search

thesofproject / qemu

Official QEMU mirror. Please see http://wiki.qemu.org/Contribute/SubmitAPatch for how to submit changes to QEMU. Pull Requests are ignored. Please only use release tarballs from the QEMU website. SOF - Please use sof-stable branch for SOF development.
http://www.qemu.org
Other
4 stars 14 forks source link

adsp: common: Use new firmware header file #25

Closed dianacretu closed 4 years ago

dianacretu commented 4 years ago

After commit 3d3e6f8cdb7f136 ("rimage: Refine sof fw blk type") struct snd_sof_fw_blk_type has been changed to reflect the of type of memory instead of memory section type.

Use the newly created fw.h file from SOF and remove file_format.h which basically is the same file but is not up to date.

Signed-off-by: Diana Cretu diana.cretu@nxp.com

lgirdwood commented 4 years ago

@dianacretu thanks, btw we are moving to the latest qemu-4.2 release soon. You should be able to apply on to of the sof-v4.2 @xiulipan btw, we should validat and move to sof-v4.2 branch soon for CI

xiulipan commented 4 years ago

@lgirdwood Will try to test with sof-v4.2 branch and I will send the PR for the QEMU Dockerfile update.

lgirdwood commented 4 years ago

@xiulipan thanks, it will also need this update from @dianacretu merged too.

lgirdwood commented 4 years ago

@dianacretu @xiulipan now that we are running on top of latest qemu we are in a position to start upstreaming DSP support. i.e. upstreaming the xtensa ISA target configurations an machine support for each DSP would be a good starting place.

dbaluta commented 4 years ago

@lgirdwood that's great! Question about SST module. Is it even used?

@xiulipan did the tests worked fine with this change? We had some concerns about breaking Intel platforms :).

lgirdwood commented 4 years ago

@dbaluta I guess you mean SSP module ? I think we need to work on some sort of EASI, HDA, SSP etc simulated drivers that can sit on top of a generic File I/O i.e. the simulated DAI driver only does what is needed to get audio running whilst PCM audio data is read from/to a file on the host FS.

dbaluta commented 4 years ago

@lgirdwood

Firstly:

I am thinking about SST as in sst_module_memcpy function in hw/adsp/dsp/common.c. Not really sure what is the difference between SST and SSP. I know what SSP is.

Secondly, s/EASI/ESAI (Enhanced Serial Audio Interface)

Indeed, I still need to get a grasp of how this should work. File based approach looks promising. And could greatly help us with debugging.

lgirdwood commented 4 years ago

SST stuff is the FW binary file format used by the older Intel DSPs like BYT and BDW.