Hi Conor,
On 26/11/24 00:01, Conor Dooley wrote:
On Mon, Nov 25, 2024 at 02:09:14PM +0530, Vaishnav Achath wrote:
J722S/AM67 uses the same BCDMA CSIRX IP as AM62A, but it supports
TX channels as well in addition to RX.
This doesn't make sense. You say that the am62a doesn't have a tx
channel ("but it supports TX as well") but then modify the struct for
the am62a to add a tx channel. Does that not break things on the am62a?
Thank you for the review, I have sent a v2 of this series adding new
compatible as suggested, after looking at it again, the J722S BCDMA CSI
is more similar to J721S2 in terms of having RX and TX support, so
updated in that way.
The below changes did not really break AM62A since the driver checks
hardware capability registers (TCHAN_CNT) to detect presence of TX
channels and then only use the Output Event Steering(OES) data below.
V2:
https://lore.kernel.org/all/20241126125158.37744-1-vaishnav.a@xxxxxx/
Thanks and Regards,
Vaishnav
Add the BCDMA TCHAN information
in the am62a_dmss_csi_soc_data so as to support all the platforms in the
family with same compatible. UDMA_CAP2_TCHAN_CNT indicates the presence
of TX channels and it will be 0 for platforms without TX support.
Signed-off-by: Vaishnav Achath <vaishnav.a@xxxxxx>
---
CSI2RX capture test results on J722S EVM with IMX219:
https://gist.github.com/vaishnavachath/e2eaed62ee8f53428ee9b830aaa02cc3
drivers/dma/ti/k3-udma.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/dma/ti/k3-udma.c b/drivers/dma/ti/k3-udma.c
index b3f27b3f9209..4130f50979d4 100644
--- a/drivers/dma/ti/k3-udma.c
+++ b/drivers/dma/ti/k3-udma.c
@@ -4340,6 +4340,8 @@ static struct udma_match_data j721e_mcu_data = {
static struct udma_soc_data am62a_dmss_csi_soc_data = {
.oes = {
+ .bcdma_tchan_data = 0x800,
+ .bcdma_tchan_ring = 0xa00,
.bcdma_rchan_data = 0xe00,
.bcdma_rchan_ring = 0x1000,
},
--
2.34.1
