Skip to contentSkip to conversation

coresight: tmc: Fix TMC mode read in tmc_read_unprepare_etb()

PATCH In review Sai Prakash Ranjan 3 participants 3 messages Jun 16, 2020 – Jun 23, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 coresight: tmc: Fix TMC mode read in tmc_read_unprepare_etb()

Reading TMC mode register without proper coresight power management can lead to exceptions like the one in the call trace below in tmc_read_unprepare_etb() when the trace data is read after the sink is disabled. So fix this by having a check for coresight sysfs mode before reading TMC mode management register in tmc_read_unprepare_etb() similar to tmc_read_prepare_etb().

SError Interrupt on CPU6, code 0xbe000411 -- SError
pstate: 80400089 (Nzcv daIf +PAN -UAO)

  pc : tmc_read_unprepare_etb+0x74/0x108
  lr : tmc_read_unprepare_etb+0x54/0x108
  sp : ffffff80d9507c30
  x29: ffffff80d9507c30 x28: ffffff80b3569a0c
  x27: 0000000000000000 x26: 00000000000a0001
  x25: ffffff80cbae9550 x24: 0000000000000010
  x23: ffffffd07296b0f0 x22: ffffffd0109ee028
  x21: 0000000000000000 x20: ffffff80d19e70e0
  x19: ffffff80d19e7080 x18: 0000000000000000
  x17: 0000000000000000 x16: 0000000000000000
  x15: 0000000000000000 x14: 0000000000000000
  x13: 0000000000000000 x12: 0000000000000000
  x11: 0000000000000000 x10: dfffffd000000001
  x9 : 0000000000000000 x8 : 0000000000000002
  x7 : ffffffd071d0fe78 x6 : 0000000000000000
  x5 : 0000000000000080 x4 : 0000000000000001
  x3 : ffffffd071d0fe98 x2 : 0000000000000000
  x1 : 0000000000000004 x0 : 0000000000000001
  Kernel panic - not syncing: Asynchronous SError Interrupt
Fixes 4525412a5046 ("coresight: tmc: making prepare/unprepare functions generic")
---
 drivers/hwtracing/coresight/coresight-tmc-etf.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)
drivers/hwtracing/coresight/coresight-tmc-etf.c +6 -7
index 36cce2bfb744..6375504ba8b0 100644
@@ -639,15 +639,14 @@ int tmc_read_unprepare_etb(struct tmc_drvdata *drvdata)
 
 	spin_lock_irqsave(&drvdata->spinlock, flags);
 
-	/* There is no point in reading a TMC in HW FIFO mode */
-	mode = readl_relaxed(drvdata->base + TMC_MODE);
-	if (mode != TMC_MODE_CIRCULAR_BUFFER) {
-		spin_unlock_irqrestore(&drvdata->spinlock, flags);
-		return -EINVAL;
-	}
-
 	/* Re-enable the TMC if need be */
 	if (drvdata->mode == CS_MODE_SYSFS) {
+		/* There is no point in reading a TMC in HW FIFO mode */
+		mode = readl_relaxed(drvdata->base + TMC_MODE);
+		if (mode != TMC_MODE_CIRCULAR_BUFFER) {
+			spin_unlock_irqrestore(&drvdata->spinlock, flags);
+			return -EINVAL;
+		}
 		/*
 		 * The trace run will continue with the same allocated trace
 		 * buffer. As such zero-out the buffer so that we don't end

base-commit: 3d439a6c349778f129de19595db564a8366c3634

Signature · 2 lines
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member
of Code Aurora Forum, hosted by The Linux Foundation
Tested-by on v1 1/1
#1 lore raw
Quoted 65 lines · Sai Prakash Ranjan

Reading TMC mode register without proper coresight power management can lead to exceptions like the one in the call trace below in tmc_read_unprepare_etb() when the trace data is read after the sink is disabled. So fix this by having a check for coresight sysfs mode before reading TMC mode management register in tmc_read_unprepare_etb() similar to tmc_read_prepare_etb().

SError Interrupt on CPU6, code 0xbe000411 -- SError
pstate: 80400089 (Nzcv daIf +PAN -UAO)

  pc : tmc_read_unprepare_etb+0x74/0x108
  lr : tmc_read_unprepare_etb+0x54/0x108
  sp : ffffff80d9507c30
  x29: ffffff80d9507c30 x28: ffffff80b3569a0c
  x27: 0000000000000000 x26: 00000000000a0001
  x25: ffffff80cbae9550 x24: 0000000000000010
  x23: ffffffd07296b0f0 x22: ffffffd0109ee028
  x21: 0000000000000000 x20: ffffff80d19e70e0
  x19: ffffff80d19e7080 x18: 0000000000000000
  x17: 0000000000000000 x16: 0000000000000000
  x15: 0000000000000000 x14: 0000000000000000
  x13: 0000000000000000 x12: 0000000000000000
  x11: 0000000000000000 x10: dfffffd000000001
  x9 : 0000000000000000 x8 : 0000000000000002
  x7 : ffffffd071d0fe78 x6 : 0000000000000000
  x5 : 0000000000000080 x4 : 0000000000000001
  x3 : ffffffd071d0fe98 x2 : 0000000000000000
  x1 : 0000000000000004 x0 : 0000000000000001
  Kernel panic - not syncing: Asynchronous SError Interrupt
Fixes 4525412a5046 ("coresight: tmc: making prepare/unprepare functions generic")
---
 drivers/hwtracing/coresight/coresight-tmc-etf.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)
drivers/hwtracing/coresight/coresight-tmc-etf.c +6 -7
index 36cce2bfb744..6375504ba8b0 100644
@@ -639,15 +639,14 @@ int tmc_read_unprepare_etb(struct tmc_drvdata *drvdata)
 
        spin_lock_irqsave(&drvdata->spinlock, flags);
 
