linux-omap.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] ARM: OMAP5: select SCU
@ 2013-05-14 16:57 Vincent Stehlé
  2013-05-15 12:36 ` Santosh Shilimkar
  0 siblings, 1 reply; 6+ messages in thread
From: Vincent Stehlé @ 2013-05-14 16:57 UTC (permalink / raw)
  To: Tony Lindgren; +Cc: linux-omap, linux-kernel, Vincent Stehlé

From: Vincent Stehlé <v-stehle@ti.com>

OMAP5 needs SCU in SMP.

This fixes the following link errors:

  arch/arm/mach-omap2/built-in.o: In function `scu_gp_set':
  arch/arm/mach-omap2/sleep44xx.S:132: undefined reference to `scu_power_mode'
  arch/arm/mach-omap2/built-in.o: In function `scu_gp_clear':
  arch/arm/mach-omap2/sleep44xx.S:229: undefined reference to `scu_power_mode'
  arch/arm/mach-omap2/built-in.o: In function `omap4_smp_prepare_cpus':
  arch/arm/mach-omap2/omap-smp.c:211: undefined reference to `scu_enable'
  arch/arm/mach-omap2/built-in.o: In function `omap4_smp_init_cpus':
  arch/arm/mach-omap2/omap-smp.c:185: undefined reference to `scu_get_core_count'

Signed-off-by: Vincent Stehlé <v-stehle@ti.com>
---
 arch/arm/mach-omap2/Kconfig |    1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/mach-omap2/Kconfig b/arch/arm/mach-omap2/Kconfig
index f49cd51..81690a2 100644
--- a/arch/arm/mach-omap2/Kconfig
+++ b/arch/arm/mach-omap2/Kconfig
@@ -109,6 +109,7 @@ config SOC_OMAP5
 	select ARM_CPU_SUSPEND if PM
 	select ARM_GIC
 	select CPU_V7
+	select HAVE_ARM_SCU if SMP
 	select HAVE_SMP
 	select COMMON_CLK
 	select HAVE_ARM_ARCH_TIMER
-- 
1.7.10.4

^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH] ARM: OMAP5: select SCU
  2013-05-14 16:57 [PATCH] ARM: OMAP5: select SCU Vincent Stehlé
@ 2013-05-15 12:36 ` Santosh Shilimkar
  2013-05-16 13:10   ` Vincent Stehlé
  2013-05-16 16:00   ` [RFC PATCH v2] ARM: OMAP4+: protect SCU code compilation Vincent Stehlé
  0 siblings, 2 replies; 6+ messages in thread
From: Santosh Shilimkar @ 2013-05-15 12:36 UTC (permalink / raw)
  To: Vincent Stehlé
  Cc: Tony Lindgren, linux-omap, linux-kernel, Vincent Stehlé

On Tuesday 14 May 2013 10:27 PM, Vincent Stehlé wrote:
> From: Vincent Stehlé <v-stehle@ti.com>
> 
> OMAP5 needs SCU in SMP.
> 
> This fixes the following link errors:
> 
>   arch/arm/mach-omap2/built-in.o: In function `scu_gp_set':
>   arch/arm/mach-omap2/sleep44xx.S:132: undefined reference to `scu_power_mode'
>   arch/arm/mach-omap2/built-in.o: In function `scu_gp_clear':
>   arch/arm/mach-omap2/sleep44xx.S:229: undefined reference to `scu_power_mode'
>   arch/arm/mach-omap2/built-in.o: In function `omap4_smp_prepare_cpus':
>   arch/arm/mach-omap2/omap-smp.c:211: undefined reference to `scu_enable'
>   arch/arm/mach-omap2/built-in.o: In function `omap4_smp_init_cpus':
>   arch/arm/mach-omap2/omap-smp.c:185: undefined reference to `scu_get_core_count'
> 
> Signed-off-by: Vincent Stehlé <v-stehle@ti.com>
> ---
OMAP5 doesn't need the SCU. So Nak for the patch. I haven't noticed
this earlier. Do you know which commit broke this ?

