Re: [PATCH] coresight: Fix data argument to coresight_enable_helpers

[Jie Gan <jie.gan@oss.qualcomm.com>]

On 9/17/2025 6:44 AM, Carl Worth wrote:
> In the commit being fixed, coresight_enable_path() was changed to call
> coresight_enable_helpers() with a final argument of 'path' rather than
> 'sink_data'. This was invalid, resulting in derefencing a pointer to
> a 'struct coresight_path' as if it were a 'struct perf_output_handle'.
> 
> The compiler could not flag the error since there are several layers
> of function calls treating the pointer as void*.
> 
> Fix to correctly pass the sink_data pointer as the final argument to
> coresight_enable_helpers(), exactly as it was before the buggy commit.
> 
> Bug can be reproduced with:
> 
> $ perf record -e cs_etm//k -C 0-9 dd if=/dev/zero of=/dev/null
> 
> Showing an oops as follows:
> 
> [   88.696535] Unable to handle kernel paging request at virtual address 000f6e84934ed19e
> ...
> [   88.911293] Call trace:
> [   88.913728]  tmc_etr_get_buffer+0x30/0x80 [coresight_tmc] (P)
> [   88.919463]  catu_enable_hw+0xbc/0x3d0 [coresight_catu]
> [   88.924677]  catu_enable+0x70/0xe0 [coresight_catu]
> [   88.929542]  coresight_enable_path+0xb0/0x258 [coresight]
> 
> Fixes: 080ee83cc361 ("Coresight: Change functions to accept the coresight_path")
> Signed-off-by: Carl Worth <carl@os.amperecomputing.com>
> ---
>   drivers/hwtracing/coresight/coresight-core.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index fa758cc21827..b1077d73c988 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -510,7 +510,7 @@ int coresight_enable_path(struct coresight_path *path, enum cs_mode mode,
>   		type = csdev->type;
>   
>   		/* Enable all helpers adjacent to the path first */
> -		ret = coresight_enable_helpers(csdev, mode, path);
> +		ret = coresight_enable_helpers(csdev, mode, sink_data);

I dont think we can change back to sink_data since we introduced 
coresight_path to wrap 'data' which is needed by the path.

I suggest you to add the struct perf_output_handle to the 
coresight_path, then retrieving it with data->perf_handle in 
tmc_etr_get_buffer.

before:
struct perf_output_handle *handle = data;

after:
struct coresight_path *path = data;
struct perf_output_handle *handle = path->perf_handle;

We can assign the perf_output_handle to the coresight_path after we 
constructed the coresight_path in perf mode.


Thanks,
Jie

>   		if (ret)
>   			goto err_disable_path;
>   		/*