[PATCH v3 00/46] pxe: Support read_all() for extlinux and PXE

Sun Feb 16 17:15:16 CET 2025

On Sun, Feb 16, 2025 at 07:09:57AM -0700, Simon Glass wrote:
> Hi Tom,
> 
> On Sat, 15 Feb 2025 at 11:21, Tom Rini <trini at konsulko.com> wrote:
> >
> > On Sat, Feb 15, 2025 at 10:24:08AM -0700, Simon Glass wrote:
> > > Hi Tom,
> > >
> > > On Sat, 15 Feb 2025 at 07:46, Tom Rini <trini at konsulko.com> wrote:
> > > >
> > > > On Sat, Feb 15, 2025 at 05:06:42AM -0700, Simon Glass wrote:
> > > > > Hi Tom,
> > > > >
> > > > > On Fri, 24 Jan 2025 at 12:18, Tom Rini <trini at konsulko.com> wrote:
> > > > > >
> > > > > > On Thu, Dec 05, 2024 at 07:35:39PM -0700, Simon Glass wrote:
> > > > > >
> > > > > > > This series implements read_all() so that it is possible to read all the
> > > > > > > files relating to a bootflow, make adjustments and then boot.
> > > > > > >
> > > > > > > Unfortunately quite a few things stand in the way, so this series
> > > > > > > finishes off several pending items: zboot without CONFIG_CMDLINE,
> > > > > > > required support in pxe_utils and the differing code in the extlinux and
> > > > > > > PXE bootmeths. There is very little new code, but quite a lot of
> > > > > > > refactoring.
> > > > > > >
> > > > > > > The bootm, booti and bootz commands have all been refactored previously,
> > > > > > > so that they can operate without needing CONFIG_CMDLINE to be enabled.
> > > > > > > At the, time, zboot was left alone, since it is x86-specific and a bit
> > > > > > > more trouble.
> > > > > > >
> > > > > > > However it turns out that the booti support doesn't work with compressed
> > > > > > > booti images, so this series resolved that problem too.
> > > > > > >
> > > > > > > This series adds a programatic API for zboot and uses the forthcoming
> > > > > > > bootstd 'image list' to collect information from an extlinux file. It is
> > > > > > > therefore possible to parse the file, load the resulting images and then
> > > > > > > examine/adjust the resulting bootflow, before booting.
> > > > > > >
> > > > > > > The addition of options to extlinux resulted in the PXE and extlinux
> > > > > > > bootmeth having slightly different features. This is tidied up in this
> > > > > > > series, with common functions for both. This allows the same features
> > > > > > > (loading images as a separate step) to be provided for PXE.
> > > > > >
> > > > > > With v4 of 45/46, this fails with lwIP on Pi in the PXE tests now. You
> > > > > > likely don't need more than CONFIG_NET_LWIP=y being added but for
> > > > > > completeness here's everything:
> > > > > > CONFIG_UNIT_TEST=y
> > > > > > # CONFIG_CMD_EFIDEBUG is not set
> > > > > > CONFIG_CMD_BOOTMENU=y
> > > > > > CONFIG_CMD_LOG=y
> > > > > > # CONFIG_CMD_BOOTEFI_SELFTEST is not set
> > > > > > CONFIG_IPV6=y
> > > > > > CONFIG_IPV6_ROUTER_DISCOVERY=y
> > > > > > CONFIG_CMD_TFTPPUT=y
> > > > > > CONFIG_CMD_WGET=y
> > > > > > CONFIG_FIT=y
> > > > > > CONFIG_FIT_SIGNATURE=y
> > > > > > CONFIG_FIT_BEST_MATCH=y
> > > > > > CONFIG_SYS_BOOTM_LEN=0x4000000
> > > > > > CONFIG_NET_LWIP=y
> > > > > > CONFIG_BOOTSTAGE_STASH_ADDR=0x02400000
> > > > > > CONFIG_BOOTSTAGE=y
> > > > > > CONFIG_BOOTSTAGE_STASH=y
> > > > > > CONFIG_CMD_BOOTSTAGE=y
> > > > > > on top of rpi_3_defconfig.
> > > > >
> > > > > I just noticed this response (a bit behind on email), but I don't
> > > > > think it is reasonable to require out-of-tree changes. We just need CI
> > > > > to pass, along with any labs. This whole series seems to have been
> > > > > blocked?
> > > > >
> > > > > If you like, I would be willing to debug it if this series is applied,
> > > > > and send a follow-on patch. I have a second rpi3 in my lab which I
> > > > > could add to gitlab, along with the changes above, so we can catch
> > > > > this sort of thing in future. It would have to be added with some
> > > > > documentation, so others can figure out what is going on. Just to be
> > > > > clear, I have no issue with LWIP, just with this series being blocked
> > > > > due to out-of-tree changes.
> > > >
> > > > I don't know what "out of tree" changes you're referring to. I guess you
> > > > mean the test I noted above? It's only out of tree because you didn't
> > > > follow up with Heinrich's buildman patch to allow using fragments. If
> > > > that was applied then it would be worthwhile to have the Pi targets in
> > > > your lab do more tests, including the above. And please note this is my
> > > > lab that fails.
> > >
> > > I suggested a file format to specify combinations that we test. You
> > > wanted them to be created as separate boards. If you don't want to
> > > reconsider my proposal[1] you could create a separate in-tree rpi3
> > > board with these options. That would actually be easier for my lab,
> > > although with Heinrich's patch it wouldn't matter much, as you say.
> > >
> > > Heinrich's patch[2] is fine but needs a test. Are you asking me to write it?
> >
> > Yes, I much prefer what Heinrich did. As he's apparently been too busy
> > to address your feedback, it would be helpful if you finished it.
> 
> Yes I can do that - so are you OK to apply this first?

No, I am not. It (a) doesn't apply to master (or more appropriately now,
next) and (b) breaks my lab. I'm actually now a little confused about
why qemu_arm64_lwip doesn't break in CI too. I bet it's because while we
do:
    - ln -s conf.qemu_arm64_na /tmp/uboot-test-hooks/bin/travis-ci/conf.qemu_arm64_lwip_na
we don't have a similar link for the boardenv problem (so yes, your
patch to make it say something would be good, but also, ugh, that
https://source.denx.de/u-boot/u-boot/-/jobs/1025936 shows 6 skipped,
nothing passed means we didn't *look* for anything more than green light
/ red light).

So I'm going to go and fix things so qemu_arm64_lwip runs tests for
real, and CI would fail on this series. And Michal, is there a reason we
don't enable network tests in the boardenv files for xilinx? That's
using lwIP now too, but since most of the tests don't run, it would also
not fail there due to being skipped.  You can use the qemu boardenv
files as examples of how to automatically get something available for
tftp.

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 659 bytes
Desc: not available
URL: <https://lists.denx.de/pipermail/u-boot/attachments/20250216/c1b07b1e/attachment-0001.sig>