Skip to contentSkip to conversation

Fix watchdog on Renesas R-Car Gen2 and RZ/G1

PATCH Superseded Fabrizio Castro 6 participants 65 messages Feb 12, 2018 – May 2, 2018

Subsystems (7)
lore
Share & tools
lore
#0 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

Dear All,

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock
clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

 .../devicetree/bindings/watchdog/renesas-wdt.txt   | 19 ++++++--
 arch/arm/boot/dts/r8a7743-iwg20m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7743.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7745-iwg22m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7745.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7790.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7791.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7792.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7793.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7794.dtsi                     | 12 ++++-
 arch/arm/configs/shmobile_defconfig                |  1 +
 arch/arm/mach-shmobile/common.h                    |  6 +++
 arch/arm/mach-shmobile/headsmp.S                   | 55 ++++++++++++++++++++++
 arch/arm/mach-shmobile/platsmp-apmu.c              |  1 +
 arch/arm/mach-shmobile/pm-rcar-gen2.c              | 15 ++++--
 drivers/clk/renesas/r8a7743-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7745-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7790-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7791-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7794-cpg-mssr.c             |  2 +
 drivers/soc/renesas/rcar-rst.c                     | 35 +++++++++++---
 drivers/watchdog/renesas_wdt.c                     | 39 +++++++++++++--
 22 files changed, 229 insertions(+), 26 deletions(-)
Signature · 1 line
2.7.4
#1 lore raw

PATCH v5 1/26 ARM: shmobile: Add watchdog support

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

---
v4->v5:
* no change

 arch/arm/mach-shmobile/common.h  |  6 +++++
 arch/arm/mach-shmobile/headsmp.S | 55 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 61 insertions(+)
arch/arm/mach-shmobile/common.h +6 -0
index a8fa4f7..d8c11f5 100644
@@ -7,6 +7,12 @@ extern void shmobile_init_delay(void);
 extern void shmobile_boot_vector(void);
 extern unsigned long shmobile_boot_fn;
 extern unsigned long shmobile_boot_size;
+#ifdef CONFIG_ARCH_RCAR_GEN2
+extern void shmobile_boot_vector_gen2(void);
+extern unsigned long shmobile_boot_fn_gen2;
+extern unsigned long shmobile_boot_cpu_gen2;
+extern unsigned long shmobile_boot_size_gen2;
+#endif /* CONFIG_ARCH_RCAR_GEN2 */
 extern void shmobile_smp_boot(void);
 extern void shmobile_smp_sleep(void);
 extern void shmobile_smp_hook(unsigned int cpu, unsigned long fn,
arch/arm/mach-shmobile/headsmp.S +55 -0

67 lines not on this page · raw message

Signature · 1 line
2.7.4
#2 lore raw

PATCH v5 1/26 ARM: shmobile: Add watchdog support

Hi Fabrizio,

Quoted 28 lines · Fabrizio Castro

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

arch/arm/mach-shmobile/common.h extern void shmobile_init_delay(void); · line 10

Quoted 4 lines
arch/arm/mach-shmobile/common.h +1 -0
@@ -7,6 +7,12 @@ extern void shmobile_init_delay(void);
 extern void shmobile_boot_vector(void);
 extern unsigned long shmobile_boot_fn;
 extern unsigned long shmobile_boot_size;
+#ifdef CONFIG_ARCH_RCAR_GEN2

I think this #ifdef can be removed.

arch/arm/mach-shmobile/common.h extern void shmobile_init_delay(void); · line 18

Quoted 8 lines
arch/arm/mach-shmobile/common.h +5 -0
+extern void shmobile_boot_vector_gen2(void);
+extern unsigned long shmobile_boot_fn_gen2;
+extern unsigned long shmobile_boot_cpu_gen2;
+extern unsigned long shmobile_boot_size_gen2;
+#endif /* CONFIG_ARCH_RCAR_GEN2 */
 extern void shmobile_smp_boot(void);
 extern void shmobile_smp_sleep(void);
 extern void shmobile_smp_hook(unsigned int cpu, unsigned long fn,

Apart from that:

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#3 lore raw

PATCH v5 1/26 ARM: shmobile: Add watchdog support

Hi Geert,

thank you for your feedback!

Quoted 31 lines

Subject: Re: [PATCH v5 01/26] ARM: shmobile: Add watchdog support

Hi Fabrizio,

Quoted 24 lines · Fabrizio Castro

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

Quoted 9 lines

arch/arm/mach-shmobile/common.h extern void shmobile_init_delay(void); · line 10

Quoted 7 lines
arch/arm/mach-shmobile/common.h +1 -0
@@ -7,6 +7,12 @@ extern void shmobile_init_delay(void);
 extern void shmobile_boot_vector(void);
 extern unsigned long shmobile_boot_fn;
 extern unsigned long shmobile_boot_size;
+#ifdef CONFIG_ARCH_RCAR_GEN2

I think this #ifdef can be removed.

I'll send a new version of this patch without ifdefs.

Thanks,
Fab

Quoted 15 lines

arch/arm/mach-shmobile/common.h extern void shmobile_init_delay(void); · line 18

Quoted 9 lines
arch/arm/mach-shmobile/common.h +5 -0
+extern void shmobile_boot_vector_gen2(void);
+extern unsigned long shmobile_boot_fn_gen2;
+extern unsigned long shmobile_boot_cpu_gen2;
+extern unsigned long shmobile_boot_size_gen2;
+#endif /* CONFIG_ARCH_RCAR_GEN2 */
 extern void shmobile_smp_boot(void);
 extern void shmobile_smp_sleep(void);
 extern void shmobile_smp_hook(unsigned int cpu, unsigned long fn,

Apart from that:

Gr{oetje,eeting}s,

Quoted 8 lines

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

Renesas Electronics Europe Ltd, Dukes Meadow, Millboard Road, Bourne End, Buckinghamshire, SL8 5FH, UK. Registered in England & Wales under Registered No. 04586709.

#4 lore raw

PATCH v6 1/26 ARM: shmobile: Add watchdog support

In reply to Geert Uytterhoeven

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

---
v5->v6:
* taken ifdefs out as per Geert's suggestion.

 arch/arm/mach-shmobile/common.h  |  4 +++
 arch/arm/mach-shmobile/headsmp.S | 53 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 57 insertions(+)
arch/arm/mach-shmobile/common.h +4 -0
index a8fa4f7..43c1ac69 100644
@@ -7,6 +7,10 @@ extern void shmobile_init_delay(void);
 extern void shmobile_boot_vector(void);
 extern unsigned long shmobile_boot_fn;
 extern unsigned long shmobile_boot_size;
+extern void shmobile_boot_vector_gen2(void);
+extern unsigned long shmobile_boot_fn_gen2;
+extern unsigned long shmobile_boot_cpu_gen2;
+extern unsigned long shmobile_boot_size_gen2;
 extern void shmobile_smp_boot(void);
 extern void shmobile_smp_sleep(void);
 extern void shmobile_smp_hook(unsigned int cpu, unsigned long fn,
arch/arm/mach-shmobile/headsmp.S +53 -0

65 lines not on this page · raw message

Signature · 1 line
2.7.4
#5 lore raw

PATCH v6 1/26 ARM: shmobile: Add watchdog support

Quoted 18 lines · Fabrizio Castro

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Quoted 6 lines

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

Thanks, I have marked this as deferred as I would like dependencies to be merged first.

#6 lore raw

PATCH v6 1/26 ARM: shmobile: Add watchdog support

In reply to Fabrizio Castro

Hi Fabrizio,

Quoted 22 lines · Fabrizio Castro

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

Quoted 6 lines

---
v5->v6:
* taken ifdefs out as per Geert's suggestion.