Regards,
Santosh

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] ARM: OMAP5: select SCU
  2013-05-15 12:36 ` Santosh Shilimkar
@ 2013-05-16 13:10   ` Vincent Stehlé
  2013-05-16 14:29     ` Vincent Stehlé
  2013-05-16 16:00   ` [RFC PATCH v2] ARM: OMAP4+: protect SCU code compilation Vincent Stehlé
  1 sibling, 1 reply; 6+ messages in thread
From: Vincent Stehlé @ 2013-05-16 13:10 UTC (permalink / raw)
  To: Santosh Shilimkar
  Cc: Vincent Stehlé, Tony Lindgren, linux-omap, linux-kernel

On 05/15/2013 02:36 PM, Santosh Shilimkar wrote:
> OMAP5 doesn't need the SCU. So Nak for the patch.

Hi,

I agree that it is a pure link issue.

> I haven't noticed
> this earlier. Do you know which commit broke this ?

After some bisecting I arrived at this merge commit:

  5497e7b Merge branch 'late/fixes' into fixes

Indeed, if I revert it (-m 1) the result does link properly. I'll try to
isolate it further.

Best regards,

V.

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] ARM: OMAP5: select SCU
  2013-05-16 13:10   ` Vincent Stehlé
@ 2013-05-16 14:29     ` Vincent Stehlé
  2013-05-17  7:16       ` Shilimkar, Santosh
  0 siblings, 1 reply; 6+ messages in thread
From: Vincent Stehlé @ 2013-05-16 14:29 UTC (permalink / raw)
  To: Santosh Shilimkar
  Cc: Vincent Stehlé, Tony Lindgren, linux-omap, linux-kernel

So the exact commit breaking OMAP5 link for me is:

  883a106 ARM: default machine descriptor for multiplatform

The breakage seems to be a side effect of not selecting ARCH_VEXPRESS
any more, which causes HAVE_ARM_SCU to not be selected any more, too,
when compiling only for OMAP5. IMHO this only reveals a deeper issue.

Probably the ideal fix would be to remove all references to the scu_
functions for OMAP5, but those seem to exist in files and functions,
common to OMAP4 and 5 (sleep44xx.S, omap-smp.c).

How would you deal with that, please?

Best regards,

V.

^ permalink raw reply	[flat|nested] 6+ messages in thread

* [RFC PATCH v2] ARM: OMAP4+: protect SCU code compilation
  2013-05-15 12:36 ` Santosh Shilimkar
  2013-05-16 13:10   ` Vincent Stehlé
@ 2013-05-16 16:00   ` Vincent Stehlé
  1 sibling, 0 replies; 6+ messages in thread
From: Vincent Stehlé @ 2013-05-16 16:00 UTC (permalink / raw)
  To: Santosh Shilimkar; +Cc: Tony Lindgren, linux-omap, linux-kernel, v-stehle

From: Vincent Stehlé <v-stehle@ti.com>

This fixes the following link error when compiling only for OMAP5:

  arch/arm/mach-omap2/built-in.o: In function `scu_gp_set':
  arch/arm/mach-omap2/sleep44xx.S:132: undefined reference to `scu_power_mode'
  arch/arm/mach-omap2/built-in.o: In function `scu_gp_clear':
  arch/arm/mach-omap2/sleep44xx.S:229: undefined reference to `scu_power_mode'
  arch/arm/mach-omap2/built-in.o: In function `omap4_smp_prepare_cpus':
  arch/arm/mach-omap2/omap-smp.c:211: undefined reference to `scu_enable'
  arch/arm/mach-omap2/built-in.o: In function `omap4_smp_init_cpus':
  arch/arm/mach-omap2/omap-smp.c:185: undefined reference to `scu_get_core_count'

Signed-off-by: Vincent Stehlé <v-stehle@ti.com>
---


Hi Santosh,

Here is a try at "gating" all SCU references from OMAP4/5 common code. That
does build and run for me on OMAP4 and 5. On OMAP4, System.map files are even
identical with and without the patch.

