Skip to contentSkip to conversation

mips: mt76xx: Implement new d-cache fix in last_stage_init()

PATCH New Stefan Roese 2 participants 3 messages May 28, 2019 – Jul 5, 2019 MIPS

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mips: mt76xx: Implement new d-cache fix in last_stage_init()

With commit 06985289d452 ("watchdog: Implement generic watchdog_reset() version") the init sequence has changed in arch_misc_init(), resulting in a re-appearance of the d-cache issue on MT7688 boards (e.g. gardena). When this happens, the first (or sometimes later ones as well) TFTP command hangs and does not complete correctly. This leads to the assumption that the d-cache is not in a clean state once the ethernet driver is called (d-cache is used here for the buffers). The old work- around with the cache flush somehow does not work any more now with the new code change.

Unfortunately adding CONFIG_SYS_MALLOC_CLEAR_ON_INIT also did not fix this issue. With v2019.07-rc3 it shows again. The time of accessing the data seems to be very important here. It needs to be "very late" in the boot process.

Testing has shown, that copying a 64KiB area in DDR at a very late bootup time, directly before calling into the prompt, fixes this issue. Flushing of the complete d-cache does not seem to necessary, as this copy alone seems to fix this problem.

Cc Daniel Schwierzeck <[email protected]>
---
 arch/mips/Kconfig           |  2 +-
 arch/mips/mach-mtmips/cpu.c | 26 ++++++++++++++++++++++++++
 2 files changed, 27 insertions(+), 1 deletion(-)
arch/mips/Kconfig +1 -1
index 5cb9bdf2ee..e3e7945567 100644
@@ -84,13 +84,13 @@ config ARCH_MTMIPS
 	select DM_SERIAL
 	imply DM_SPI
 	imply DM_SPI_FLASH
+	select LAST_STAGE_INIT
 	select MIPS_TUNE_24KC
 	select OF_CONTROL
 	select ROM_EXCEPTION_VECTORS
 	select SUPPORTS_CPU_MIPS32_R1
 	select SUPPORTS_CPU_MIPS32_R2
 	select SUPPORTS_LITTLE_ENDIAN
-	select SYS_MALLOC_CLEAR_ON_INIT
 	select SYSRESET
 
 config ARCH_JZ47XX
arch/mips/mach-mtmips/cpu.c +26 -0
index b0a6397d68..7afc2c5940 100644
@@ -68,3 +68,29 @@ int print_cpuinfo(void)
 
 	return 0;
 }
+
+int last_stage_init(void)
+{
+	void *src, *dst;
+
+	src = malloc(SZ_64K);
+	dst = malloc(SZ_64K);
+	if (!src || !dst) {
+		printf("Can't allocate buffer for cache cleanup copy!\n");
+		return 0;
+	}
+
+	/*
+	 * It has been noticed, that sometimes the d-cache is not in a
+	 * "clean-state" when U-Boot is running on MT7688. This was
+	 * detected when using the ethernet driver (which uses d-cache)
+	 * and a TFTP command does not complete. Copying an area of 64KiB
+	 * in DDR at a very late bootup time in U-Boot, directly before
+	 * calling into the prompt, seems to fix this issue.
+	 */
+	memcpy(dst, src, SZ_64K);
+	free(src);
+	free(dst);
+
+	return 0;
+}
Signature · 1 line
2.21.0
#1 lore raw

PATCH 1/1 mips: mt76xx: Implement new d-cache fix in last_stage_init()

Hi Daniel,

Quoted 15 lines · Stefan Roese

With commit 06985289d452 ("watchdog: Implement generic watchdog_reset() version") the init sequence has changed in arch_misc_init(), resulting in a re-appearance of the d-cache issue on MT7688 boards (e.g. gardena). When this happens, the first (or sometimes later ones as well) TFTP command hangs and does not complete correctly. This leads to the assumption that the d-cache is not in a clean state once the ethernet driver is called (d-cache is used here for the buffers). The old work- around with the cache flush somehow does not work any more now with the new code change.

Unfortunately adding CONFIG_SYS_MALLOC_CLEAR_ON_INIT also did not fix this issue. With v2019.07-rc3 it shows again. The time of accessing the data seems to be very important here. It needs to be "very late" in the boot process.

Quoted 7 lines

Testing has shown, that copying a 64KiB area in DDR at a very late bootup time, directly before calling into the prompt, fixes this issue. Flushing of the complete d-cache does not seem to necessary, as this copy alone seems to fix this problem.

Cc Daniel Schwierzeck <[email protected]>

Could you please review this patch and if possible push it for the upcoming release? Without it, we experience the ugly hickups in the tftp tranfers again.