My intention was to remove the #ifdefs from the header file only, not from arch/arm/mach-shmobile/headsmp.S, as the latter is used for other SoC families, too.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#7 lore raw

PATCH v6 1/26 ARM: shmobile: Add watchdog support

Hi Geert,

Quoted 35 lines

Subject: Re: [PATCH v6 01/26] ARM: shmobile: Add watchdog support

Hi Fabrizio,

Quoted 22 lines · Fabrizio Castro

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

Quoted 6 lines

---
v5->v6:
* taken ifdefs out as per Geert's suggestion.

Quoted 3 lines

My intention was to remove the #ifdefs from the header file only, not from arch/arm/mach-shmobile/headsmp.S, as the latter is used for other SoC families, too.

Ah ok, I'll send another version to restore the ifdef within headsmp.S then.

Thanks,
Fab

Quoted 11 lines

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

Renesas Electronics Europe Ltd, Dukes Meadow, Millboard Road, Bourne End, Buckinghamshire, SL8 5FH, UK. Registered in England & Wales under Registered No. 04586709.

#8 lore raw

PATCH v7 1/1 ARM: shmobile: Add watchdog support

In reply to Geert Uytterhoeven

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

---
v6->v7:
* restored ifdef within arch/arm/mach-shmobile/headsmp.S

 arch/arm/mach-shmobile/common.h  |  4 +++
 arch/arm/mach-shmobile/headsmp.S | 55 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 59 insertions(+)
arch/arm/mach-shmobile/common.h +4 -0
index a8fa4f7..43c1ac69 100644
@@ -7,6 +7,10 @@ extern void shmobile_init_delay(void);
 extern void shmobile_boot_vector(void);
 extern unsigned long shmobile_boot_fn;
 extern unsigned long shmobile_boot_size;
+extern void shmobile_boot_vector_gen2(void);
+extern unsigned long shmobile_boot_fn_gen2;
+extern unsigned long shmobile_boot_cpu_gen2;
+extern unsigned long shmobile_boot_size_gen2;
 extern void shmobile_smp_boot(void);
 extern void shmobile_smp_sleep(void);
 extern void shmobile_smp_hook(unsigned int cpu, unsigned long fn,
arch/arm/mach-shmobile/headsmp.S +55 -0

67 lines not on this page · raw message

Signature · 1 line
2.7.4
#9 lore raw

PATCH v7 1/1 ARM: shmobile: Add watchdog support

Quoted 22 lines · Fabrizio Castro

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

Quoted 6 lines

---
v6->v7:
* restored ifdef within arch/arm/mach-shmobile/headsmp.S

Thanks. I believe that this is the same as the hand modified version of v5 that I applied yesterday.

#10 lore raw

PATCH v7 1/1 ARM: shmobile: Add watchdog support

Hello Simon,

Quoted 32 lines

Subject: Re: [PATCH v7] ARM: shmobile: Add watchdog support

Quoted 22 lines · Fabrizio Castro

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

Quoted 6 lines

---
v6->v7:
* restored ifdef within arch/arm/mach-shmobile/headsmp.S

Quoted 2 lines

Thanks. I believe that this is the same as the hand modified version of v5 that I applied yesterday.

yes, I believe v5 modified and applied by you is the same as v7. Thank you for taking the patch.

Fab

Renesas Electronics Europe Ltd, Dukes Meadow, Millboard Road, Bourne End, Buckinghamshire, SL8 5FH, UK. Registered in England & Wales under Registered No. 04586709.

#11 lore raw

PATCH v7 1/1 ARM: shmobile: Add watchdog support

Quoted 34 lines · Fabrizio Castro

Hello Simon,

Quoted 32 lines

Subject: Re: [PATCH v7] ARM: shmobile: Add watchdog support

Quoted 28 lines · Fabrizio Castro

On R-Car Gen2 and RZ/G1 platforms, we use the SBAR registers to make non boot CPUs run a routine designed to bring up SMP and deal with hot plug. The value contained in the SBAR registers is not initialized by a WDT triggered reset, which means that after a WDT triggered reset we jump to the SMP bring up routine, preventing the system from executing the bootrom code.

The purpose of this patch is to jump to the bootrom code in case of a WDT triggered reset, and keep the SMP functionality untouched. In order to tell if the code had been called due to the WDT overflowing we are testing WOVF from register RWTCSRA.

The new function shmobile_boot_vector_gen2 isn't replacing shmobile_boot_vector for backward compatibility reasons. The kernel will install the best option (either shmobile_boot_vector or shmobile_boot_vector_gen2) to ICRAM1 after parsing the device tree, according to the amount of memory available.

Since shmobile_boot_vector has become bigger, "reg" property of nodes compatible with "renesas,smp-sram" now need to be set to a value greater or equal to "<0 0x60>".

---
v6->v7:
* restored ifdef within arch/arm/mach-shmobile/headsmp.S

Quoted 5 lines
Quoted 2 lines

Thanks. I believe that this is the same as the hand modified version of v5 that I applied yesterday.

yes, I believe v5 modified and applied by you is the same as v7. Thank you for taking the patch.

Likewise, thanks for the patches and for confirming this one is correct.

#12 lore raw

PATCH v5 2/26 ARM: dts: r8a7743: Adjust SMP routine size

In reply to Fabrizio Castro

This patch adjusts the definition of the SMP routine size according to the latest changes made by commit:
"ARM: shmobile: Add watchdog support"

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7743.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/r8a7743.dtsi +1 -1
index 47b62f6..c2943d7 100644
@@ -973,7 +973,7 @@
 
 			smp-sram at 0 {
 				compatible = "renesas,smp-sram";
-				reg = <0 0x10>;
+				reg = <0 0x100>;
 			};
 		};
 
Signature · 1 line
2.7.4
#13 lore raw

PATCH v5 3/26 ARM: dts: r8a7745: Adjust SMP routine size

In reply to Fabrizio Castro

This patch adjusts the definition of the SMP routine size according to the latest changes made by commit:
"ARM: shmobile: Add watchdog support"

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7745.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/r8a7745.dtsi +1 -1
index ddc3da2..6bc1cf4 100644
@@ -755,7 +755,7 @@
 
 			smp-sram at 0 {
 				compatible = "renesas,smp-sram";
-				reg = <0 0x10>;
+				reg = <0 0x100>;
 			};
 		};
 
Signature · 1 line
2.7.4
#14 lore raw

PATCH v5 4/26 ARM: dts: r8a7790: Adjust SMP routine size

In reply to Fabrizio Castro

This patch adjusts the definition of the SMP routine size according to the latest changes made by commit:
"ARM: shmobile: Add watchdog support"

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7790.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/r8a7790.dtsi +1 -1
index 3bbcc0b..a84bf35 100644
@@ -443,7 +443,7 @@
 
 			smp-sram at 0 {
 				compatible = "renesas,smp-sram";
-				reg = <0 0x10>;
+				reg = <0 0x100>;
 			};
 		};
 
Signature · 1 line
2.7.4
#15 lore raw

PATCH v5 5/26 ARM: dts: r8a7791: Adjust SMP routine size

In reply to Fabrizio Castro

This patch adjusts the definition of the SMP routine size according to the latest changes made by commit:
"ARM: shmobile: Add watchdog support"

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7791.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/r8a7791.dtsi +1 -1
index dc65935..d2a4546 100644
@@ -407,7 +407,7 @@
 
 			smp-sram at 0 {
 				compatible = "renesas,smp-sram";
-				reg = <0 0x10>;
+				reg = <0 0x100>;
 			};
 		};
 
