Stephen,
On Wed, Jan 16, 2013 at 3:42 PM, Stephen Warren swarren@wwwdotorg.org wrote:
On 01/16/2013 02:14 PM, Tom Warren wrote:
This provides SPL support for T114 boards - AVP early init, plus CPU (A15) init/jump to main U-Boot.
diff --git a/arch/arm/cpu/arm720t/tegra-common/cpu.c b/arch/arm/cpu/arm720t/tegra-common/cpu.c
if (chip_id == 0x30)
if (chip_id >= 0x30) return TEGRA_FAMILY_T3x; else return TEGRA_FAMILY_T2x;Shouldn't that comparison use CHIPID_TEGRA30?
Shouldn't there be a TEGRA_FAMILY_T11x for Tegra114; anything that cares about >=Tegra30 can compare the family with >= not ==.
Hmm. It seems the only use of the FAMILY value is get_num_cpus() right below; why not have that just switch on the same chip_id variable that everything else uses?
Sure, I can refine this code. I didn't want to spend too much time obsessively polishing my 'precious' and miss the upstreaming dates I'd committed to.
But it's worth a look for V2.
diff --git a/arch/arm/cpu/arm720t/tegra-common/cpu.h b/arch/arm/cpu/arm720t/tegra-common/cpu.h
+#ifndef TRUE +#define TRUE 1 +#endif +#ifndef FALSE +#define FALSE 0 +#endif
Surely those are in a standard header somewhere; we shouldn't create yet another duplicate of them.
Couldn't find 'em on a quick search (grep define.TRUE) except in places like scsi.h and ext4_common.h and fpga.h. If you have a standard header that you know of, let me know.
diff --git a/arch/arm/cpu/arm720t/tegra114/cpu.c b/arch/arm/cpu/arm720t/tegra114/cpu.c
+static int is_partition_powered(u32 mask)
reg = readl(&pmc->pmc_pwrgate_status);if ((reg & mask) == mask)return TRUE;return FALSE;The last 4 lines are just "return reg & mask;" or "return (reg & mask) == mask;". Same in the next function.
I'm porting internal bootloader bringup code here, and trying to avoid unnecessary changes, but I can modify these to remove the TRUE/FALSE in V2.
+void powerup_cpus(void) +{
debug("powerup_cpus entry\n");/* Are we booting to the fast cluster? */if (get_cluster_id() == 0) {Why would we ever boot on anything other than the fast cluster? I would assume that the kernel assume it will always get booted on the fast cluster, which would then imply that U-Boot had to boot on or switch to the fast cluster.
Again, this is from internal NV bootloader code that I know works. I don't know the circumstances where we might be booted on the LP cluster, but I figured if the internal bootloader code thought it worth checking, so should I. If you have unimpeachable evidence to the contrary, I can optimize this out.
Tom