Many thanks,
Stefan

Quoted 49 lines
---
  arch/mips/Kconfig           |  2 +-
  arch/mips/mach-mtmips/cpu.c | 26 ++++++++++++++++++++++++++
  2 files changed, 27 insertions(+), 1 deletion(-)
arch/mips/Kconfig +1 -1
index 5cb9bdf2ee..e3e7945567 100644
@@ -84,13 +84,13 @@ config ARCH_MTMIPS
  	select DM_SERIAL
  	imply DM_SPI
  	imply DM_SPI_FLASH
+	select LAST_STAGE_INIT
  	select MIPS_TUNE_24KC
  	select OF_CONTROL
  	select ROM_EXCEPTION_VECTORS
  	select SUPPORTS_CPU_MIPS32_R1
  	select SUPPORTS_CPU_MIPS32_R2
  	select SUPPORTS_LITTLE_ENDIAN
-	select SYS_MALLOC_CLEAR_ON_INIT
  	select SYSRESET
 
  config ARCH_JZ47XX
arch/mips/mach-mtmips/cpu.c +16 -0
index b0a6397d68..7afc2c5940 100644
@@ -68,3 +68,29 @@ int print_cpuinfo(void)
 
  	return 0;
  }
+
+int last_stage_init(void)
+{
+	void *src, *dst;
+
+	src = malloc(SZ_64K);
+	dst = malloc(SZ_64K);
+	if (!src || !dst) {
+		printf("Can't allocate buffer for cache cleanup copy!\n");
+		return 0;
+	}
+
+	/*
+	 * It has been noticed, that sometimes the d-cache is not in a
+	 * "clean-state" when U-Boot is running on MT7688. This was
+	 * detected when using the ethernet driver (which uses d-cache)

arch/mips/mach-mtmips/cpu.c int print_cpuinfo(void) · line 96

Quoted 10 lines
arch/mips/mach-mtmips/cpu.c +10 -0
+	 * and a TFTP command does not complete. Copying an area of 64KiB
+	 * in DDR at a very late bootup time in U-Boot, directly before
+	 * calling into the prompt, seems to fix this issue.
+	 */
+	memcpy(dst, src, SZ_64K);
+	free(src);
+	free(dst);
+
+	return 0;
+}

Viele Grüße,
Stefan

Signature · 3 lines
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr at denx.de
#2 lore raw

PATCH 1/1 mips: mt76xx: Implement new d-cache fix in last_stage_init()

Hi Stefan,

Am 05.07.19 um 16:42 schrieb Stefan Roese:

Quoted 26 lines

Hi Daniel,

Quoted 15 lines · Stefan Roese

With commit 06985289d452 ("watchdog: Implement generic watchdog_reset() version") the init sequence has changed in arch_misc_init(), resulting in a re-appearance of the d-cache issue on MT7688 boards (e.g. gardena). When this happens, the first (or sometimes later ones as well) TFTP command hangs and does not complete correctly. This leads to the assumption that the d-cache is not in a clean state once the ethernet driver is called (d-cache is used here for the buffers). The old work- around with the cache flush somehow does not work any more now with the new code change.

Unfortunately adding CONFIG_SYS_MALLOC_CLEAR_ON_INIT also did not fix this issue. With v2019.07-rc3 it shows again. The time of accessing the data seems to be very important here. It needs to be "very late" in the boot process.

Quoted 7 lines

Testing has shown, that copying a 64KiB area in DDR at a very late bootup time, directly before calling into the prompt, fixes this issue. Flushing of the complete d-cache does not seem to necessary, as this copy alone seems to fix this problem.

Cc Daniel Schwierzeck <[email protected]>
Quoted 6 lines

Could you please review this patch and if possible push it for the upcoming release? Without it, we experience the ugly hickups in the tftp tranfers again.

Many thanks,
Stefan

sorry, I forgot about that one ;)

I'll send a PR shortly.

Quoted 62 lines
Quoted 48 lines

---
  arch/mips/Kconfig           |  2 +- arch/mips/mach-mtmips/cpu.c | 26 ++++++++++++++++++++++++++

  2 files changed, 27 insertions(+), 1 deletion(-)
arch/mips/Kconfig +0 -0
index 5cb9bdf2ee..e3e7945567 100644
@@ -84,13 +84,13 @@ config ARCH_MTMIPS

      select DM_SERIAL
      imply DM_SPI
      imply DM_SPI_FLASH
+    select LAST_STAGE_INIT
      select MIPS_TUNE_24KC
      select OF_CONTROL
      select ROM_EXCEPTION_VECTORS
      select SUPPORTS_CPU_MIPS32_R1
      select SUPPORTS_CPU_MIPS32_R2
      select SUPPORTS_LITTLE_ENDIAN
-    select SYS_MALLOC_CLEAR_ON_INIT
      select SYSRESET
    config ARCH_JZ47XX

arch/mips/mach-mtmips/cpu.c +17 -0
index b0a6397d68..7afc2c5940 100644
@@ -68,3 +68,29 @@ int print_cpuinfo(void)
         return 0;
   }
+
+int last_stage_init(void)
+{
+    void *src, *dst;
+
+    src = malloc(SZ_64K);
+    dst = malloc(SZ_64K);
+    if (!src || !dst) {
+        printf("Can't allocate buffer for cache cleanup copy!\n");
+        return 0;
+    }
+
+    /*
+     * It has been noticed, that sometimes the d-cache is not in a
+     * "clean-state" when U-Boot is running on MT7688. This was
+     * detected when using the ethernet driver (which uses d-cache)
+     * and a TFTP command does not complete. Copying an area of 64KiB

arch/mips/mach-mtmips/cpu.c int print_cpuinfo(void) · line 96

Quoted 10 lines
arch/mips/mach-mtmips/cpu.c +9 -0
+     * in DDR at a very late bootup time in U-Boot, directly before
+     * calling into the prompt, seems to fix this issue.
+     */
+    memcpy(dst, src, SZ_64K);
+    free(src);
+    free(dst);
+
+    return 0;
+}
 

Viele Grüße,
Stefan

Signature · 1 line
- Daniel