Uh oh!
There was an error while loading. Please reload this page.
LLEXT: automatically calculate module addresses (first part of #9430) - #9439
Conversation
lyakh
commented
Sep 5, 2024
CI: IIUC https://sof-ci.01.org/sofpr/PR9439/build7624/devicetest/index.html unrelated, https://sof-ci.01.org/sofpr/PR9439/build7626/devicetest/index.html - one device missing |
lyakh
commented
Sep 5, 2024
Since we decided to split #9430 and since this passes tests, let's merge it |
lgirdwood
commented
Sep 9, 2024
@lyakh can you give it a quick fixup for the conflict. Thanks |
Add a Kconfig option that can be used to flip all LLEXT-supporting code to a modular build. Use that global flag for smart-amp-test, mixin-mixout and eq-iir. Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Currently LLEXT module starting addresses are hard-coded in their respective CMakeLists.txt files. This is very wasteful, since it's unknown in what order modules are loaded, inflexible and not easily extendible to other platforms. Switch to calculating addresses automatically based on a single per-platform Kconfig value. Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
LNL modules overlay is missing eq_iir and mixin_mixout entries, add them there. Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
lyakh
commented
Sep 9, 2024
pr-build seems stuck, restarting |
lyakh
commented
Sep 9, 2024
SOFCI TEST |
lyakh
commented
Sep 9, 2024
CI: unrelated LNL failures https://sof-ci.01.org/sofpr/PR9439/build7780/devicetest/index.html |
abonislawski
commented
Sep 11, 2024
Are we now unable to set module address to IMR for one module and HPSRAM for another? |
lyakh
commented
Sep 11, 2024
@abonislawski with this - no, with this module addresses are calculated automatically. Running in IMR needs to be added separately |
The important change here is an automatic module address calculation, simplifies modularisation a lot.