How would you like this, please?

Best regards,

V.


 arch/arm/mach-omap2/omap-smp.c  |   10 ++++++++++
 arch/arm/mach-omap2/sleep44xx.S |   10 ++++++++++
 2 files changed, 20 insertions(+)

diff --git a/arch/arm/mach-omap2/omap-smp.c b/arch/arm/mach-omap2/omap-smp.c
index 2a551f9..45c8b5d 100644
--- a/arch/arm/mach-omap2/omap-smp.c
+++ b/arch/arm/mach-omap2/omap-smp.c
@@ -21,7 +21,9 @@
 #include <linux/io.h>
 #include <linux/irqchip/arm-gic.h>
 
+#ifdef CONFIG_HAVE_ARM_SCU
 #include <asm/smp_scu.h>
+#endif
 
 #include "omap-secure.h"
 #include "omap-wakeupgen.h"
@@ -41,15 +43,19 @@
 
 u16 pm44xx_errata;
 
+#ifdef CONFIG_HAVE_ARM_SCU
 /* SCU base address */
 static void __iomem *scu_base;
+#endif
 
 static DEFINE_SPINLOCK(boot_lock);
 
+#ifdef CONFIG_HAVE_ARM_SCU
 void __iomem *omap4_get_scu_base(void)
 {
 	return scu_base;
 }
+#endif
 
 static void __cpuinit omap4_secondary_init(unsigned int cpu)
 {
@@ -176,6 +182,7 @@ static void __init omap4_smp_init_cpus(void)
 	/* Use ARM cpuid check here, as SoC detection will not work so early */
 	cpu_id = read_cpuid_id() & CPU_MASK;
 	if (cpu_id == CPU_CORTEX_A9) {
+#ifdef CONFIG_HAVE_ARM_SCU
 		/*
 		 * Currently we can't call ioremap here because
 		 * SoC detection won't work until after init_early.
@@ -183,6 +190,7 @@ static void __init omap4_smp_init_cpus(void)
 		scu_base =  OMAP2_L4_IO_ADDRESS(scu_a9_get_base());
 		BUG_ON(!scu_base);
 		ncores = scu_get_core_count(scu_base);
+#endif
 	} else if (cpu_id == CPU_CORTEX_A15) {
 		ncores = OMAP5_CORE_COUNT;
 	}
@@ -203,12 +211,14 @@ static void __init omap4_smp_prepare_cpus(unsigned int max_cpus)
 	void *startup_addr = omap_secondary_startup;
 	void __iomem *base = omap_get_wakeupgen_base();
 
+#ifdef CONFIG_HAVE_ARM_SCU
 	/*
 	 * Initialise the SCU and wake up the secondary core using
 	 * wakeup_secondary().
 	 */
 	if (scu_base)
 		scu_enable(scu_base);
+#endif
 
 	if (cpu_is_omap446x()) {
 		startup_addr = omap_secondary_startup_4460;
diff --git a/arch/arm/mach-omap2/sleep44xx.S b/arch/arm/mach-omap2/sleep44xx.S
index 88ff83a..3e73a32 100644
--- a/arch/arm/mach-omap2/sleep44xx.S
+++ b/arch/arm/mach-omap2/sleep44xx.S
@@ -10,7 +10,11 @@
  */
 
 #include <linux/linkage.h>
+
+#ifdef CONFIG_HAVE_ARM_SCU
 #include <asm/smp_scu.h>
+#endif
+
 #include <asm/memory.h>
 #include <asm/hardware/cache-l2x0.h>
 
@@ -70,6 +74,7 @@ ENTRY(omap4_finish_suspend)
 	 * Flush all data from the L1 data cache before disabling
 	 * SCTLR.C bit.
 	 */
+#ifdef CONFIG_HAVE_ARM_SCU
 	bl	omap4_get_sar_ram_base
 	ldr	r9, [r0, #OMAP_TYPE_OFFSET]
 	cmp	r9, #0x1			@ Check for HS device
@@ -81,6 +86,7 @@ ENTRY(omap4_finish_suspend)
 	DO_SMC
 	ldmfd   r13!, {r4-r12, r14}
 skip_secure_l1_clean:
+#endif
 	bl	v7_flush_dcache_all
 
