From: Pascal Fontain <pascal.fontain@bachmann.info>

The sram subsystem might use the dma-buf subsystem which only gets
initialized in subsys_initcall phase. Because of this we move the sram
initcall into the device phase.

Signed-off-by: Pascal Fontain <pascal.fontain@bachmann.info>
---
 drivers/misc/sram.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/misc/sram.c b/drivers/misc/sram.c
index e248c0a8882f..632f90d9bcea 100644
--- a/drivers/misc/sram.c
+++ b/drivers/misc/sram.c
@@ -461,4 +461,4 @@ static int __init sram_init(void)
 	return platform_driver_register(&sram_driver);
 }
 
-postcore_initcall(sram_init);
+device_initcall(sram_init);
-- 
2.39.2

On Tue, Apr 9, 2024, at 14:06, Pascal FONTAIN wrote:
> From: Pascal Fontain <pascal.fontain@bachmann.info>
>
> The sram subsystem might use the dma-buf subsystem which only gets
> initialized in subsys_initcall phase. Because of this we move the sram
> initcall into the device phase.
>
> Signed-off-by: Pascal Fontain <pascal.fontain@bachmann.info>

This does not look safe unless you check all users of the sram
driver to ensure they are not needed early and they can handle
deferred probing.

      Arnd