Repository navigation
[BUG] 2 Sue Creek regressions #1575
Description
SOF support for Sue Creek doesn't have a high priority, but so far it is the only platform, that can be used for testing of the Cascading Interrupt Controller patch series #1053 .
Two regressions have been identified in the current SOF master branch (as of 0dd6c23) on Sue Creek:
-
PR memory: new memory layout #1483 broke bootloader startup on Sue Creek. The easiest fix might be reverting those changes for Sue Creek only, if possible.
-
With reverted memory: new memory layout #1483 booting still fails sporadically. In these cases the SPI communication with the host stops after the initial 4 IPC messages. This results in the following output on the host:
[ 538.231905] sof-audio sof-audio: ipc rx: 0x70000000: FW_READY
[ 538.231930] sof-audio sof-audio: Firmware info: version 0:0-0 build 291 on Jun 18 2019:15:05:39
[ 538.231940] sof-audio sof-audio: booting DSP firmware completed
[ 538.231958] sof-audio sof-audio: ipc rx done: 0x70000000: FW_READY
[ 538.231975] sof-audio sof-audio: firmware boot complete
[ 538.232405] sof-audio sof-audio: generating page table for 50e73d5d size 0x10000 pages 16
[ 538.232421] sof-audio sof-audio: dma_trace_pages: 16
[ 538.232474] sof-audio sof-audio: stream_tag: 0
[ 538.232516] sof-audio sof-audio: ipc tx: 0x90010000: GLB_TRACE_MSG
[ 538.540225] sof-audio sof-audio: error: ipc timed out for 0x90010000 size 0x28
[ 538.540253] sof-audio sof-audio: error: can't set params for DMA for trace -110
whereas the normal case looks as follows:
[ 61.684563] sof-audio sof-audio: ipc rx: 0x70000000: FW_READY
[ 61.684584] sof-audio sof-audio: Firmware info: version 0:0-0 build 291 on Jun 18 2019:15:05:39
[ 61.684594] sof-audio sof-audio: booting DSP firmware completed
[ 61.684615] sof-audio sof-audio: ipc rx done: 0x70000000: FW_READY
[ 61.684637] sof-audio sof-audio: firmware boot complete
[ 61.685315] sof-audio sof-audio: generating page table for (ptrval) size 0x10000 pages 16
[ 61.685331] sof-audio sof-audio: dma_trace_pages: 16
[ 61.685383] sof-audio sof-audio: stream_tag: 0
[ 61.685427] sof-audio sof-audio: ipc tx: 0x90010000: GLB_TRACE_MSG
[ 61.685690] sof-audio sof-audio: ipc rx: 0x10000000: GLB_REPLY
[ 61.685701] sof-audio sof-audio: error: ipc reply unknown
[ 61.685712] sof-audio sof-audio: ipc rx done: 0x10000000: GLB_REPLY
[ 61.685735] sof-audio sof-audio: ipc tx succeeded: 0x90010000: GLB_TRACE_MSG
[ 61.686046] sof-audio sof-audio: created machine sof-nocodec
Activity
lgirdwood commentedon Jun 18, 2019
@lyakh this should be 2 issues with descriptive titles. Have you tried booting on qemu ?
abonislawski commentedon Jun 18, 2019
As I already commented in this PR: there is obvious bug for suecreek and it is already fixed in PR #1567
We cannot test suecreek platform, that's why we didnt spotted it before
lgirdwood commentedon Jun 20, 2019
@xiulipan I guess qemu CI did not detect this issue ? I would have expected boot failure.
xiulipan commentedon Jun 20, 2019
@lgirdwood
We did not enable QEMU with Suecreek yet, is the QEMU support it now?
abonislawski commentedon Jun 20, 2019
@xiulipan @lgirdwood it would be great to enable boot test for sue creek if possible
lyakh commentedon Jun 21, 2019
@abonislawski no, #1567 doesn't fix it. I'm surprised we don't have Sue Creek in our CI, @xiulipan hasn't it been once included there?
jajanusz commentedon Jun 21, 2019
@lyakh There is building in CI, but no boot test. I guess that's because sue boot process is different than on other platforms. So there is more work to be done.
abonislawski commentedon Jun 21, 2019
@lyakh yes, #1567 fix boot problem because boot stack was placed at address "0". If there is another bug then we need a platform to test it.
lgirdwood commentedon Jun 23, 2019
@xiulipan yes we have sue creek qemu support, it should boot.
xiulipan commentedon Jun 24, 2019
@lgirdwood
Let me try and add support into CI. And I think I need to test with latest QEMU also.
lyakh commentedon Jun 24, 2019
@lgirdwood would you have a pointer to up-to-date qemu instructions?
lgirdwood commentedon Jun 24, 2019
@lyakh I guess qemu is not building for you ?
https://alsa-project.org/wiki/Firmware#Using_the_Qemu_DSP_emulator
There are differences with dependencies compared with the old instruction on alsa wiki on newer distros, can you try configuring with
Please also add the latest build instruction to sof-docs.
lyakh commentedon Jun 25, 2019
@lgirdwood sorry, I just never tried using qemu for a Sue Creek emulation. Would that only run the boot simulation, since I don't see support for an SPI connection between the host and the xtensa?
lgirdwood commentedon Jun 25, 2019
@lyakh yes, that's correct there is no SPI simulation yet (although I have not checked if DW SPI is used by other qemu targets). Boot test is all we are doing, so FW will write to SPI regs (and wont crash).
lyakh commentedon Jun 25, 2019
@abonislawski I just retested - the current top of "master" branch with this your PR applied still doesn't boot. It might well fix one issue with the memory-layout commit on Sue Creek, but there are definitely more problems. I would prefer a complete fix over a patch that improves the situation theoretically, but still cannot be successfully tested.
tlauda commentedon Mar 12, 2020
@xiulipan Is there any progress with adding sue to qemu boot test?
slawblauciak commentedon Sep 15, 2020
@lgirdwood Sue Creek support is not complete in SOF as a whole and there are no use cases/commitment. Do you think we could close this for the time being if not outright remove all relevant code for Sue Creek?
lgirdwood commentedon Sep 15, 2020
@slawblauciak @lyakh Luckily enough Sue creek is supported by Zephyr so we can remove the IRQ code.
Lets keep the SUE target and platform stuff atm though as this may be needed to complete the Zephyr integration. We can remove this when we know it's no longer needed (probably after v1.6 and after Zephyr 2.5 releases)..
mengdonglin commentedon Sep 15, 2020
@xiulipan Can you help disable Sue Creek building in Travis CI?
marc-hb commentedon Jun 15, 2021
@lyakh do you still care about this?
lyakh commentedon Jun 16, 2021
I still have two Sue Creek boards! :-D But frankly - their support has never been completed, it's just sitting there bit-rotting by the minute. So I'd be grateful to anyone submitting a clean PR to completely but cleanly wipe out Sue Creek support from SOF.
github-actions commentedon Jun 20, 2025
This issue has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this issue will automatically be closed in 14 days. Note, that you can always re-open a closed issue at any time.