Skip to contentSkip to conversation

ARM: tegra: remove the ifdef of ARCH SoC in the tegra_resume

PATCH New Joseph Lo 3 participants 5 messages Jun 3, 2013 – Jun 5, 2013

Subsystems (3)
Arm (32-bit) Arm sub-architectures Tegra architecture support
lore
Share & tools
lore

Joseph Lo

author
#0 lore raw

PATCH 1/1 ARM: tegra: remove the ifdef of ARCH SoC in the tegra_resume

Removing the ifdef of ARCH_TEGRA_SoC in the tegra_resume function. Because we always build with all Tegra SoCs support and had a runtime chip detection code there. And we expect most of the chips would need the code in the future.

We also fix a typo of a macro name that cause a build error.

Reported-by Arnd Bergmann <[email protected]>
Signed-off-by Joseph Lo <[email protected]>
---
 arch/arm/mach-tegra/reset-handler.S | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
arch/arm/mach-tegra/reset-handler.S +1 -3
index d2042ac..39dc9e7 100644
@@ -54,12 +54,11 @@ ENTRY(tegra_resume)
 	bne	cpu_resume			@ no
 no_cpu0_chk:
 
-#ifndef CONFIG_ARCH_TEGRA_2x_SOC
 	/* Are we on Tegra20? */
 	cmp	r6, #TEGRA20
 	beq	1f				@ Yes
 	/* Clear the flow controller flags for this CPU. */
-	cpu_to_csr_req r1, r0
+	cpu_to_csr_reg r1, r0
 	mov32	r2, TEGRA_FLOW_CTRL_BASE
 	ldr	r1, [r2, r1]
 	/* Clear event & intr flag */

@@ -70,7 +69,6 @@ no_cpu0_chk:
 	bic	r1, r1, r0
 	str	r1, [r2]
 1:
-#endif
 
 	check_cpu_part_num 0xc09, r8, r9
 	bne	not_ca9
Signature · 1 line
1.8.3
#1 lore raw

PATCH 1/1 ARM: tegra: remove the ifdef of ARCH SoC in the tegra_resume

Quoted 6 lines · Joseph Lo

Removing the ifdef of ARCH_TEGRA_SoC in the tegra_resume function. Because we always build with all Tegra SoCs support and had a runtime chip detection code there. And we expect most of the chips would need the code in the future.

We also fix a typo of a macro name that cause a build error.

OK, this fixes the build issue, but this patch raises some questions about the code...

arch/arm/mach-tegra/reset-handler.S ENTRY(tegra_resume) · line 59

Quoted 6 lines
arch/arm/mach-tegra/reset-handler.S +0 -1
 
-#ifndef CONFIG_ARCH_TEGRA_2x_SOC
 	/* Are we on Tegra20? */
 	cmp	r6, #TEGRA20
 	beq	1f				@ Yes

Was that ifdef completely incorrect before? I can see why the cmp/beq might be ifdef'd (although it's not worth it), but I assume the code after that beq was intended to run on all chips after Tegra20. The ifdef as it was written does something rather different; it prevents any of that code from running unless the kernel doesn't have Tegra20 support. So, I think that the removal of the ifdef is more of a bug-fix that "because we always build with all Tegra SoCs support. Let me know, and I'll re-write the commit description to something more accurate...

Quoted 14 lines
arch/arm/mach-tegra/reset-handler.S +1 -2
 	/* Clear the flow controller flags for this CPU. */
-	cpu_to_csr_req r1, r0
+	cpu_to_csr_reg r1, r0
 	mov32	r2, TEGRA_FLOW_CTRL_BASE
 	ldr	r1, [r2, r1]
 	/* Clear event & intr flag */

@@ -70,7 +69,6 @@ no_cpu0_chk:
 	bic	r1, r1, r0
 	str	r1, [r2]
 1:
-#endif
 
 	check_cpu_part_num 0xc09, r8, r9
 	bne	not_ca9
#2 lore raw

PATCH 1/1 ARM: tegra: remove the ifdef of ARCH SoC in the tegra_resume

Quoted 8 lines · Stephen Warren

Was that ifdef completely incorrect before? I can see why the cmp/beq might be ifdef'd (although it's not worth it), but I assume the code after that beq was intended to run on all chips after Tegra20. The ifdef as it was written does something rather different; it prevents any of that code from running unless the kernel doesn't have Tegra20 support. So, I think that the removal of the ifdef is more of a bug-fix that "because we always build with all Tegra SoCs support. Let me know, and I'll re-write the commit description to something more accurate...

The #ifdef was recently changed from #if TEGRA3 to #if !TEGRA2, which support for TEGRA4 was added. I think this is where it broke.

Arnd

author
#3 lore raw

PATCH 1/1 ARM: tegra: remove the ifdef of ARCH SoC in the tegra_resume

Quoted 10 lines · Arnd Bergmann
Quoted 8 lines · Stephen Warren

Was that ifdef completely incorrect before? I can see why the cmp/beq might be ifdef'd (although it's not worth it), but I assume the code after that beq was intended to run on all chips after Tegra20. The ifdef as it was written does something rather different; it prevents any of that code from running unless the kernel doesn't have Tegra20 support. So, I think that the removal of the ifdef is more of a bug-fix that "because we always build with all Tegra SoCs support. Let me know, and I'll re-write the commit description to something more accurate...

Quoted 3 lines

The #ifdef was recently changed from #if TEGRA3 to #if !TEGRA2, which support for TEGRA4 was added. I think this is where it broke.

Yes, exactly. Sorry again for the error code.

Joseph

Applied
#4 lore raw

PATCH 1/1 ARM: tegra: remove the ifdef of ARCH SoC in the tegra_resume

In reply to Joseph Lo
Quoted 6 lines · Joseph Lo

Removing the ifdef of ARCH_TEGRA_SoC in the tegra_resume function. Because we always build with all Tegra SoCs support and had a runtime chip detection code there. And we expect most of the chips would need the code in the future.

We also fix a typo of a macro name that cause a build error.

Applied to Tegra's for-3.11/soc, with re-written commit description.