[Sound-open-firmware] [PATCH] arch: xtensa: set SRAM window error codes during bootloader

Mon Jun 11 17:41:49 CEST 2018


On 6/11/2018 22:17, sound-open-firmware-bounces at alsa-project.org wrote:
>> -----Original Message-----
>> From: sound-open-firmware-bounces at alsa-project.org [mailto:sound-open-
>> firmware-bounces at alsa-project.org] On Behalf Of Pan, Xiuli
>> Sent: Monday, June 11, 2018 5:52 PM
>> To: Keyon Jie <yang.jie at linux.intel.com>; sound-open-firmware at alsa-
>> project.org
>> Subject: Re: [Sound-open-firmware] [PATCH] arch: xtensa: set SRAM window
>> error codes during bootloader
>>
>>
>>
>> On 6/11/2018 16:38, Keyon Jie wrote:
>>>
>>> On 2018年06月11日 16:29, Pan, Xiuli wrote:
>>>>
>>>> On 6/11/2018 15:58, Lauda, Tomasz wrote:
>>>>> FW sets MW0 to different address than ROM We are setting status 0x05
>>>>> from bootloader, so it's the correct assumption, that ROM jumped out
>>>>> to FW correctly.
>>>> So could we disable the MW0 setting in the boot loader? This seems to
>>>> broke the status check as we want to check the ROM code.
>>>>
>>>> And BTW, I checked the memory window data and found they are fulled
>>>> with some data that I can find in the rodata. Should this happen?
>>>> If the memory data can be some none zero value, i suggest that we add
>>>> a memset for the memory we used for memory windows
>>> We previously memset all MW datas to 0s at MW setup, except those
>>> registers reserved for ROM status, does anything of that changed?
>>>
>> I tried to add memset in platform_memory_windows_init, but there still some
>> unknown data.
>> After comparing the data, I found that the memory in the MW3 (from file
>> /sys/kernel/debug/sof/etrace) is the exactly the same data in the sof-apl.ri binary
> >from offset 0x38000 to 0x3a000.
>> I could not find any related with these two values, but there are two pages of the
>> data is exactly same. We must have something wrong with memory settings.
>> Either memory window or the memory mapping.
>> Anyone have any idea?
> Since DMA trace works, can you use DMA trace to dump this MW3 region at different FW initial stage, to see when it is changed?
> e.g. at boot loader stage, or platform initial stage, or after FW_READY?
>
DMA trace will not work after FW ready.
It will need IPC mailbox to send the offset.

