Re: [PATCH v7 7/8] media: qcom: camss: Account for C-PHY when calculating link frequency

On 10/06/2026 10:23, Wenmeng Liu wrote:
> 
> 
> On 6/5/2026 9:14 PM, David Heidelberg via B4 Relay wrote:
>> From: David Heidelberg <[email protected]>
>>
>> Ensure that the link frequency divider correctly accounts for C-PHY
>> operation. The divider differs between D-PHY and C-PHY, as described
>> in the MIPI CSI-2 specification.
>>
>> For more details, see:
>> https://docs.kernel.org/driver-api/media/tx-rx.html#pixel-rate
>>
>> Suggested-by: Sakari Ailus <[email protected]>
>> Acked-by: Cory Keitz <[email protected]>
>> Tested-by: Cory Keitz <[email protected]>
>> Reviewed-by: Bryan O'Donoghue <[email protected]>
>> Link: https://docs.kernel.org/driver-api/media/tx-rx.html#pixel-rate
>> Signed-off-by: David Heidelberg <[email protected]>
>> ---
>>   drivers/media/platform/qcom/camss/camss-csid.c   |  7 +++++--
>>   drivers/media/platform/qcom/camss/camss-csiphy.c |  6 ++----
>>   drivers/media/platform/qcom/camss/camss.c        | 18 +++++++++++++++---
>>   drivers/media/platform/qcom/camss/camss.h        |  2 +-
>>   4 files changed, 23 insertions(+), 10 deletions(-)
>>
>> diff --git a/drivers/media/platform/qcom/camss/camss-csid.c b/drivers/media/ 
>> platform/qcom/camss/camss-csid.c
>> index bcc34ac9dd212..7415e811082da 100644
>> --- a/drivers/media/platform/qcom/camss/camss-csid.c
>> +++ b/drivers/media/platform/qcom/camss/camss-csid.c
>> @@ -534,25 +534,28 @@ const struct csid_format_info *csid_get_fmt_entry(const 
>> struct csid_format_info
>>   /*
>>    * csid_set_clock_rates - Calculate and set clock rates on CSID module
>>    * @csiphy: CSID device
>>    */
>>   static int csid_set_clock_rates(struct csid_device *csid)
>>   {
>>       struct device *dev = csid->camss->dev;
>> +    struct csiphy_device *csiphy = &csid->camss->csiphy[csid->phy.csiphy_id];
> 
> The PHY indices are not necessarily contiguous on some platforms, such as Hamoa 
> or Purwa.
> 
> For example, on Hamoa, we have csiphy0, csiphy1, csiphy2, and csiphy4. Using 
> csiphy->id directly would lead to an array out-of-bounds access.
> 

Nice catch, thanks. I thought I'll make it cleaner, but I guess this chance 
could be pain point. Reverted to passing (number of) `lanes` and just added cphy 
bool instead of passing whole struct.

David

> Thanks,
> Wenmeng
>> +    struct csiphy_lanes_cfg *lane_cfg = &csiphy->cfg.csi2->lane_cfg;
>>       const struct csid_format_info *fmt;
>> +
>>       s64 link_freq;
>>       int i, j;
>>       int ret;
> 

-- 
David Heidelberg