Signature · 1 line
2.7.4
#16 lore raw

PATCH v5 6/26 ARM: dts: r8a7792: Adjust SMP routine size

In reply to Fabrizio Castro

This patch adjusts the definition of the SMP routine size according to the latest changes made by commit:
"ARM: shmobile: Add watchdog support"

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7792.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/r8a7792.dtsi +1 -1
index 268987f..bea0f12 100644
@@ -341,7 +341,7 @@
 
 			smp-sram at 0 {
 				compatible = "renesas,smp-sram";
-				reg = <0 0x10>;
+				reg = <0 0x100>;
 			};
 		};
 
Signature · 1 line
2.7.4
#17 lore raw

PATCH v5 7/26 ARM: dts: r8a7793: Adjust SMP routine size

In reply to Fabrizio Castro

This patch adjusts the definition of the SMP routine size according to the latest changes made by commit:
"ARM: shmobile: Add watchdog support"

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7793.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/r8a7793.dtsi +1 -1
index aa7d779..bde6969 100644
@@ -392,7 +392,7 @@
 
 			smp-sram at 0 {
 				compatible = "renesas,smp-sram";
-				reg = <0 0x10>;
+				reg = <0 0x100>;
 			};
 		};
 
Signature · 1 line
2.7.4
#18 lore raw

PATCH v5 8/26 ARM: dts: r8a7794: Adjust SMP routine size

In reply to Fabrizio Castro

This patch adjusts the definition of the SMP routine size according to the latest changes made by commit:
"ARM: shmobile: Add watchdog support"

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7794.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/r8a7794.dtsi +1 -1
index d588efa..e47c5f4 100644
@@ -348,7 +348,7 @@
 
 			smp-sram at 0 {
 				compatible = "renesas,smp-sram";
-				reg = <0 0x10>;
+				reg = <0 0x100>;
 			};
 		};
 
Signature · 1 line
2.7.4
#19 lore raw

PATCH v5 9/26 soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2

In reply to Fabrizio Castro

This patch allows for platform specific quirks as some of the SoC need further customization for the watchdog to work properly, like for R-Car Gen2 and for RZ/G.

---
v4->v5:
* no change

 drivers/soc/renesas/rcar-rst.c | 35 +++++++++++++++++++++++++++++------
 1 file changed, 29 insertions(+), 6 deletions(-)
drivers/soc/renesas/rcar-rst.c +29 -6
index 3316b02..93febe6 100644
@@ -13,8 +13,18 @@
 #include <linux/of_address.h>
 #include <linux/soc/renesas/rcar-rst.h>
 
+#define WDTRSTCR_RESET		0xA55A0002
+#define WDTRSTCR		0x0054
+
+static int rcar_rst_enable_wdt_reset(void __iomem *base)
+{
+	iowrite32(WDTRSTCR_RESET, base + WDTRSTCR);
+	return 0;
+}
+
 struct rst_config {
-	unsigned int modemr;	/* Mode Monitoring Register Offset */
+	unsigned int modemr;		/* Mode Monitoring Register Offset */
+	int (*configure)(void *base);	/* Platform specific configuration */
 };
 
 static const struct rst_config rcar_rst_gen1 __initconst = {
@@ -23,6 +33,11 @@ static const struct rst_config rcar_rst_gen1 __initconst = {
 
 static const struct rst_config rcar_rst_gen2 __initconst = {
 	.modemr = 0x60,
+	.configure = rcar_rst_enable_wdt_reset,
+};
+
+static const struct rst_config rcar_rst_gen3 __initconst = {
+	.modemr = 0x60,
 };
 
 static const struct of_device_id rcar_rst_matches[] __initconst = {
@@ -38,11 +53,11 @@ static const struct of_device_id rcar_rst_matches[] __initconst = {
 	{ .compatible = "renesas,r8a7792-rst", .data = &rcar_rst_gen2 },
 	{ .compatible = "renesas,r8a7793-rst", .data = &rcar_rst_gen2 },
 	{ .compatible = "renesas,r8a7794-rst", .data = &rcar_rst_gen2 },
-	/* R-Car Gen3 is handled like R-Car Gen2 */
-	{ .compatible = "renesas,r8a7795-rst", .data = &rcar_rst_gen2 },
-	{ .compatible = "renesas,r8a7796-rst", .data = &rcar_rst_gen2 },
-	{ .compatible = "renesas,r8a77970-rst", .data = &rcar_rst_gen2 },
-	{ .compatible = "renesas,r8a77995-rst", .data = &rcar_rst_gen2 },
+	/* R-Car Gen3 */
+	{ .compatible = "renesas,r8a7795-rst", .data = &rcar_rst_gen3 },
+	{ .compatible = "renesas,r8a7796-rst", .data = &rcar_rst_gen3 },
+	{ .compatible = "renesas,r8a77970-rst", .data = &rcar_rst_gen3 },
+	{ .compatible = "renesas,r8a77995-rst", .data = &rcar_rst_gen3 },
 	{ /* sentinel */ }
 };
 
@@ -71,6 +86,14 @@ static int __init rcar_rst_init(void)
 	rcar_rst_base = base;
 	cfg = match->data;
 	saved_mode = ioread32(base + cfg->modemr);
+	if (cfg->configure) {
+		error = cfg->configure(base);
+		if (error) {
+			pr_warn("%pOF: Cannot run SoC specific configuration\n",
+				np);
+			goto out_put;
+		}
+	}
 
 	pr_debug("%pOF: MODE = 0x%08x\n", np, saved_mode);
 
Signature · 1 line
2.7.4
#20 lore raw

PATCH v5 9/26 soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2

Quoted 6 lines · Fabrizio Castro

This patch allows for platform specific quirks as some of the SoC need further customization for the watchdog to work properly, like for R-Car Gen2 and for RZ/G.

I am fine with this patch but please rebase it on top of renesas-devel-20180212-v4.16-rc1 (or later).

#21 lore raw

PATCH v5 9/26 soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2

Hello Simon,

Quoted 7 lines

Subject: Re: [PATCH v5 09/26] soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2

Quoted 4 lines · Fabrizio Castro

This patch allows for platform specific quirks as some of the SoC need further customization for the watchdog to work properly, like for R-Car Gen2 and for RZ/G.

Quoted 5 lines
Quoted 2 lines

I am fine with this patch but please rebase it on top of renesas-devel-20180212-v4.16-rc1 (or later).

sure, I will.

Thanks,
Fab

Renesas Electronics Europe Ltd, Dukes Meadow, Millboard Road, Bourne End, Buckinghamshire, SL8 5FH, UK. Registered in England & Wales under Registered No. 04586709.

#22 lore raw

PATCH v6 9/26 soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2

In reply to Simon Horman

This patch allows for platform specific quirks as some of the SoC need further customization for the watchdog to work properly, like for R-Car Gen2 and for RZ/G.

