[PATCH u-boot-marvell 00/10] tools: kwboot: Fix boot and terminal mode

[Tony Dinh]

Hi Pali,

It seems the problem was with the ActionTec MI424WR-I board quirk. It
has nothing to do with the new kwboot on Avanta SoC. Sorry for the
false alarm!

I spent a couple hours running with various delays for the option -s,
even with 50, and it all works fine!  Uhm... So I went back to the
default run (without -s) and it also worked. All these tests were done
with a cold start each time.

So that's when I realized  it was probably a quirk with this board.
Sometimes the warm boot caused problems for the serial port. But the
new kwboot will work whenever I coldstart (power cycle).

Thanks for a new and improved kwboot!

Tested-by: Tony Dinh <mibodhi at gmail.com>

Thanks,
Tony




On Thu, Mar 3, 2022 at 4:05 PM Pali Rohár <pali at kernel.org> wrote:
>
> Hello!
>
> On Thursday 03 March 2022 15:58:41 Tony Dinh wrote:
> > Hi Pali,
> >
> > Here is the test report.
> >
> > 1. Kirkwood SoC
> >
> > Tested with Seagate GoFlex Home board (88F6281):
> > - Everything is working fine as before.
> > - kwboot -d also works to get BootROM into debug mode.
> > - At the BootROM command line, x 0x5 selected boot source NAND, and
> > started booting from u-boot in NAND successfully.
> >
> > <BEGIN LOG>
> > ./kwboot -t -B 115200 /dev/ttyUSB0 -d
> > kwboot version 2022.04-rc3-00002-gf64aac4a69-dirty
> > Sending debug message. Please reboot the target.../
> > Clearing input buffer...
> > [Type Ctrl-\ + c to quit]
> > >x 0x5
> > <END LOG>
> >
> > 2. Armada SoC
> >
> > Tested with Globalscale Technologies Mirabox (Armada 370)
> > - Everything is working fine as before
> > - kwboot -d also works to get BootROM into debug mode
> >
> > <BEGIN LOG>
> > ./kwboot -t  -B 115200 /dev/ttyUSB2 -d
> > kwboot version 2022.04-rc3-00002-gf64aac4a69-dirty
> > Sending debug message. Please reboot the target.../
> > Clearing input buffer...
> > [Type Ctrl-\ + c to quit]
> >
> > >help
> > h/? - print this help screen
> > r <0xXXXXXXXX> - read memory address 0xXXXXXXXX
> > w <0xXXXXXXXX> <0xYYYYYYYY> - write 0xYYY
> > YYYYY to address 0xXXXXXXXX
> > j <0xXXXXXXXX> - jump to address 0xXXXXXXXX
> > x <X> - select SAR boot source X and restart
> > >
> > <END LOG>
>
> Thank you for testing. Nice to see that -d is working!
>
> > 3. Dove SoC
> >
> > Tested with HP Thin Client T5335z (Dove 88AP510)
> > - Everything is working as before.
> > - No change in behavior since the last test. IOW, u-boot image was
> > transferred successfully, but the BootROM did not start u-boot, as
> > described here:
> > https://lists.denx.de/pipermail/u-boot/2022-February/475494.html
> >
> > 4. Avanta ȘoC
> >
> > Tested with the ActionTec MI424WR-I (88F6560 A0)
> >
> > - This kwboot version is broken for this board. The header and image
> > were not transferred at all.
> > - The last tested and working kwboot version was built on Feb 15th (up
> > until and including patch
> > tools-kwbimage-Support-for-parsing-extended-v0-format.patch)
>
> Ou. This needs to be fixed.
>
> Could you try adjust "-s" argument? Default value is 50. My tests showed
> that Armada 385 is better detected with value 24. And some old
> documentation says that Armada XP needs value 1000.
>
> >
> > Thanks,
> > Tony
> >
> >
> > On Wed, Mar 2, 2022 at 1:18 PM Tony Dinh <mibodhi at gmail.com> wrote:
> > >
> > > Hi Pali,
> > >
> > >
> > > On Wed, Mar 2, 2022 at 1:03 PM Pali Rohár <pali at kernel.org> wrote:
> > > >
> > > > On Wednesday 02 March 2022 12:51:25 Tony Dinh wrote:
> > > > > Hi Pali,
> > > > >
> > > > > The patch did not apply cleanly to the master branch (I downloaded the
> > > > > series from patchwork). Do you have a link where I can use it?
> > > >
> > > > Hello, this patch series depends on following kwboot patch:
> > > > https://patchwork.ozlabs.org/project/uboot/patch/20220218112413.10009-1-pali@kernel.org/
> > >
> > > That was it. I will try to kwboot a few different boxes and let you know.
> > >
> > > Thanks,
> > > Tony
> > >
> > > > > Thanks,
> > > > > Tony
> > > > >
> > > > >
> > > > >
> > > > > On Wed, Mar 2, 2022 at 2:52 AM Pali Rohár <pali at kernel.org> wrote:
> > > > > >
> > > > > > This patch series fixes sending boot and debug patterns by doing it in
> > > > > > separate thread. Entering BootROM debug mode via '-d' option is now more
> > > > > > stable and working fine on Armada 385. There is also support for backspace
> > > > > > key and updated documentation.
> > > > > >
> > > > > > Stefan and Tony, could you test this patch series on more Marvell boards
> > > > > > which you have? Specially if '-d' option is working too.
> > > > > >
> > > > > > *** BLURB HERE ***
> > > > > >
> > > > > > Pali Rohár (10):
> > > > > >   tools: kwboot: Check for return value of kwboot_tty_send() and
> > > > > >     tcflush()
> > > > > >   tools: kwboot: Remove msg_req_delay
> > > > > >   tools: kwboot: Cleanup bootmsg and debugmsg variables
> > > > > >   tools: kwboot: Use separate thread for sending boot message pattern
> > > > > >   tools: kwboot: Fix sending and processing debug message pattern (-d
> > > > > >     option)
> > > > > >   tools: kwboot: Add support for backspace key in mini terminal
> > > > > >   tools: kwboot: Update usage
> > > > > >   tools: kwboot: Update manpage
> > > > > >   tools: kwboot: Update doc about Avanta
> > > > > >   tools: kwboot: Update references with public links
> > > > > >
> > > > > >  doc/kwboot.1   | 105 ++++++++++-
> > > > > >  tools/Makefile |   3 +
> > > > > >  tools/kwboot.c | 462 ++++++++++++++++++++++++++++++++++++++++---------
> > > > > >  3 files changed, 483 insertions(+), 87 deletions(-)
> > > > > >
> > > > > > --
> > > > > > 2.20.1
> > > > > >