[U-Boot] [PATCH] rpi: increase SYS_BOOTM_LEN to 64M
On AArch64, kernel images are not self-decompressing and easily exceed the 8MB limit.
Signed-off-by: Laurent Bonnans laurent.bonnans@here.com --- include/configs/rpi.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/include/configs/rpi.h b/include/configs/rpi.h index 9ce41767a9..1a5ae26ae7 100644 --- a/include/configs/rpi.h +++ b/include/configs/rpi.h @@ -55,6 +55,7 @@ #define CONFIG_SYS_MEMTEST_START 0x00100000 #define CONFIG_SYS_MEMTEST_END 0x00200000 #define CONFIG_LOADADDR 0x00200000 +#define CONFIG_SYS_BOOTM_LEN SZ_64M
/* Devices */ /* GPIO */
Hi Laurent,
On 07/06/2019 17:04, Bonnans, Laurent wrote:
On AArch64, kernel images are not self-decompressing and easily exceed the 8MB limit.
Signed-off-by: Laurent Bonnans laurent.bonnans@here.com
include/configs/rpi.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/include/configs/rpi.h b/include/configs/rpi.h index 9ce41767a9..1a5ae26ae7 100644 --- a/include/configs/rpi.h +++ b/include/configs/rpi.h @@ -55,6 +55,7 @@ #define CONFIG_SYS_MEMTEST_START 0x00100000 #define CONFIG_SYS_MEMTEST_END 0x00200000 #define CONFIG_LOADADDR 0x00200000 +#define CONFIG_SYS_BOOTM_LEN SZ_64M
Sorry for the late reply. If it's only needed for ARM64 then we should make it set for this architecture, right?
Next time you send a patch, plese add the maintainers mailinglist. This way I'll find the patch easier.
Thanks, Matthias
On 7/31/19 9:12 AM, Matthias Brugger wrote:
Hi Laurent,
On 07/06/2019 17:04, Bonnans, Laurent wrote:
On AArch64, kernel images are not self-decompressing and easily exceed the 8MB limit.
Signed-off-by: Laurent Bonnans laurent.bonnans@here.com
include/configs/rpi.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/include/configs/rpi.h b/include/configs/rpi.h index 9ce41767a9..1a5ae26ae7 100644 --- a/include/configs/rpi.h +++ b/include/configs/rpi.h @@ -55,6 +55,7 @@ #define CONFIG_SYS_MEMTEST_START 0x00100000 #define CONFIG_SYS_MEMTEST_END 0x00200000 #define CONFIG_LOADADDR 0x00200000 +#define CONFIG_SYS_BOOTM_LEN SZ_64M
Sorry for the late reply. If it's only needed for ARM64 then we should make it set for this architecture, right?
That seems to be the best call indeed. Then I wonder if we should just set the default larger if CONFIG_ARM64 because the problem should arise almost in all case, I believe. There is a similar define in sunxi-common.h for example.
But anyway, that's a larger debate, I'll send another patch with the ARM64 check in the header file.
Next time you send a patch, plese add the maintainers mailinglist. This way I'll find the patch easier.
Yes, sorry about that. Only realized it later...
Laurent
participants (2)
-
Bonnans, Laurent -
Matthias Brugger