---
v5->v6:
* rebased on top of renesas-devel-20180213-v4.16-rc1

 drivers/soc/renesas/rcar-rst.c | 37 ++++++++++++++++++++++++++++++-------
 1 file changed, 30 insertions(+), 7 deletions(-)
drivers/soc/renesas/rcar-rst.c +30 -7
index e2340eb..3413666 100644
@@ -13,8 +13,18 @@
 #include <linux/of_address.h>
 #include <linux/soc/renesas/rcar-rst.h>
 
+#define WDTRSTCR_RESET		0xA55A0002
+#define WDTRSTCR		0x0054
+
+static int rcar_rst_enable_wdt_reset(void __iomem *base)
+{
+	iowrite32(WDTRSTCR_RESET, base + WDTRSTCR);
+	return 0;
+}
+
 struct rst_config {
-	unsigned int modemr;	/* Mode Monitoring Register Offset */
+	unsigned int modemr;		/* Mode Monitoring Register Offset */
+	int (*configure)(void *base);	/* Platform specific configuration */
 };
 
 static const struct rst_config rcar_rst_gen1 __initconst = {
@@ -23,6 +33,11 @@ static const struct rst_config rcar_rst_gen1 __initconst = {
 
 static const struct rst_config rcar_rst_gen2 __initconst = {
 	.modemr = 0x60,
+	.configure = rcar_rst_enable_wdt_reset,
+};
+
+static const struct rst_config rcar_rst_gen3 __initconst = {
+	.modemr = 0x60,
 };
 
 static const struct of_device_id rcar_rst_matches[] __initconst = {
@@ -38,12 +53,12 @@ static const struct of_device_id rcar_rst_matches[] __initconst = {
 	{ .compatible = "renesas,r8a7792-rst", .data = &rcar_rst_gen2 },
 	{ .compatible = "renesas,r8a7793-rst", .data = &rcar_rst_gen2 },
 	{ .compatible = "renesas,r8a7794-rst", .data = &rcar_rst_gen2 },
-	/* R-Car Gen3 is handled like R-Car Gen2 */
-	{ .compatible = "renesas,r8a7795-rst", .data = &rcar_rst_gen2 },
-	{ .compatible = "renesas,r8a7796-rst", .data = &rcar_rst_gen2 },
-	{ .compatible = "renesas,r8a77970-rst", .data = &rcar_rst_gen2 },
-	{ .compatible = "renesas,r8a77980-rst", .data = &rcar_rst_gen2 },
-	{ .compatible = "renesas,r8a77995-rst", .data = &rcar_rst_gen2 },
+	/* R-Car Gen3 */
+	{ .compatible = "renesas,r8a7795-rst", .data = &rcar_rst_gen3 },
+	{ .compatible = "renesas,r8a7796-rst", .data = &rcar_rst_gen3 },
+	{ .compatible = "renesas,r8a77970-rst", .data = &rcar_rst_gen3 },
+	{ .compatible = "renesas,r8a77980-rst", .data = &rcar_rst_gen3 },
+	{ .compatible = "renesas,r8a77995-rst", .data = &rcar_rst_gen3 },
 	{ /* sentinel */ }
 };
 
@@ -72,6 +87,14 @@ static int __init rcar_rst_init(void)
 	rcar_rst_base = base;
 	cfg = match->data;
 	saved_mode = ioread32(base + cfg->modemr);
+	if (cfg->configure) {
+		error = cfg->configure(base);
+		if (error) {
+			pr_warn("%pOF: Cannot run SoC specific configuration\n",
+				np);
+			goto out_put;
+		}
+	}
 
 	pr_debug("%pOF: MODE = 0x%08x\n", np, saved_mode);
 
Signature · 1 line
2.7.4
Applied
#23 lore raw

PATCH v6 9/26 soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2

Quoted 9 lines · Fabrizio Castro

This patch allows for platform specific quirks as some of the SoC need further customization for the watchdog to work properly, like for R-Car Gen2 and for RZ/G.

---
v5->v6:
* rebased on top of renesas-devel-20180213-v4.16-rc1

Thanks, applied.

#24 lore raw

PATCH v5 10/26 ARM: shmobile: rcar-gen2: Add watchdog support

In reply to Fabrizio Castro

This patch adds watchdog support by installing shmobile_boot_vector_gen2 to ICRAM1 when enough memory is available, in which case we also keep a copy of MPIDR to complete the reset vector logic.

---
v4->v5:
* reworked comment style

 arch/arm/mach-shmobile/platsmp-apmu.c |  1 +
 arch/arm/mach-shmobile/pm-rcar-gen2.c | 15 ++++++++++++---
 2 files changed, 13 insertions(+), 3 deletions(-)
arch/arm/mach-shmobile/platsmp-apmu.c +1 -0
index 4422b61..ba732ef 100644
@@ -191,6 +191,7 @@ static void __init shmobile_smp_apmu_setup_boot(void)
 {
 	/* install boot code shared by all CPUs */
 	shmobile_boot_fn = __pa_symbol(shmobile_smp_boot);
+	shmobile_boot_fn_gen2 = shmobile_boot_fn;
 }
 
 void __init shmobile_smp_apmu_prepare_cpus(unsigned int max_cpus,
arch/arm/mach-shmobile/pm-rcar-gen2.c +12 -3
index e5f215c..5a798b4 100644
@@ -17,6 +17,7 @@
 #include <linux/smp.h>
 #include <linux/soc/renesas/rcar-sysc.h>
 #include <asm/io.h>
+#include <asm/cputype.h>
 #include "common.h"
 #include "rcar-gen2.h"
 
@@ -37,7 +38,6 @@
 #define CA7RESCNT_CODE	0x5a5a0000
 #define CA7RESCNT_CPUS	0xf		/* CPU0-3 */
 
-
 /* On-chip RAM */
 #define ICRAM1		0xe63c0000	/* Inter Connect RAM1 (4 KiB) */
 
@@ -119,8 +119,17 @@ void __init rcar_gen2_pm_init(void)
 	p = ioremap(res.start, resource_size(&res));
 	if (!p)
 		return;
-
-	memcpy_toio(p, shmobile_boot_vector, shmobile_boot_size);
+	/*
+	 * install the reset vector, use the largest version if we have enough
+	 * memory available
+	 */
+	if (resource_size(&res) >= shmobile_boot_size_gen2) {
+		shmobile_boot_cpu_gen2 = read_cpuid_mpidr();
+		memcpy_toio(p, shmobile_boot_vector_gen2,
+			    shmobile_boot_size_gen2);
+	} else {
+		memcpy_toio(p, shmobile_boot_vector, shmobile_boot_size);
+	}
 	iounmap(p);
 
 	/* setup reset vectors */
Signature · 1 line
2.7.4
#25 lore raw

PATCH v5 10/26 ARM: shmobile: rcar-gen2: Add watchdog support

Quoted 6 lines · Fabrizio Castro

This patch adds watchdog support by installing shmobile_boot_vector_gen2 to ICRAM1 when enough memory is available, in which case we also keep a copy of MPIDR to complete the reset vector logic.

I am marking this patch as deferred as it has dependencies on watchdog driver and in turn clock changes in this series.

#26 lore raw

PATCH v5 11/26 dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support

In reply to Fabrizio Castro

This commit documents the compatibility with R-Car Gen2 and RZ/G devices by defining the generic compatible string "renesas,rcar-gen2-wdt". Also, this patch expands the list of SoC-specific compatible strings to include RZ/G and R-Car Gen2 devices.

