Skip to contentSkip to conversation

Due to some broken bitfiles, we can't trust IntCtl

PATCH In review Raghu Gandham 2 participants 3 messages Jul 10, 2009 MIPS

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Due to some broken bitfiles, we can't trust IntCtl

From: Robin Randhawa <[email protected]>

---

 arch/mips/include/asm/irq.h |    1 +
 arch/mips/kernel/traps.c    |    4 ++++
 2 files changed, 5 insertions(+), 0 deletions(-)
arch/mips/include/asm/irq.h +1 -0
index 09b08d0..ca0b5ed 100644
@@ -158,6 +158,7 @@ extern void free_irqno(unsigned int irq);
  * IE7.  Since R2 their number has to be read from the c0_intctl register.
  */
 #define CP0_LEGACY_COMPARE_IRQ 7
+#define CP0_LEGACY_PERFCNT_IRQ 7
 
 extern int cp0_compare_irq;
 extern int cp0_perfcount_irq;
arch/mips/kernel/traps.c +4 -0
index 08f1edf..0b6e328 100644
@@ -1538,7 +1538,11 @@ void __cpuinit per_cpu_trap_init(void)
 	 */
 	if (cpu_has_mips_r2) {
 		cp0_compare_irq = (read_c0_intctl() >> 29) & 7;
+		if (!cp0_compare_irq)
+			cp0_compare_irq = CP0_LEGACY_COMPARE_IRQ;
 		cp0_perfcount_irq = (read_c0_intctl() >> 26) & 7;
+		if (!cp0_perfcount_irq)
+			cp0_perfcount_irq = CP0_LEGACY_PERFCNT_IRQ;
 		if (cp0_perfcount_irq == cp0_compare_irq)
 			cp0_perfcount_irq = -1;
 	} else {
#1 lore raw
Quoted 18 lines · Raghu Gandham
arch/mips/include/asm/irq.h +1 -0
index 09b08d0..ca0b5ed 100644
@@ -158,6 +158,7 @@ extern void free_irqno(unsigned int irq);
  * IE7.  Since R2 their number has to be read from the c0_intctl register.
  */
 #define CP0_LEGACY_COMPARE_IRQ 7
+#define CP0_LEGACY_PERFCNT_IRQ 7
 
 extern int cp0_compare_irq;
 extern int cp0_perfcount_irq;
arch/mips/kernel/traps.c +0 -0
index 08f1edf..0b6e328 100644
@@ -1538,7 +1538,11 @@ void __cpuinit per_cpu_trap_init(void)
 	 */

arch/mips/kernel/traps.c void __cpuinit per_cpu_trap_init(void) · line 1548

Quoted 10 lines
arch/mips/kernel/traps.c +4 -0
 	if (cpu_has_mips_r2) {
 		cp0_compare_irq = (read_c0_intctl() >> 29) & 7;
+		if (!cp0_compare_irq)
+			cp0_compare_irq = CP0_LEGACY_COMPARE_IRQ;
 		cp0_perfcount_irq = (read_c0_intctl() >> 26) & 7;
+		if (!cp0_perfcount_irq)
+			cp0_perfcount_irq = CP0_LEGACY_PERFCNT_IRQ;
 		if (cp0_perfcount_irq == cp0_compare_irq)
 			cp0_perfcount_irq = -1;
 	} else {

Is there still any point in applying this patch? I thought only a bunch of early bitfiles were affected but it was never actually taped out?

Ralf

#2 lore raw
Quoted 6 lines

Is there still any point in applying this patch? I thought only a bunch of early bitfiles were affected but it was never actually taped out?

The earlier bitfiles which had the problem never taped out. Please skip this patch.

Thanks,
Raghu

Quoted 3 lines

Ralf