-       /* There is no point in reading a TMC in HW FIFO mode */
-       mode = readl_relaxed(drvdata->base + TMC_MODE);
-       if (mode != TMC_MODE_CIRCULAR_BUFFER) {
-               spin_unlock_irqrestore(&drvdata->spinlock, flags);
-               return -EINVAL;
-       }
-
        /* Re-enable the TMC if need be */
        if (drvdata->mode == CS_MODE_SYSFS) {
+               /* There is no point in reading a TMC in HW FIFO mode */
+               mode = readl_relaxed(drvdata->base + TMC_MODE);
+               if (mode != TMC_MODE_CIRCULAR_BUFFER) {
+                       spin_unlock_irqrestore(&drvdata->spinlock, flags);
+                       return -EINVAL;
+               }
                /*
                 * The trace run will continue with the same allocated trace
                 * buffer. As such zero-out the buffer so that we don't end
Quoted 5 lines

base-commit: 3d439a6c349778f129de19595db564a8366c3634
--
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

Tested OK on my DB410c / Linux 5.8-rc1

Signature · 8 lines
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#2 lore raw
In reply to Sai Prakash Ranjan
Quoted 30 lines · Sai Prakash Ranjan

Reading TMC mode register without proper coresight power management can lead to exceptions like the one in the call trace below in tmc_read_unprepare_etb() when the trace data is read after the sink is disabled. So fix this by having a check for coresight sysfs mode before reading TMC mode management register in tmc_read_unprepare_etb() similar to tmc_read_prepare_etb().

SError Interrupt on CPU6, code 0xbe000411 -- SError
pstate: 80400089 (Nzcv daIf +PAN -UAO)

  pc : tmc_read_unprepare_etb+0x74/0x108
  lr : tmc_read_unprepare_etb+0x54/0x108
  sp : ffffff80d9507c30
  x29: ffffff80d9507c30 x28: ffffff80b3569a0c
  x27: 0000000000000000 x26: 00000000000a0001
  x25: ffffff80cbae9550 x24: 0000000000000010
  x23: ffffffd07296b0f0 x22: ffffffd0109ee028
  x21: 0000000000000000 x20: ffffff80d19e70e0
  x19: ffffff80d19e7080 x18: 0000000000000000
  x17: 0000000000000000 x16: 0000000000000000
  x15: 0000000000000000 x14: 0000000000000000
  x13: 0000000000000000 x12: 0000000000000000
  x11: 0000000000000000 x10: dfffffd000000001
  x9 : 0000000000000000 x8 : 0000000000000002
  x7 : ffffffd071d0fe78 x6 : 0000000000000000
  x5 : 0000000000000080 x4 : 0000000000000001
  x3 : ffffffd071d0fe98 x2 : 0000000000000000
  x1 : 0000000000000004 x0 : 0000000000000001
  Kernel panic - not syncing: Asynchronous SError Interrupt
Quoted 7 lines
Fixes 4525412a5046 ("coresight: tmc: making prepare/unprepare functions generic")
---
 drivers/hwtracing/coresight/coresight-tmc-etf.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

And this one too.

Quoted 32 lines
drivers/hwtracing/coresight/coresight-tmc-etf.c +6 -7
index 36cce2bfb744..6375504ba8b0 100644
@@ -639,15 +639,14 @@ int tmc_read_unprepare_etb(struct tmc_drvdata *drvdata)
 
 	spin_lock_irqsave(&drvdata->spinlock, flags);
 
-	/* There is no point in reading a TMC in HW FIFO mode */
-	mode = readl_relaxed(drvdata->base + TMC_MODE);
-	if (mode != TMC_MODE_CIRCULAR_BUFFER) {
-		spin_unlock_irqrestore(&drvdata->spinlock, flags);
-		return -EINVAL;
-	}
-
 	/* Re-enable the TMC if need be */
 	if (drvdata->mode == CS_MODE_SYSFS) {
+		/* There is no point in reading a TMC in HW FIFO mode */
+		mode = readl_relaxed(drvdata->base + TMC_MODE);
+		if (mode != TMC_MODE_CIRCULAR_BUFFER) {
+			spin_unlock_irqrestore(&drvdata->spinlock, flags);
+			return -EINVAL;
+		}
 		/*
 		 * The trace run will continue with the same allocated trace
 		 * buffer. As such zero-out the buffer so that we don't end

base-commit: 3d439a6c349778f129de19595db564a8366c3634

Signature · 2 lines
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member
of Code Aurora Forum, hosted by The Linux Foundation