---
v4->v5:
* no change

 .../devicetree/bindings/watchdog/renesas-wdt.txt      | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)
Documentation/devicetree/bindings/watchdog/renesas-wdt.txt +14 -5
index bf6d1ca..4815128 100644
@@ -1,17 +1,26 @@
 Renesas Watchdog Timer (WDT) Controller
 
 Required properties:
-- compatible : Should be "renesas,<soctype>-wdt", and
-	       "renesas,rcar-gen3-wdt" or "renesas,rza-wdt" as fallback.
+ - compatible : Must be "renesas,<soctype>-wdt", followed by a generic
+		fallback compatible string when compatible with the generic
+		version.
 	       Examples with soctypes are:
+		 - "renesas,r8a7743-wdt" (RZ/G1M)
+		 - "renesas,r8a7745-wdt" (RZ/G1E)
+	         - "renesas,r8a7790-wdt" (R-Car H2)
+	         - "renesas,r8a7791-wdt" (R-Car M2-W)
+	         - "renesas,r8a7792-wdt" (R-Car V2H)
+	         - "renesas,r8a7793-wdt" (R-Car M2-N)
+	         - "renesas,r8a7794-wdt" (R-Car E2)
 	         - "renesas,r8a7795-wdt" (R-Car H3)
 	         - "renesas,r8a7796-wdt" (R-Car M3-W)
 	         - "renesas,r8a77995-wdt" (R-Car D3)
 	         - "renesas,r7s72100-wdt" (RZ/A1)
+		The generic compatible string must be:
+		 - "renesas,rza-wdt" for RZ/A
+		 - "renesas,rcar-gen2-wdt" for R-Car Gen2 and RZ/G
+		 - "renesas,rcar-gen3-wdt" for R-Car Gen3
 
-  When compatible with the generic version, nodes must list the SoC-specific
-  version corresponding to the platform first, followed by the generic
-  version.
 
 - reg : Should contain WDT registers location and length
 - clocks : the clock feeding the watchdog timer.
Signature · 1 line
2.7.4
#27 lore raw

PATCH v5 12/26 watchdog: renesas_wdt: Add R-Car Gen2 support

In reply to Fabrizio Castro

Due to commits:
* "ARM: shmobile: Add watchdog support",
* "ARM: shmobile: rcar-gen2: Add watchdog support", and
* "soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2", we now have everything we needed for the watchdog to work on Gen2 and RZ/G1.

This commit adds "renesas,rcar-gen2-wdt" as compatible string for R-Car Gen2 and RZ/G1, and since on those platforms the rwdt clock needs to be always ON, when suspending to RAM we need to explicitly disable the counting by clearing TME from RWTCSRA.

---
v4->v5:
* various improvements suggested by Wolfram

 drivers/watchdog/renesas_wdt.c | 27 ++++++++++++++++++++++-----
 1 file changed, 22 insertions(+), 5 deletions(-)
drivers/watchdog/renesas_wdt.c +22 -5
index 831ef83..e972b7b 100644
@@ -203,13 +203,29 @@ static int rwdt_remove(struct platform_device *pdev)
 	return 0;
 }
 
-/*
- * This driver does also fit for R-Car Gen2 (r8a779[0-4]) WDT. However, for SMP
- * to work there, one also needs a RESET (RST) driver which does not exist yet
- * due to HW issues. This needs to be solved before adding compatibles here.
- */
+static int __maybe_unused rwdt_suspend(struct device *dev)
+{
+	struct rwdt_priv *priv = dev_get_drvdata(dev);
+
+	if (watchdog_active(&priv->wdev))
+		rwdt_write(priv, priv->cks, RWTCSRA);
+	return 0;
+}
+
+static int __maybe_unused rwdt_resume(struct device *dev)
+{
+	struct rwdt_priv *priv = dev_get_drvdata(dev);
+
+	if (watchdog_active(&priv->wdev))
+		rwdt_write(priv, priv->cks | RWTCSRA_TME, RWTCSRA);
+	return 0;
+}
+
+static SIMPLE_DEV_PM_OPS(rwdt_pm_ops, rwdt_suspend, rwdt_resume);
+
 static const struct of_device_id rwdt_ids[] = {
 	{ .compatible = "renesas,rcar-gen3-wdt", },
+	{ .compatible = "renesas,rcar-gen2-wdt", },
 	{ /* sentinel */ }
 };
 MODULE_DEVICE_TABLE(of, rwdt_ids);
@@ -218,6 +234,7 @@ static struct platform_driver rwdt_driver = {
 	.driver = {
 		.name = "renesas_wdt",
 		.of_match_table = rwdt_ids,
+		.pm = &rwdt_pm_ops,
 	},
 	.probe = rwdt_probe,
 	.remove = rwdt_remove,
Signature · 1 line
2.7.4
maintainer
#28 lore raw

PATCH v5 12/26 watchdog: renesas_wdt: Add R-Car Gen2 support

Quoted 7 lines · Fabrizio Castro

Due to commits:
* "ARM: shmobile: Add watchdog support",
* "ARM: shmobile: rcar-gen2: Add watchdog support", and
* "soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2", we now have everything we needed for the watchdog to work on Gen2 and RZ/G1.

Quoted 7 lines

This commit adds "renesas,rcar-gen2-wdt" as compatible string for R-Car Gen2 and RZ/G1, and since on those platforms the rwdt clock needs to be always ON, when suspending to RAM we need to explicitly disable the counting by clearing TME from RWTCSRA.

Quoted 57 lines

---
v4->v5:
* various improvements suggested by Wolfram

 drivers/watchdog/renesas_wdt.c | 27 ++++++++++++++++++++++-----
 1 file changed, 22 insertions(+), 5 deletions(-)
drivers/watchdog/renesas_wdt.c +22 -5
index 831ef83..e972b7b 100644
@@ -203,13 +203,29 @@ static int rwdt_remove(struct platform_device *pdev)
 	return 0;
 }
 
-/*
- * This driver does also fit for R-Car Gen2 (r8a779[0-4]) WDT. However, for SMP
- * to work there, one also needs a RESET (RST) driver which does not exist yet
- * due to HW issues. This needs to be solved before adding compatibles here.
- */
+static int __maybe_unused rwdt_suspend(struct device *dev)
+{
+	struct rwdt_priv *priv = dev_get_drvdata(dev);
+
+	if (watchdog_active(&priv->wdev))
+		rwdt_write(priv, priv->cks, RWTCSRA);
+	return 0;
+}
+
+static int __maybe_unused rwdt_resume(struct device *dev)
+{
+	struct rwdt_priv *priv = dev_get_drvdata(dev);
+
+	if (watchdog_active(&priv->wdev))
+		rwdt_write(priv, priv->cks | RWTCSRA_TME, RWTCSRA);
+	return 0;
+}
+
+static SIMPLE_DEV_PM_OPS(rwdt_pm_ops, rwdt_suspend, rwdt_resume);
+
 static const struct of_device_id rwdt_ids[] = {
 	{ .compatible = "renesas,rcar-gen3-wdt", },
+	{ .compatible = "renesas,rcar-gen2-wdt", },
 	{ /* sentinel */ }
 };
 MODULE_DEVICE_TABLE(of, rwdt_ids);
@@ -218,6 +234,7 @@ static struct platform_driver rwdt_driver = {
 	.driver = {
 		.name = "renesas_wdt",
 		.of_match_table = rwdt_ids,
+		.pm = &rwdt_pm_ops,
 	},
 	.probe = rwdt_probe,
 	.remove = rwdt_remove,
Signature · 1 line
2.7.4
#29 lore raw

PATCH v5 12/26 watchdog: renesas_wdt: Add R-Car Gen2 support

In reply to Fabrizio Castro
Quoted 12 lines · Fabrizio Castro

Due to commits:
* "ARM: shmobile: Add watchdog support",
* "ARM: shmobile: rcar-gen2: Add watchdog support", and
* "soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2", we now have everything we needed for the watchdog to work on Gen2 and RZ/G1.

This commit adds "renesas,rcar-gen2-wdt" as compatible string for R-Car Gen2 and RZ/G1, and since on those platforms the rwdt clock needs to be always ON, when suspending to RAM we need to explicitly disable the counting by clearing TME from RWTCSRA.

