Skip to contentSkip to conversation

MIPS: Alchemy: sleepcode without compile-time cputype dependencies

PATCH RESEND In review Manuel Lauss 2 participants 2 messages May 24, 2010 – May 27, 2010 MIPS

lore
Share & tools
lore
#0 lore raw

PATCH RESEND 1/1 MIPS: Alchemy: sleepcode without compile-time cputype dependencies

Split the low-level sleepcode into per-cpu functions instead of relying on compile-time-defined cpu type.

---
This patch didn't make it to 2.6.35 so I'm resending it.

 arch/mips/alchemy/common/power.c           |   12 +++-
 arch/mips/alchemy/common/sleeper.S         |   81 ++++++++++++++++++----------
 arch/mips/include/asm/mach-au1x00/au1000.h |    3 +-
 3 files changed, 64 insertions(+), 32 deletions(-)
arch/mips/alchemy/common/power.c +9 -3
index 14eb8c4..5ef06a1 100644
@@ -193,9 +193,15 @@ static void restore_core_regs(void)
 
 void au_sleep(void)
 {
-	save_core_regs();
-	au1xxx_save_and_sleep();
-	restore_core_regs();
+	int cpuid = alchemy_get_cputype();
+	if (cpuid != ALCHEMY_CPU_UNKNOWN) {
+		save_core_regs();
+		if (cpuid <= ALCHEMY_CPU_AU1500)
+			alchemy_sleep_au1000();
+		else if (cpuid <= ALCHEMY_CPU_AU1200)
+			alchemy_sleep_au1550();
+		restore_core_regs();
+	}
 }
 
 #endif	/* CONFIG_PM */
arch/mips/alchemy/common/sleeper.S +53 -28

122 lines not on this page · raw message

arch/mips/include/asm/mach-au1x00/au1000.h +2 -1
index e76941d..a697661 100644
@@ -188,7 +188,8 @@ extern unsigned long get_au1x00_uart_baud_base(void);
 extern unsigned long au1xxx_calc_clock(void);
 
 /* PM: arch/mips/alchemy/common/sleeper.S, power.c, irq.c */
-void au1xxx_save_and_sleep(void);
+void alchemy_sleep_au1000(void);
+void alchemy_sleep_au1550(void);
 void au_sleep(void);
 
 
Signature · 1 line
1.7.1
#1 lore raw
Quoted 6 lines · Manuel Lauss

Split the low-level sleepcode into per-cpu functions instead of relying on compile-time-defined cpu type.

---
This patch didn't make it to 2.6.35 so I'm resending it.

Sorry, seems I or my patch-o-matic scripts somehow dropped
http://patchwork.linux-mips.org/patch/1038/ on the ground. That patch was supposed to be applied. Anyway, I marked the old patch as superseeded and I'm applying this one now.

Ralf