Thanks
Xiuli
> Thanks,
> ~Keyon
>
>>
>> HEXDUMP for /sys/kernel/debug/sof/etrace
>>
>> 0000000 f99f ff17 7557 019e 6ccc fd7d b651 03a3
>> 0000010 7d57 fae3 c582 071f a85a f5e4 ac82 0f2e
>> 0000020 7e76 e584 5b06 513e 5b06 513e 7e76 e584
>> 0000030 ac82 0f2e a85a f5e4 c582 071f 7d57 fae3
>> 0000040 b651 03a3 6ccc fd7d 7557 019e f99f ff17
>> 0000050 bae7 0055 cf09 001e d41d ff85 e0f6 00bf
>> 0000060 2f5e ff0d 78f7 0115 e827 fed5 bf27 0132
>> 0000070 a101 fece c9e6 0127 4ac4 fee8 b78a 0102
>> 0000080 ba22 ff15 b0da 00cf dde1 ff4b 9a5b 0098
>> 0000090 0fc4 ff82 d063 0064 4196 ffb2 16d8 0039
>> 00000a0 edf4 ffd8 c7d9 0017 cc3f fff4 3993 0001
>> 00000b0 55cd 0006 4871 fff4 3257 000f ee1f ffee
>> 00000c0 a32a 0011 cef0 ffee 019f 0010 ac0c fff1
>> 00000d0 5ed0 000c b043 fff5 4ac8 0008 9548 fff9
>> 00000e0 c1aa 0004 a5fe fffc 378c 0002 a739 fffe
>> 00000f0 26e6 0000 2d62 0000 3301 ffff ccc8 0001
>> 0000100 bec0 fffc 3d0a 0005 30e5 fff8 00c7 000b
>>
>> HEXDUMP for sof-apl.ri
>>
>> 0038000 f99f ff17 7557 019e 6ccc fd7d b651 03a3
>> 0038010 7d57 fae3 c582 071f a85a f5e4 ac82 0f2e
>> 0038020 7e76 e584 5b06 513e 5b06 513e 7e76 e584
>> 0038030 ac82 0f2e a85a f5e4 c582 071f 7d57 fae3
>> 0038040 b651 03a3 6ccc fd7d 7557 019e f99f ff17
>> 0038050 bae7 0055 cf09 001e d41d ff85 e0f6 00bf
>> 0038060 2f5e ff0d 78f7 0115 e827 fed5 bf27 0132
>> 0038070 a101 fece c9e6 0127 4ac4 fee8 b78a 0102
>> 0038080 ba22 ff15 b0da 00cf dde1 ff4b 9a5b 0098
>> 0038090 0fc4 ff82 d063 0064 4196 ffb2 16d8 0039
>> 00380a0 edf4 ffd8 c7d9 0017 cc3f fff4 3993 0001
>> 00380b0 55cd 0006 4871 fff4 3257 000f ee1f ffee
>> 00380c0 a32a 0011 cef0 ffee 019f 0010 ac0c fff1
>> 00380d0 5ed0 000c b043 fff5 4ac8 0008 9548 fff9
>> 00380e0 c1aa 0004 a5fe fffc 378c 0002 a739 fffe
>> 00380f0 26e6 0000 2d62 0000 3301 ffff ccc8 0001
>> 0038100 bec0 fffc 3d0a 0005 30e5 fff8 00c7 000b
>>
>>
>>
>>> Thanks,
>>> ~Keyon
>>>
>>>> Thanks
>>>> Xiuli
>>>>> Tomek
>>>>>
>>>>>
>>>>> -----Original Message-----
>>>>> From: sound-open-firmware-bounces at alsa-project.org
>>>>> [mailto:sound-open-firmware-bounces at alsa-project.org] On Behalf Of
>>>>> Pan, Xiuli
>>>>> Sent: Monday, June 11, 2018 7:11 AM
>>>>> To: Liam Girdwood <liam.r.girdwood at linux.intel.com>;
>>>>> sound-open-firmware at alsa-project.org
>>>>> Cc: Kamil Kulesza <kamil.kulesza at linux.intel.com>
>>>>> Subject: Re: [Sound-open-firmware] [PATCH] arch: xtensa: set SRAM
>>>>> window error codes during bootloader
>>>>>
>>>>>
>>>>>
>>>>> On 6/9/2018 03:06, Liam Girdwood wrote:
>>>>>> On Fri, 2018-06-08 at 19:42 +0100, Liam Girdwood wrote:
>>>>>>> From: Kamil Kulesza <kamil.kulesza at linux.intel.com>
>>>>>>>
>>>>>>> Set status (0x05) and error (0x00) code in Memory Window 0 when
>>>>>>> the bootloader starts.
>>>>>>> boot_entry.S - set status (0x05) and error (0x00) code before wnd0
>>>>>>> reprogram platform/memory.h - increase bootloader size
>>>>>>>
>>>>>>> Signed-off-by: Kamil Kulesza <kamil.kulesza at linux.intel.com>
>>>>>>> ---
>>>>>>>     src/arch/xtensa/boot_entry.S                  | 22
>>>>>>> +++++++++++++++++++
>>>>>>>     .../apollolake/include/platform/memory.h      |  2 +-
>>>>>>>     .../cannonlake/include/platform/memory.h      |  2 +-
>>>>>>>     3 files changed, 24 insertions(+), 2 deletions(-)
>>>>>>>
>>>>>>>
>>>>> This patch seems to be a workaround for the firmware boot up. The
>>>>> status
>>>>> (0x05) should be set by the rom.
>>>>> With the rmbox and dump data from status registers. It seems all
>>>>> memory window is fulled with rodata of the firmware. I think there
>>>>> is some potential bugs for memory window.
>>>>> If this is not a bug maybe we need to add memset for memory windows.
>>>>>
>>>>> Thanks
>>>>> Xiuli
>>>>>
>>>>>> Applied.
>>>>>>
>>>>>> Thanks
>>>>>>
>>>>>> Liam
>>>>>> _______________________________________________
>>>>>> Sound-open-firmware mailing list
>>>>>> Sound-open-firmware at alsa-project.org
>>>>>> http://mailman.alsa-project.org/mailman/listinfo/sound-open-firmwar
>>>>>> e
>>>>> _______________________________________________
>>>>> Sound-open-firmware mailing list
>>>>> Sound-open-firmware at alsa-project.org
>>>>> http://mailman.alsa-project.org/mailman/listinfo/sound-open-firmware
>>>>> --------------------------------------------------------------------
>>>>>
>>>>> Intel Technology Poland sp. z o.o.
>>>>> ul. Slowackiego 173 | 80-298 Gdansk | Sad Rejonowy Gdansk Polnoc |
>>>>> VII Wydzial Gospodarczy Krajowego Rejestru Sadowego - KRS 101882 |
>>>>> NIP 957-07-52-316 | Kapital zakladowy 200.000 PLN.
>>>>>
>>>>> Ta wiadomosc wraz z zalacznikami jest przeznaczona dla okreslonego
>>>>> adresata i moze zawierac informacje poufne. W razie przypadkowego
>>>>> otrzymania tej wiadomosci, prosimy o powiadomienie nadawcy oraz
>>>>> trwale jej usuniecie; jakiekolwiek przegladanie lub
>>>>> rozpowszechnianie jest zabronione.
>>>>> This e-mail and any attachments may contain confidential material
>>>>> for the sole use of the intended recipient(s). If you are not the
>>>>> intended recipient, please contact the sender and delete all copies;
>>>>> any review or distribution by others is strictly prohibited.
>>>>>
>>>>> _______________________________________________
>>>>> Sound-open-firmware mailing list
>>>>> Sound-open-firmware at alsa-project.org
>>>>> http://mailman.alsa-project.org/mailman/listinfo/sound-open-firmware
>>>> _______________________________________________
>>>> Sound-open-firmware mailing list
>>>> Sound-open-firmware at alsa-project.org
>>>> http://mailman.alsa-project.org/mailman/listinfo/sound-open-firmware
>>>>
>>> _______________________________________________
>>> Sound-open-firmware mailing list
>>> Sound-open-firmware at alsa-project.org
>>> http://mailman.alsa-project.org/mailman/listinfo/sound-open-firmware
>> _______________________________________________
>> Sound-open-firmware mailing list
>> Sound-open-firmware at alsa-project.org
>> http://mailman.alsa-project.org/mailman/listinfo/sound-open-firmware