Tony Dinh <mibodhi at gmail.com> wrote:
> > > > >
> > > > > Hi Simon,
> > > > >
> > > > > On Sun, Oct 1, 2023 at 6:22 PM Simon Glass <sjg at chromium.org> wrote:
> > > > > >
> > > > > > Hi Tony,
> > > > > >
> > > > > > On Tue, 26 Sept 2023 at 13:40, Tony Dinh <mibodhi at gmail.com> wrote:
> > > > > > >
> > > > > > > Hi Simon,
> > > > > > >
> > > > > > > On Tue, Sep 26, 2023 at 4:37 AM Simon Glass <sjg at chromium.org> wrote:
> > > > > > > >
> > > > > > > > Hi Tony,
> > > > > > > >
> > > > > > > > On Mon, 25 Sept 2023 at 14:02, Tony Dinh <mibodhi at gmail.com> wrote:
> > > > > > > > >
> > > > > > > > > Hi Simon,
> > > > > > > > >
> > > > > > > > > Here is an observation during testing the bootflow command.
> > > > > > > > >
> > > > > > > > > If there is a SCSI bootflow, scanning for USB bootflow will remove that existing
> > > > > > > > > SCSI bootflow. To bring it back, I scanned for SCSI bootflow again, and it was
> > > > > > > > > back to normal. Perhaps there is some kind of indexing problem?
> > > > > > > >
> > > > > > > > Yes that's right. The 'botflow scan' command is not additive. The
> > > > > > > > first thing it does is removing existing bootflows.
> > > > > > >
> > > > > > > Thanks for clarifying that. I assumed it is additive, because the
> > > > > > > existing USB bootflow was not removed when I did a "bootflow scan
> > > > > > > scsi" immediately after (see the end of the log).
> > > > > >
> > > > > > Yes, but I'm not sure what is going on there. Perhaps it is a bug?
> > > > > > When you scan SCSI it should not also scan USB.
> > > > >
> > > > > Yes, it looks like a bug. I think I can see the problem and am working
> > > > > on a patch.
> > > >
> > > > Just to let you know, I'm out of time to work on this topic. I will
> > > > revisit it some other time, if you have not already tracked it down.
> > >
> > > If you are able to use 'bootflow scan -l' in each case, then I might
> > > be able to see what it is scanning USB, when you are asking for SCSI.
> > > I also wonder if this bug may have been fixed due to other changes?
> >
> > I'm wondering about that, too. I will rerun the test with the master
> > branch with debug trace. If we cannot see it, I'll do it again when
> > other bootstd patches are merged. Right now I am not sure which of
> > your bootstd patches I should apply to the tree for testing, do you
> > have a list for that?
>
> Right now the only outstanding bootstd patch I know of is your recent
> one, as I was waiting for Simon's review (which he did today) before
> merging and applying.

Thanks Tom! I'll run the test in my tree as is then.

All the best,
Tony