The crash occurs for the same reason.
In this case, on the really_probe() failure path, we are still
the DMA ops prior to releasing the device's managed memories.
This patch fixes this issue by reordering the DMA ops teardown
call to devres_release_all() on the failure path.
Reported-by: Xiang Chen <chenxiang66@xxxxxxxxxxxxx>
Tested-by: Xiang Chen <chenxiang66@xxxxxxxxxxxxx>
Signed-off-by: John Garry <john.garry@xxxxxxxxxx>
So does this "fix" 376991db4b64? If so, should this be added to the
patch and also backported to the stable trees?
No, I don't think so. I'd say it supplements it. Here I'm trying to fix
up another path in which we tear down the DMA ops prior to releasing
I didn't add a fixes tag as 376991db4b64 didn't have one either. It
need to be backported to stable, I figure the same as 376991db4b64.
I still think that we should consider merging this patch.
Bah, sorry, I thought I'd replied to this thread already, but apparently
I don't have a particularly strong opinion, but at the moment I am
leaning slightly towards this being a parallel fix for another part of
09515ef5ddad, rather than a specific fix of the other fix. Either way,
you can have a:
Reviewed-by: Robin Murphy <robin.murphy@xxxxxxx>