Skip to content

Conversation

@marcinszkudlinski
Copy link
Contributor

… init data"

This reverts commit 4849a71.

The commit has been merged with CI test fail
The test is now failing constantly making CI unusabe

Either CI test or the code requires changes before merging

FW logs:

[ 0.122168] init: print_version_banner: FW ABI 0x301b000 DBG ABI 0x5003000 tags SOF:v2.5-stable-branch-733-g8683d2d265d1 zephyr:zephyr-v3.4.0-496-g431108d89e17 src hash 0x4b68d789 (ref hash 0x4b68d789)
[ 0.122183] clock: clock_set_freq: clock 0 set freq 400000000Hz freq_idx 2
[ 0.122190] clock: clock_set_freq: clock 1 set freq 400000000Hz freq_idx 2
[ 0.122196] clock: clock_set_freq: clock 2 set freq 400000000Hz freq_idx 2
*** Booting Zephyr OS build zephyr-v3.4.0-496-g431108d89e17 ***
[ 0.122378] main: main: SOF on intel_adsp_ace15_mtpm
[ 0.122386] main: main: SOF initialized
[ 0.157836] ipc: ipc_cmd: rx : 0x43000000|0x30701000
[ 0.160618] ipc: ipc_cmd: rx : 0x43000000|0x30801000
[ 0.171581] ipc: ipc_cmd: rx : 0x44000000|0x3070000c
[ 0.171590] basefw: basefw_set_large_config: returning success for Set FW_CONFIG without handling it
[ 0.174833] ipc: ipc_cmd: rx : 0x44000000|0x3070000c
[ 0.174846] basefw: basefw_set_large_config: returning success for Set FW_CONFIG without handling it
[ 0.179795] ipc: ipc_cmd: rx : 0x44000000|0x31400008
[ 0.181086] ipc: ipc_cmd: rx : 0x44000000|0x30600064
[ 0.241513] ipc: ipc_cmd: rx : 0x11000006|0x0
[ 0.241518] pipe: pipeline_new: pipeline new pipe_id 0 priority 0
[ 0.242786] ipc: ipc_cmd: rx : 0x40000004|0x15
[ 0.242833] dma: dma_get: dma_get() ID 0 sref = 1 busy channels 0
[ 0.271751] ipc: ipc_cmd: rx : 0x40000008|0x1000000b
[ 0.271781] src: src_init: comp:0 0x80000 src_init(): Missing or bad size (44) init data
[ 0.271786] module_adapter: module_init: comp:0 0x80000 module_init() error -22: module specific init failed, comp id 524288
[ 0.271790] module_adapter: module_adapter_new: comp:0 0x80000 module_adapter_new() -22: module initialization failed
[ 0.271801] ipc: ipc4_init_module_instance: error: failed to init module 8 : 0
[ 0.271811] ipc: ipc_cmd: ipc4: MODULE_MSG failed with err 104
[ 0.272753] ipc: ipc_cmd: rx : 0x46000004|0x8
[ 0.272763] ipc: ipc_comp_disconnect: failed to find src 40000, or dst 80000
[ 0.272766] ipc: ipc_cmd: ipc4: MODULE_MSG failed with err 9

… init data"

This reverts commit 4849a71.

The commit has been merged with CI test fail
The test is now failing constantly making CI unusabe

Either CI test or the code requires changes before merging

Signed-off-by: Marcin Szkudlinski <marcin.szkudlinski@intel.com>
@mszleszy
Copy link

I've checked SRC module tests, init instance ipc messages payload looks proper.

@marcinszkudlinski marcinszkudlinski marked this pull request as ready for review July 26, 2023 09:29
@marcinszkudlinski marcinszkudlinski requested review from cujomalainey and lgirdwood and removed request for abonislawski and singalsu July 26, 2023 09:29
@singalsu
Copy link
Collaborator

singalsu commented Jul 26, 2023

Which of the tests causes the fail?

For IPC4 the config should be this:

struct ipc4_config_src {
struct ipc4_base_module_cfg base;
uint32_t sink_rate;
};

Though it's strange that this is defined in locally in src.c module. All IPC headers should be in src/include/ipc4.

@marcinszkudlinski
Copy link
Contributor Author

@singalsu it does not matter now, CI is down and we DO NEED IT WORKING - and we need it now
We'll look for a root case later. As I said - I'm not sure if the code or the test should be fixed.
Pls talk to @mszleszy

@singalsu
Copy link
Collaborator

@singalsu it does not matter now, CI is down and we DO NEED IT WORKING - and we need it now We'll look for a root case later. As I said - I'm not sure if the code or the test should be fixed. Pls talk to @mszleszy

Not disagreeing, good that we know there's an issue somewhere. This patch was supposed to fix an issue found by fuzzing.

@singalsu
Copy link
Collaborator

Also confirming that this same issue happens with Linux driver and topology (#7547).

@lgirdwood lgirdwood merged commit 55d8f41 into thesofproject:main Jul 26, 2023
@lgirdwood
Copy link
Member

@mszleszy @singalsu I'm guessing we dont have a test for this in CI today ?

@cujomalainey
Copy link
Contributor

@marcinszkudlinski thanks for root causing, I am guessing the results in the original PR were so stale that this bug slipped in because it looks fine in #7830

@cujomalainey
Copy link
Contributor

@jsarha FYI we will need to resubmit the fuzzer fix

@singalsu
Copy link
Collaborator

@jsarha FYI we will need to resubmit the fuzzer fix

The check for dev->ipc_config.type != SOF_COMP_SRC can't be done in IPC4 because the component is SOF_COMP_NONE. Can just omitting that check avoid the fuzzer found issues?

@lgirdwood
Copy link
Member

@singalsu we can have an ifdef here for ipc version if needed. Once fuzzer gets full IPC4 support we can revisist if needed.

@cujomalainey
Copy link
Contributor

@jsarha FYI we will need to resubmit the fuzzer fix

The check for dev->ipc_config.type != SOF_COMP_SRC can't be done in IPC4 because the component is SOF_COMP_NONE. Can just omitting that check avoid the fuzzer found issues?

The fuzzer issue is a security issue, it should not be ignored. @lgirdwood is right that we can just ifdef it for ipc3

@jsarha
Copy link
Contributor

jsarha commented Aug 1, 2023

@jsarha FYI we will need to resubmit the fuzzer fix

The check for dev->ipc_config.type != SOF_COMP_SRC can't be done in IPC4 because the component is SOF_COMP_NONE. Can just omitting that check avoid the fuzzer found issues?

The fuzzer issue is a security issue, it should not be ignored. @lgirdwood is right that we can just ifdef it for ipc3

I am now back from vacation. If no one has not done this yet, I can do this tomorrow.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

8 participants