Quoted 6 lines

---
v4->v5:
* various improvements suggested by Wolfram

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180212/961780c2/attachment.sig>

#30 lore raw

PATCH 1/1 watchdog: renesas_wdt: Blacklist early R-Car Gen2 SoCs

In reply to Fabrizio Castro

On early revisions of some R-Car Gen2 SoCs, and depending on SMP configuration, the system may fail to restart on watchdog time-out, and lock up instead.

Specifically:
- On R-Car H2 ES1.0 and M2-W ES1.0, watchdog restart fails unless only the first CPU core is in use (using e.g. the "maxcpus=1" kernel commandline option).
- On R-Car V2H ES1.1, watchdog restart fails unless SMP is disabled completely (using CONFIG_SMP=n during build configuration, or using the "nosmp" or "maxcpus=0" kernel commandline options).

Prevent using the watchdog in impacted cases by blacklisting the affected SoCs, using the minimum known working revisions (ES2.0 on R-Car H2, and ES3.0 on M2-W), and taking the actual SMP software configuration into account.

---
To be folded into Fabrizio Castro's "watchdog: renesas_wdt: Add R-Car Gen2 support".

Note that I cannot use IS_ENABLED(CONFIG_SMP), as setup_max_cpus does not exist for CONFIG_SMP=n.

Any reports on R-Car M2-W ES2.0 and V2H ES2.0+ are welcomed!

As the failure is due to an integration issue, and the watchdog itself is working fine, an alternative solution would be to move the check to the code that installs the reset trigger ("soc: renesas: rcar-rst:
Enable watchdog as reset trigger for Gen2").
However, doing so would mean that:
1. The user could enable and seemingly use the watchdog, but watchdog timeout would not restart the system,
2. The same check should be done before installing the new reset vector ("ARM: shmobile: rcar-gen2: Add watchdog support"), too, else onlining CPU0 would fail once the watchdog has timed out.

---
 drivers/watchdog/renesas_wdt.c | 43 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)
drivers/watchdog/renesas_wdt.c +43 -0
index 0f88614797c36022..4c8e8d2600a922a5 100644
@@ -16,6 +16,8 @@
 #include <linux/of.h>
 #include <linux/platform_device.h>
 #include <linux/pm_runtime.h>
+#include <linux/smp.h>
+#include <linux/sys_soc.h>
 #include <linux/watchdog.h>
 
 #define RWTCNT		0
@@ -131,6 +133,44 @@ static const struct watchdog_ops rwdt_ops = {
 	.restart = rwdt_restart,
 };
 
+#if defined(CONFIG_ARCH_RCAR_GEN2) && defined(CONFIG_SMP)
+/*
+ * Watchdog-reset integration is broken on early revisions of R-Car Gen2 SoCs
+ */
+static const struct soc_device_attribute rwdt_quirks_match[] = {
+	{
+		.soc_id = "r8a7790",
+		.revision = "ES1.*",
+		.data = (void *)1,	/* needs single CPU */
+	}, {
+		.soc_id = "r8a7791",
+		.revision = "ES[12].*",
+		.data = (void *)1,	/* needs single CPU */
+	}, {
+		.soc_id = "r8a7792",
+		.revision = "*",
+		.data = (void *)0,	/* needs SMP disabled */
+	},
+	{ /* sentinel */ }
+};
+
+static bool rwdt_blacklisted(struct device *dev)
+{
+	const struct soc_device_attribute *attr;
+
+	attr = soc_device_match(rwdt_quirks_match);
+	if (attr && setup_max_cpus > (uintptr_t)attr->data) {
+		dev_info(dev, "Watchdog blacklisted on %s %s\n", attr->soc_id,
+			 attr->revision);
+		return true;
+	}
+
+	return false;
+}
+#else /* !CONFIG_ARCH_RCAR_GEN2 || !CONFIG_SMP */
+static inline bool rwdt_blacklisted(struct device *dev) { return false; }
+#endif /* !CONFIG_ARCH_RCAR_GEN2 || !CONFIG_SMP */
+
 static int rwdt_probe(struct platform_device *pdev)
 {
 	struct rwdt_priv *priv;
@@ -139,6 +179,9 @@ static int rwdt_probe(struct platform_device *pdev)
 	unsigned long clks_per_sec;
 	int ret, i;
 
+	if (rwdt_blacklisted(&pdev->dev))
+		return -ENODEV;
+
 	priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
 	if (!priv)
 		return -ENOMEM;
Signature · 1 line
2.7.4
Reviewed-by on v1 1/1
#31 lore raw

Re: [PATCH] watchdog: renesas_wdt: Blacklist early R-Car Gen2 SoCs

Quoted 17 lines · Geert Uytterhoeven

On early revisions of some R-Car Gen2 SoCs, and depending on SMP configuration, the system may fail to restart on watchdog time-out, and lock up instead.

Specifically:
- On R-Car H2 ES1.0 and M2-W ES1.0, watchdog restart fails unless only the first CPU core is in use (using e.g. the "maxcpus=1" kernel commandline option).
- On R-Car V2H ES1.1, watchdog restart fails unless SMP is disabled completely (using CONFIG_SMP=n during build configuration, or using the "nosmp" or "maxcpus=0" kernel commandline options).

Prevent using the watchdog in impacted cases by blacklisting the affected SoCs, using the minimum known working revisions (ES2.0 on R-Car H2, and ES3.0 on M2-W), and taking the actual SMP software configuration into account.

Quoted 7 lines

---
To be folded into Fabrizio Castro's "watchdog: renesas_wdt: Add R-Car Gen2 support".

Note that I cannot use IS_ENABLED(CONFIG_SMP), as setup_max_cpus does not exist for CONFIG_SMP=n.

Thanks, I was going to ask about that.

Quoted 87 lines

Any reports on R-Car M2-W ES2.0 and V2H ES2.0+ are welcomed!

As the failure is due to an integration issue, and the watchdog itself is working fine, an alternative solution would be to move the check to the code that installs the reset trigger ("soc: renesas: rcar-rst:
Enable watchdog as reset trigger for Gen2").
However, doing so would mean that:
1. The user could enable and seemingly use the watchdog, but watchdog timeout would not restart the system,
2. The same check should be done before installing the new reset vector ("ARM: shmobile: rcar-gen2: Add watchdog support"), too, else onlining CPU0 would fail once the watchdog has timed out.

---
 drivers/watchdog/renesas_wdt.c | 43 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)
drivers/watchdog/renesas_wdt.c +43 -0
index 0f88614797c36022..4c8e8d2600a922a5 100644
@@ -16,6 +16,8 @@
 #include <linux/of.h>
 #include <linux/platform_device.h>
 #include <linux/pm_runtime.h>
+#include <linux/smp.h>
+#include <linux/sys_soc.h>
 #include <linux/watchdog.h>
 
 #define RWTCNT		0
@@ -131,6 +133,44 @@ static const struct watchdog_ops rwdt_ops = {
 	.restart = rwdt_restart,
 };
 
