Message-ID: <50fb2dd6-4e8f-a550-6eda-073beb86f2ff@xilinx.com>
Date:   Wed, 17 Jun 2020 09:56:59 +0200
From:   Michal Simek <michal.simek@...inx.com>
To:     Nathan Chancellor <natechancellor@...il.com>,
        Michal Simek <michal.simek@...inx.com>
Cc:     Christophe Leroy <christophe.leroy@...roup.eu>,
        Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Paul Mackerras <paulus@...ba.org>,
        Michael Ellerman <mpe@...erman.id.au>, arnd@...db.de,
        linuxppc-dev@...ts.ozlabs.org, linux-kernel@...r.kernel.org,
        clang-built-linux@...glegroups.com
Subject: Re: [PATCH v5 01/13] powerpc: Remove Xilinx PPC405/PPC440 support

Hi Nathan,


On 16. 06. 20 20:16, Nathan Chancellor wrote:
> Hi Michal,
> 
> On Tue, Jun 16, 2020 at 04:45:20PM +0200, Michal Simek wrote:
>>
>>
>> On 16. 06. 20 2:27, Nathan Chancellor wrote:
>>> On Thu, May 21, 2020 at 04:55:52PM +0000, Christophe Leroy wrote:
>>>> From: Michal Simek <michal.simek@...inx.com>
>>>>
>>>> The latest Xilinx design tools called ISE and EDK has been released in
>>>> October 2013. New tool doesn't support any PPC405/PPC440 new designs.
>>>> These platforms are no longer supported and tested.
>>>>
>>>> PowerPC 405/440 port is orphan from 2013 by
>>>> commit cdeb89943bfc ("MAINTAINERS: Fix incorrect status tag") and
>>>> commit 19624236cce1 ("MAINTAINERS: Update Grant's email address and maintainership")
>>>> that's why it is time to remove the support fot these platforms.
>>>>
>>>> Signed-off-by: Michal Simek <michal.simek@...inx.com>
>>>> Acked-by: Arnd Bergmann <arnd@...db.de>
>>>> Signed-off-by: Christophe Leroy <christophe.leroy@...roup.eu>
>>>
>>> This patch causes qemu-system-ppc to fail to load ppc44x_defconfig:
>>>
>>> $ make -skj"$(nproc)" ARCH=powerpc CROSS_COMPILE=powerpc-linux- O=out/ppc distclean ppc44x_defconfig zImage
>>>
>>> $ timeout --foreground 30s unbuffer \
>>> qemu-system-ppc \
>>> -machine bamboo \
>>
>> Did you bisect it that you found that this patch is causing problem for
>> you on any bamboo machine?
>>
>> Or this was caused by the whole series?
>>
>> Thanks,
>> Michal
> 
> Yes, this conclusion was the result of the following bisect:
> 
> $ cat test.sh
> #!/usr/bin/env bash
> 
> cd "${HOME}"/src/linux || exit 125
> 
> set -x
> 
> PATH=${HOME}/toolchains/gcc/10.1.0/bin:${PATH} \
> make -skj"$(nproc)" ARCH=powerpc CROSS_COMPILE=powerpc-linux- O=out/ppc32 distclean ppc44x_defconfig zImage || exit 125
> 
> "${HOME}"/cbl/github/boot-utils/boot-qemu.sh -a ppc32 -k out/ppc32 -t 30s
> 
> $ git bisect start v5.8-rc1 v5.7
> ...
> 
> $ git bisect run test.sh
> ...
> 
> $ git bisect log
> # bad: [b3a9e3b9622ae10064826dccb4f7a52bd88c7407] Linux 5.8-rc1
> # good: [3d77e6a8804abcc0504c904bd6e5cdf3a5cf8162] Linux 5.7
> git bisect start 'v5.8-rc1' 'v5.7'
> # good: [ee01c4d72adffb7d424535adf630f2955748fa8b] Merge branch 'akpm' (patches from Andrew)
> git bisect good ee01c4d72adffb7d424535adf630f2955748fa8b
> # bad: [6f2dc3d335457d9c815be9f4fd3dc8eff92fcef7] Merge tag 'dma-mapping-5.8-2' of git://git.infradead.org/users/hch/dma-mapping
> git bisect bad 6f2dc3d335457d9c815be9f4fd3dc8eff92fcef7
> # skip: [828f3e18e1cb98c68fc6db4d5113513d4a267775] Merge tag 'arm-drivers-5.8' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
> git bisect skip 828f3e18e1cb98c68fc6db4d5113513d4a267775
> # good: [c46241a370a61f0f264791abb9fc869016e749ce] powerpc/pkeys: Check vma before returning key fault error to the user
> git bisect good c46241a370a61f0f264791abb9fc869016e749ce
> # good: [3f0be4df50a7854a831c80a74d7cf2cfd61f2fde] Merge tag 'versatile-dts-v5.8-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-integrator into arm/dt
> git bisect good 3f0be4df50a7854a831c80a74d7cf2cfd61f2fde
> # bad: [bd55e792de0844631d34487d43eaf3f13294ebfe] powerpc/module_64: Use special stub for _mcount() with -mprofile-kernel
> git bisect bad bd55e792de0844631d34487d43eaf3f13294ebfe
> # good: [303e6a9ddcdc168e92253c78cdb4bbe1e10d78b3] powerpc/watchpoint: Convert thread_struct->hw_brk to an array
> git bisect good 303e6a9ddcdc168e92253c78cdb4bbe1e10d78b3
> # good: [0755e85570a4615ca674ad6489d44d63916f1f3e] powerpc/xive: Do not expose a debugfs file when XIVE is disabled
> git bisect good 0755e85570a4615ca674ad6489d44d63916f1f3e
> # bad: [b4ac18eead28611ff470d0f47a35c4e0ac080d9c] powerpc/perf/hv-24x7: Fix inconsistent output values incase multiple hv-24x7 events run
> git bisect bad b4ac18eead28611ff470d0f47a35c4e0ac080d9c
> # bad: [3aacaa719b7bf135551cabde2480e8f7bfdf7c7d] powerpc/40x: Don't save CR in SPRN_SPRG_SCRATCH6
> git bisect bad 3aacaa719b7bf135551cabde2480e8f7bfdf7c7d
> # bad: [1b5c0967ab8aa9424cdd5108de4e055d8aeaa9d0] powerpc/40x: Remove support for IBM 403GCX
> git bisect bad 1b5c0967ab8aa9424cdd5108de4e055d8aeaa9d0
> # good: [0bdad33d6bd7b80722e2f9e588d3d7c6d6e34978] powerpc/64: Refactor interrupt exit irq disabling sequence
> git bisect good 0bdad33d6bd7b80722e2f9e588d3d7c6d6e34978
> # bad: [2c74e2586bb96012ffc05f1c819b05d9cad86d6e] powerpc/40x: Rework 40x PTE access and TLB miss
> git bisect bad 2c74e2586bb96012ffc05f1c819b05d9cad86d6e
> # bad: [7ade8495dcfd788a76e6877c9ea86f5207369ea4] powerpc: Remove Xilinx PPC405/PPC440 support
> git bisect bad 7ade8495dcfd788a76e6877c9ea86f5207369ea4
> # first bad commit: [7ade8495dcfd788a76e6877c9ea86f5207369ea4] powerpc: Remove Xilinx PPC405/PPC440 support
> 

