Hello Liu, On Thu Feb 5, 2026 at 10:26 AM CET, Liu Ying wrote: > On Thu, Feb 05, 2026 at 09:52:04AM +0100, Luca Ceresoli wrote: >> Hello Liu, > > Hello Luca, > >> >> On Wed Feb 4, 2026 at 7:27 AM CET, Liu Ying wrote: >>> Hi Luca, >>> >>> On Tue, Feb 03, 2026 at 11:35:25AM +0100, Luca Ceresoli wrote: >>>> This driver obtains a bridge pointer from of_drm_find_bridge() in the probe >>>> function and stores it until driver removal. of_drm_find_bridge() is >>>> deprecated. Move to of_drm_find_and_get_bridge() for the bridge to be >>>> refcounted and use bridge->next_bridge to put the reference on >>>> deallocation. >>>> >>>> To keep the code as simple and reliable as possible, get a reference for >>>> each pointer that stores a drm_bridge address when it is stored and release >>>> it when the pointer is overwritten or goes out of scope. Also remove the >>>> intermediate selected_bridge variable to reduce the refcounted variables in >>>> the function. The involved pointers are: >>>> >>>> * next_bridge loop-local variable: >>>> - get reference by of_drm_find_and_get_bridge() >>>> - put reference at the end of the loop iteration (__free) >>>> >>>> * pl->bridge.next_bridge, tied to struct imx8qxp_pixel_link lifetime: >>>> - get reference when assigned (by copy from next_bridge) >>>> - put reference before reassignment if reassignment happens >>>> - put reference when the struct imx8qxp_pixel_link embedding the >>>> struct drm_bridge is destroyed (struct drm_bridge::next_bridge) >>>> >>>> Additionally, split the somewhat complex if() for readability. >>>> >>>> Signed-off-by: Luca Ceresoli <[email protected]> >>>> >>>> --- >>>> >>>> Changes in v5: >>>> - rewrite commit message after Liu's review to clarify the per-pointer >>>> get/put idea >>>> - split the if()s involved in selcting the bridge >>>> - remove intermediate selected_bridge pointer >>>> - removed Maxime's R-by, patch changed >>>> --- >>>> drivers/gpu/drm/bridge/imx/imx8qxp-pixel-link.c | 17 ++++++++++------- >>>> 1 file changed, 10 insertions(+), 7 deletions(-) >>>> >>>> diff --git a/drivers/gpu/drm/bridge/imx/imx8qxp-pixel-link.c >>>> b/drivers/gpu/drm/bridge/imx/imx8qxp-pixel-link.c >>>> index 91e4f4d55469..e29e099b893a 100644 >>>> --- a/drivers/gpu/drm/bridge/imx/imx8qxp-pixel-link.c >>>> +++ b/drivers/gpu/drm/bridge/imx/imx8qxp-pixel-link.c >>>> @@ -23,7 +23,6 @@ >>>> >>>> struct imx8qxp_pixel_link { >>>> struct drm_bridge bridge; >>>> - struct drm_bridge *next_bridge; >>>> struct device *dev; >>>> struct imx_sc_ipc *ipc_handle; >>>> u8 stream_id; >>>> @@ -140,7 +139,7 @@ static int imx8qxp_pixel_link_bridge_attach(struct >>>> drm_bridge *bridge, >>>> } >>>> >>>> return drm_bridge_attach(encoder, >>>> - pl->next_bridge, bridge, >>>> + pl->bridge.next_bridge, bridge, >>>> DRM_BRIDGE_ATTACH_NO_CONNECTOR); >>>> } >>>> >>>> @@ -260,7 +259,6 @@ static int imx8qxp_pixel_link_find_next_bridge(struct >>>> imx8qxp_pixel_link *pl) >>>> { >>>> struct device_node *np = pl->dev->of_node; >>>> struct device_node *port; >>>> - struct drm_bridge *selected_bridge = NULL; >>>> u32 port_id; >>>> bool found_port = false; >>>> int reg; >>>> @@ -297,7 +295,8 @@ static int imx8qxp_pixel_link_find_next_bridge(struct >>>> imx8qxp_pixel_link *pl) >>>> continue; >>>> } >>>> >>>> - struct drm_bridge *next_bridge = of_drm_find_bridge(remote); >>>> + struct drm_bridge *next_bridge __free(drm_bridge_put) = >>>> + of_drm_find_and_get_bridge(remote); >>>> if (!next_bridge) >>>> return -EPROBE_DEFER; >>>> >>>> @@ -305,12 +304,16 @@ static int >>>> imx8qxp_pixel_link_find_next_bridge(struct imx8qxp_pixel_link *pl) >>>> * Select the next bridge with companion PXL2DPI if >>>> * present, otherwise default to the first bridge >>>> */ >>>> - if (!selected_bridge || of_property_present(remote, >>>> "fsl,companion-pxl2dpi")) >>>> - selected_bridge = next_bridge; >>>> + if (!pl->bridge.next_bridge) >>>> + pl->bridge.next_bridge = drm_bridge_get(next_bridge); >>>> + >>>> + if (of_property_present(remote, "fsl,companion-pxl2dpi")) { >>>> + drm_bridge_put(pl->bridge.next_bridge); >>>> + pl->bridge.next_bridge = drm_bridge_get(next_bridge); >>>> + } >>> >>> Can you drop the intermediate next_bridge variable to simplify the code? >>> >>> -8<- >>> if (!pl->bridge.next_bridge) { >>> pl->bridge.next_bridge = of_drm_find_and_get_bridge(remote); >>> if (!pl->bridge.next_bridge) >>> return -EPROBE_DEFER; >>> } >>> >>> if (of_property_present(remote, "fsl,companion-pxl2dpi")) { >>> drm_bridge_put(pl->bridge.next_bridge); >>> pl->bridge.next_bridge = of_drm_find_and_get_bridge(remote); >>> if (!pl->bridge.next_bridge) >>> return -EPROBE_DEFER; >>> } >>> -8<- >> >> Potentially calling of_drm_find_and_get_bridge() twice on the same node, >> with a put in the middle, looks poorly readable to me, even though it still >> looks correct code. >> >> However I think we can do even better with an 'else if': >> >> if (!pl->bridge.next_bridge) { >> pl->bridge.next_bridge = of_drm_find_and_get_bridge(remote); >> if (!pl->bridge.next_bridge) >> return -EPROBE_DEFER; >> } else if (of_property_present(remote, "fsl,companion-pxl2dpi")) { <=== >> drm_bridge_put(pl->bridge.next_bridge); >> pl->bridge.next_bridge = of_drm_find_and_get_bridge(remote); >> if (!pl->bridge.next_bridge) >> return -EPROBE_DEFER; >> } >> >> Looks OK? > > Both are fine to me. TBH, I feel my version with two 'if's is a bit easier > to read. But, I'd say up to you.
I think this is really a minor detail and there is no obvious "best" version, so I'll send v6 as I had it ready and build-tested already, i.e. with the 'else if' version. Thanks for the discussion! Luca -- Luca Ceresoli, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