+#if defined(CONFIG_ARCH_RCAR_GEN2) && defined(CONFIG_SMP)
+/*
+ * Watchdog-reset integration is broken on early revisions of R-Car Gen2 SoCs
+ */
+static const struct soc_device_attribute rwdt_quirks_match[] = {
+	{
+		.soc_id = "r8a7790",
+		.revision = "ES1.*",
+		.data = (void *)1,	/* needs single CPU */
+	}, {
+		.soc_id = "r8a7791",
+		.revision = "ES[12].*",
+		.data = (void *)1,	/* needs single CPU */
+	}, {
+		.soc_id = "r8a7792",
+		.revision = "*",
+		.data = (void *)0,	/* needs SMP disabled */
+	},
+	{ /* sentinel */ }
+};
+
+static bool rwdt_blacklisted(struct device *dev)
+{
+	const struct soc_device_attribute *attr;
+
+	attr = soc_device_match(rwdt_quirks_match);
+	if (attr && setup_max_cpus > (uintptr_t)attr->data) {
+		dev_info(dev, "Watchdog blacklisted on %s %s\n", attr->soc_id,
+			 attr->revision);
+		return true;
+	}
+
+	return false;
+}
+#else /* !CONFIG_ARCH_RCAR_GEN2 || !CONFIG_SMP */
+static inline bool rwdt_blacklisted(struct device *dev) { return false; }
+#endif /* !CONFIG_ARCH_RCAR_GEN2 || !CONFIG_SMP */
+
 static int rwdt_probe(struct platform_device *pdev)
 {
 	struct rwdt_priv *priv;
@@ -139,6 +179,9 @@ static int rwdt_probe(struct platform_device *pdev)
 	unsigned long clks_per_sec;
 	int ret, i;
 
+	if (rwdt_blacklisted(&pdev->dev))
+		return -ENODEV;
+
 	priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
 	if (!priv)
 		return -ENOMEM;
Signature · 1 line
2.7.4
author Acked-by on v1 1/1
#32 lore raw

RE: [PATCH] watchdog: renesas_wdt: Blacklist early R-Car Gen2 SoCs

In reply to Geert Uytterhoeven
Quoted 14 lines

Subject: [PATCH] watchdog: renesas_wdt: Blacklist early R-Car Gen2 SoCs

On early revisions of some R-Car Gen2 SoCs, and depending on SMP configuration, the system may fail to restart on watchdog time-out, and lock up instead.

Specifically:
- On R-Car H2 ES1.0 and M2-W ES1.0, watchdog restart fails unless only the first CPU core is in use (using e.g. the "maxcpus=1" kernel commandline option).
- On R-Car V2H ES1.1, watchdog restart fails unless SMP is disabled completely (using CONFIG_SMP=n during build configuration, or using the "nosmp" or "maxcpus=0" kernel commandline options).

Quoted 6 lines

Prevent using the watchdog in impacted cases by blacklisting the affected SoCs, using the minimum known working revisions (ES2.0 on R-Car H2, and ES3.0 on M2-W), and taking the actual SMP software configuration into account.

Quoted 92 lines

---
To be folded into Fabrizio Castro's "watchdog: renesas_wdt: Add R-Car Gen2 support".

Note that I cannot use IS_ENABLED(CONFIG_SMP), as setup_max_cpus does not exist for CONFIG_SMP=n.

Any reports on R-Car M2-W ES2.0 and V2H ES2.0+ are welcomed!

As the failure is due to an integration issue, and the watchdog itself is working fine, an alternative solution would be to move the check to the code that installs the reset trigger ("soc: renesas: rcar-rst:
Enable watchdog as reset trigger for Gen2").
However, doing so would mean that:
1. The user could enable and seemingly use the watchdog, but watchdog timeout would not restart the system,
2. The same check should be done before installing the new reset vector ("ARM: shmobile: rcar-gen2: Add watchdog support"), too, else onlining CPU0 would fail once the watchdog has timed out.

---
 drivers/watchdog/renesas_wdt.c | 43 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)
drivers/watchdog/renesas_wdt.c +43 -1
index 0f88614797c36022..4c8e8d2600a922a5 100644
@@ -16,6 +16,8 @@
 #include <linux/of.h>
 #include <linux/platform_device.h>
 #include <linux/pm_runtime.h>
+#include <linux/smp.h>
+#include <linux/sys_soc.h>
 #include <linux/watchdog.h>
 
 #define RWTCNT0
@@ -131,6 +133,44 @@ static const struct watchdog_ops rwdt_ops = {
 .restart = rwdt_restart,
 };
 
+#if defined(CONFIG_ARCH_RCAR_GEN2) && defined(CONFIG_SMP)
+/*
+ * Watchdog-reset integration is broken on early revisions of R-Car Gen2 SoCs
+ */
+static const struct soc_device_attribute rwdt_quirks_match[] = {
+{
+.soc_id = "r8a7790",
+.revision = "ES1.*",
+.data = (void *)1,/* needs single CPU */
+}, {
+.soc_id = "r8a7791",
+.revision = "ES[12].*",
+.data = (void *)1,/* needs single CPU */
+}, {
+.soc_id = "r8a7792",
+.revision = "*",
+.data = (void *)0,/* needs SMP disabled */
+},
+{ /* sentinel */ }
+};
+
+static bool rwdt_blacklisted(struct device *dev)
+{
+const struct soc_device_attribute *attr;
+
+attr = soc_device_match(rwdt_quirks_match);
+if (attr && setup_max_cpus > (uintptr_t)attr->data) {
+dev_info(dev, "Watchdog blacklisted on %s %s\n", attr->soc_id,
+ attr->revision);
+return true;
+}
+
+return false;
+}
+#else /* !CONFIG_ARCH_RCAR_GEN2 || !CONFIG_SMP */
+static inline bool rwdt_blacklisted(struct device *dev) { return false; }
+#endif /* !CONFIG_ARCH_RCAR_GEN2 || !CONFIG_SMP */
+
 static int rwdt_probe(struct platform_device *pdev)
 {
 struct rwdt_priv *priv;
@@ -139,6 +179,9 @@ static int rwdt_probe(struct platform_device *pdev)
 unsigned long clks_per_sec;
 int ret, i;
 
+if (rwdt_blacklisted(&pdev->dev))
+return -ENODEV;
+
 priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
 if (!priv)
 return -ENOMEM;
--
Quoted 1 line

2.7.4

Renesas Electronics Europe Ltd, Dukes Meadow, Millboard Road, Bourne End, Buckinghamshire, SL8 5FH, UK. Registered in England & Wales under Registered No. 04586709.

#33 lore raw

PATCH v5 12/26 watchdog: renesas_wdt: Add R-Car Gen2 support

In reply to Fabrizio Castro

Hi Fabrizio,

Quoted 7 lines · Fabrizio Castro

Due to commits:
* "ARM: shmobile: Add watchdog support",
* "ARM: shmobile: rcar-gen2: Add watchdog support", and
* "soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2", we now have everything we needed for the watchdog to work on Gen2 and RZ/G1.

Quoted 7 lines

This commit adds "renesas,rcar-gen2-wdt" as compatible string for R-Car Gen2 and RZ/G1, and since on those platforms the rwdt clock needs to be always ON, when suspending to RAM we need to explicitly disable the counting by clearing TME from RWTCSRA.

Thanks for your patch!

I verified this works on R-Car Gen2, so

Still, more comments below...

Quoted 16 lines
drivers/watchdog/renesas_wdt.c +5 -5
@@ -203,13 +203,29 @@ static int rwdt_remove(struct platform_device *pdev)
        return 0;
 }
 
-/*
- * This driver does also fit for R-Car Gen2 (r8a779[0-4]) WDT. However, for SMP
- * to work there, one also needs a RESET (RST) driver which does not exist yet
- * due to HW issues. This needs to be solved before adding compatibles here.
- */
+static int __maybe_unused rwdt_suspend(struct device *dev)
+{
+       struct rwdt_priv *priv = dev_get_drvdata(dev);
+
+       if (watchdog_active(&priv->wdev))

drivers/watchdog/renesas_wdt.c static int rwdt_remove(struct platform_device *pdev) · line 220

Quoted 10 lines
drivers/watchdog/renesas_wdt.c +10 -0
+               rwdt_write(priv, priv->cks, RWTCSRA);
+       return 0;
+}
+
+static int __maybe_unused rwdt_resume(struct device *dev)
+{
+       struct rwdt_priv *priv = dev_get_drvdata(dev);
+
+       if (watchdog_active(&priv->wdev))
+               rwdt_write(priv, priv->cks | RWTCSRA_TME, RWTCSRA);

Writing to this register is not sufficient on R-Car Gen3, where PSCI suspend powers down the whole SoC. Hence all WDT register content is lost, causing the watchdog timeout never to trigger.
Note that this issue is pre-existing, and not caused by your patch.