 	/*
@@ -108,6 +114,7 @@ skip_secure_l1_clean:
 	 * preventing the CPU from receiving cache, TLB, or BTB
 	 * maintenance operations broadcast by other CPUs in the cluster.
 	 */
+#ifdef CONFIG_HAVE_ARM_SCU
 	bl	omap4_get_sar_ram_base
 	mov	r8, r0
 	ldr	r9, [r8, #OMAP_TYPE_OFFSET]
@@ -131,6 +138,7 @@ scu_gp_set:
 	bl	omap4_get_scu_base
 	bl	scu_power_mode
 skip_scu_gp_set:
+#endif
 	mrc	p15, 0, r0, c1, c1, 2		@ Read NSACR data
 	tst	r0, #(1 << 18)
 	mrcne	p15, 0, r0, c1, c0, 1
@@ -211,6 +219,7 @@ do_WFI:
 	orreq	r0, r0, #(1 << 6)
 	mcreq	p15, 0, r0, c1, c0, 1
 	isb
+#ifdef CONFIG_HAVE_ARM_SCU
 	bl	omap4_get_sar_ram_base
 	mov	r8, r0
 	ldr	r9, [r8, #OMAP_TYPE_OFFSET]
@@ -228,6 +237,7 @@ scu_gp_clear:
 	mov	r1, #SCU_PM_NORMAL
 	bl	scu_power_mode
 skip_scu_gp_clear:
+#endif
 	isb
 	dsb
 	ldmfd	sp!, {r4-r12, pc}
-- 
1.7.10.4

--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply related	[flat|nested] 6+ messages in thread

* RE: [PATCH] ARM: OMAP5: select SCU
  2013-05-16 14:29     ` Vincent Stehlé
@ 2013-05-17  7:16       ` Shilimkar, Santosh
  0 siblings, 0 replies; 6+ messages in thread
From: Shilimkar, Santosh @ 2013-05-17  7:16 UTC (permalink / raw)
  To: Stehle, Vincent
  Cc: Vincent Stehlé, Tony Lindgren, linux-omap@vger.kernel.org,
	linux-kernel@vger.kernel.org

Sorry for top posting. Can you just add static inlines functions in header file and #ifdef it for OMAP5.

Regards,
Santosh
________________________________________
From: Stehle, Vincent
Sent: Thursday, May 16, 2013 7:59 PM
To: Shilimkar, Santosh
Cc: Vincent Stehlé; Tony Lindgren; linux-omap@vger.kernel.org; linux-kernel@vger.kernel.org
Subject: Re: [PATCH] ARM: OMAP5: select SCU

So the exact commit breaking OMAP5 link for me is:

  883a106 ARM: default machine descriptor for multiplatform

The breakage seems to be a side effect of not selecting ARCH_VEXPRESS
any more, which causes HAVE_ARM_SCU to not be selected any more, too,
when compiling only for OMAP5. IMHO this only reveals a deeper issue.

Probably the ideal fix would be to remove all references to the scu_
functions for OMAP5, but those seem to exist in files and functions,
common to OMAP4 and 5 (sleep44xx.S, omap-smp.c).

How would you deal with that, please?

Best regards,

V.

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2013-05-17  7:16 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-05-14 16:57 [PATCH] ARM: OMAP5: select SCU Vincent Stehlé
2013-05-15 12:36 ` Santosh Shilimkar
2013-05-16 13:10   ` Vincent Stehlé
2013-05-16 14:29     ` Vincent Stehlé
2013-05-17  7:16       ` Shilimkar, Santosh
2013-05-16 16:00   ` [RFC PATCH v2] ARM: OMAP4+: protect SCU code compilation Vincent Stehlé

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).