I have took a look at it and was able to run qemu and also saw your
issue. What happened is that when xilinx platforms were removed zImage
is generated but it is not u-boot legacy image.
Don't know details about zImage/uImage in ppc world but if you dump
zImage you should see this.

Before this patch:
[qemu](master)$ hexdump -C /mnt/disk/linux/arch/powerpc/boot/zImage |
head -n 20
00000000  27 05 19 56 bc 4d 17 ce  5e e9 cb 0f 00 2c 8d 98
|'..V.M..^....,..|
00000010  00 00 00 00 00 00 00 00  69 04 b6 eb 05 07 02 01
|........i.......|
00000020  4c 69 6e 75 78 2d 35 2e  37 2e 30 2d 72 63 32 2d
|Linux-5.7.0-rc2-|
00000030  30 30 32 34 39 2d 67 32  38 31 65 65 36 66 36 38
|00249-g281ee6f68|
00000040  1f 8b 08 00 00 00 00 00  02 03 ec 5a 7b 70 54 65
|...........Z{pTe|
00000050  96 3f b7 bb d3 34 49 30  8d b6 1a 16 d4 ce 43 4c
|.?...4I0......CL|
00000060  42 84 b8 cb ec de 9e 3c  b8 21 c0 dc 24 ac 26 63
|B......<.!..$.&c|
00000070  32 04 74 9c ce 36 28 49  b4 16 04 6a db f5 5e fa
|2.t..6(I...j..^.|
00000080  36 dd b8 49 00 ab 33 60  15 28 8f 46 42 17 a0 56
|6..I..3`.(.FB..V|
00000090  65 5c 58 99 42 a0 2d 20  15 02 6c 31 2b ce 04 70  |e\X.B.-
..l1+..p|


After this patch:
[qemu](master)$ hexdump -C /mnt/disk/linux/arch/powerpc/boot/zImage |
head -n 20
00000000  00 52 50 4f 00 60 00 00  00 00 17 73 00 00 00 00
|.RPO.`.....s....|
00000010  00 60 00 70 a8 22 53 af  00 00 00 00 00 00 00 00
|.`.p."S.........|
00000020  00 01 7e a4 94 21 ff f0  3c 80 08 00 7c 08 02 a6
|..~..!..<...|...|
00000030  42 9f 00 05 38 a0 00 20  38 c0 00 40 93 c1 00 08  |B...8..
8..@....|
00000040  90 01 00 14 7f c8 02 a6  80 1e ff ec 7f c0 f2 14
|................|
00000050  80 7e 80 00 7c 83 20 50  48 00 1b 8d 80 01 00 14  |.~..|.
PH.......|


It means only Xilinx platforms have been asking for uImage format and
bamboo doesn't require it. It also looks like that qemu expect uImage
format.

You should know what format qemu expects.
Anyway if you build it by make uImage and then pass it to qemu you
should boot just fine.

Or if bamboo requires uImage to be built by default you can do it via
Kconfig.

diff --git a/arch/powerpc/platforms/44x/Kconfig
b/arch/powerpc/platforms/44x/Kconfig
index 39e93d23fb38..300864d7b8c9 100644
--- a/arch/powerpc/platforms/44x/Kconfig
+++ b/arch/powerpc/platforms/44x/Kconfig
@@ -13,6 +13,7 @@ config BAMBOO
        select PPC44x_SIMPLE
        select 440EP
        select FORCE_PCI
+       select DEFAULT_UIMAGE
        help
          This option enables support for the IBM PPC440EP evaluation board.

Hope that this help with resolving your issue.

Thanks,
Michal

Powered by blists - more mailing lists