On 08/16/2012 01:01 AM, Stefano Babic wrote:
On 15/08/2012 23:30, Tom Rini wrote:
We can only attempt to setup a malloc pool if CONFIG_SYS_SPL_MALLOC_START is defined, and not all boards require it. Make the call depend on the define.
Signed-off-by: Tom Rini trini@ti.com
arch/arm/lib/spl.c | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/arch/arm/lib/spl.c b/arch/arm/lib/spl.c index 4d33f99..71a467e 100644 --- a/arch/arm/lib/spl.c +++ b/arch/arm/lib/spl.c @@ -149,8 +149,10 @@ void board_init_r(gd_t *id, ulong dummy) u32 boot_device; debug(">>spl:board_init_r()\n");
+#ifdef CONFIG_SYS_SPL_MALLOC_START mem_malloc_init(CONFIG_SYS_SPL_MALLOC_START, CONFIG_SYS_SPL_MALLOC_SIZE); +#endif
#ifdef CONFIG_SPL_BOARD_INIT spl_board_init();
Agree, mem_malloc_init should be not always called. However, how are the dependencies with other SPL configuration ? I think (but I can be wrong) that we need it if CONFIG_SPL_LIBDISK_SUPPORT or CONFIG_SPL_FAT_SUPPORT are set. Should we add them to your #ifdef case ?
I think until we have Kconfig-style dependencies available, there's going to be a certain amount of pain here. We should probably add to docs/README.SPL that SPL_FAT_SUPPORT depends on the malloc pool and the defines that calls for (since there's a few of them not referenced right were we call mem_malloc_init).