This can be fixed by replacing the RWTCSRA register writes in the suspend/resume handlers by calls to rwdt_stop() resp. rwdt_start(), like is done in the BSP in commit e406980763f18f38 ("watchdog: renesas-wdt: Support the suspend/resume"). Note that this would cause a small change in behavior on R-Car Gen2, where the timeout would be reset on resume, instead of continuing where stopped before. I don't think that hurts, though.

Since I was always a bit uncomfortable with this patch doing two things at once (1. suspend/resume, 2. "renesas,rcar-gen2-wdt" matching), I think it would be better to take the patch from the BSP first, and add support for "renesas,rcar-gen2-wdt" in a subsequent patch.

Does the above make sense?
Do you agree?

Thanks!

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#34 lore raw

PATCH v5 12/26 watchdog: renesas_wdt: Add R-Car Gen2 support

Hi Geert,

thank you for your feedback!

Quoted 60 lines

Subject: Re: [PATCH v5 12/26] watchdog: renesas_wdt: Add R-Car Gen2 support

Hi Fabrizio,

Quoted 7 lines · Fabrizio Castro

Due to commits:
* "ARM: shmobile: Add watchdog support",
* "ARM: shmobile: rcar-gen2: Add watchdog support", and
* "soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2", we now have everything we needed for the watchdog to work on Gen2 and RZ/G1.

Quoted 7 lines

This commit adds "renesas,rcar-gen2-wdt" as compatible string for R-Car Gen2 and RZ/G1, and since on those platforms the rwdt clock needs to be always ON, when suspending to RAM we need to explicitly disable the counting by clearing TME from RWTCSRA.

Thanks for your patch!

I verified this works on R-Car Gen2, so

Still, more comments below...

Quoted 16 lines
drivers/watchdog/renesas_wdt.c +5 -5
@@ -203,13 +203,29 @@ static int rwdt_remove(struct platform_device *pdev)
        return 0;
 }
 
-/*
- * This driver does also fit for R-Car Gen2 (r8a779[0-4]) WDT. However, for SMP
- * to work there, one also needs a RESET (RST) driver which does not exist yet
- * due to HW issues. This needs to be solved before adding compatibles here.
- */
+static int __maybe_unused rwdt_suspend(struct device *dev)
+{
+       struct rwdt_priv *priv = dev_get_drvdata(dev);
+
+       if (watchdog_active(&priv->wdev))

drivers/watchdog/renesas_wdt.c static int rwdt_remove(struct platform_device *pdev) · line 220

Quoted 10 lines
drivers/watchdog/renesas_wdt.c +10 -0
+               rwdt_write(priv, priv->cks, RWTCSRA);
+       return 0;
+}
+
+static int __maybe_unused rwdt_resume(struct device *dev)
+{
+       struct rwdt_priv *priv = dev_get_drvdata(dev);
+
+       if (watchdog_active(&priv->wdev))
+               rwdt_write(priv, priv->cks | RWTCSRA_TME, RWTCSRA);

Writing to this register is not sufficient on R-Car Gen3, where PSCI suspend powers down the whole SoC. Hence all WDT register content is lost, causing the watchdog timeout never to trigger.
Note that this issue is pre-existing, and not caused by your patch.

Quoted 6 lines

This can be fixed by replacing the RWTCSRA register writes in the suspend/resume handlers by calls to rwdt_stop() resp. rwdt_start(), like is done in the BSP in commit e406980763f18f38 ("watchdog: renesas-wdt: Support the suspend/resume"). Note that this would cause a small change in behavior on R-Car Gen2, where the timeout would be reset on resume, instead of continuing where stopped before. I don't think that hurts, though.

I see, well I believe we can make both of us happy by addressing Gen3 problems and preserving the original behaviour from Gen2.

I am going to send a new series to address this shortly.

Quoted 5 lines

Since I was always a bit uncomfortable with this patch doing two things at once (1. suspend/resume, 2. "renesas,rcar-gen2-wdt" matching), I think it would be better to take the patch from the BSP first, and add support for "renesas,rcar-gen2-wdt" in a subsequent patch.

I agree, I'll split the patch and submit suspend/resume first, Gen2 compatibility next.

Thanks,
Fab

Quoted 8 lines

Does the above make sense?
Do you agree?

Thanks!

Gr{oetje,eeting}s,

Quoted 8 lines

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

Renesas Electronics Europe Ltd, Dukes Meadow, Millboard Road, Bourne End, Buckinghamshire, SL8 5FH, UK. Registered in England & Wales under Registered No. 04586709.

#35 lore raw

PATCH v5 13/26 watchdog: renesas_wdt: Add restart handler

In reply to Fabrizio Castro

On iWave's boards iwg20d and iwg22d the only way to reboot the system is by means of the watchdog.
This patch adds a restart handler to rwdt_ops, and also makes sure we keep its priority to a medium level, in order to not override other more effective handlers.

---
v4->v5:
* changed rwdt_restart implementation as per Wolfram comments

 drivers/watchdog/renesas_wdt.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
drivers/watchdog/renesas_wdt.c +12 -0
index e972b7b..0f88614 100644
@@ -107,6 +107,16 @@ static unsigned int rwdt_get_timeleft(struct watchdog_device *wdev)
 	return DIV_BY_CLKS_PER_SEC(priv, 65536 - val);
 }
 
+static int rwdt_restart(struct watchdog_device *wdev, unsigned long action,
+			void *data)
+{
+	struct rwdt_priv *priv = watchdog_get_drvdata(wdev);
+
+	rwdt_start(wdev);
+	rwdt_write(priv, 0xffff, RWTCNT);
+	return 0;
+}
+
 static const struct watchdog_info rwdt_ident = {
 	.options = WDIOF_MAGICCLOSE | WDIOF_KEEPALIVEPING | WDIOF_SETTIMEOUT,
 	.identity = "Renesas WDT Watchdog",
@@ -118,6 +128,7 @@ static const struct watchdog_ops rwdt_ops = {
 	.stop = rwdt_stop,
 	.ping = rwdt_init_timeout,
 	.get_timeleft = rwdt_get_timeleft,
+	.restart = rwdt_restart,
 };
 
 static int rwdt_probe(struct platform_device *pdev)
@@ -176,6 +187,7 @@ static int rwdt_probe(struct platform_device *pdev)
 	platform_set_drvdata(pdev, priv);
 	watchdog_set_drvdata(&priv->wdev, priv);
 	watchdog_set_nowayout(&priv->wdev, nowayout);
+	watchdog_set_restart_priority(&priv->wdev, 0);
 
 	/* This overrides the default timeout only if DT configuration was found */
 	ret = watchdog_init_timeout(&priv->wdev, 0, &pdev->dev);
Signature · 1 line
2.7.4
#36 lore raw

PATCH v5 13/26 watchdog: renesas_wdt: Add restart handler

Quoted 9 lines · Fabrizio Castro

On iWave's boards iwg20d and iwg22d the only way to reboot the system is by means of the watchdog.
This patch adds a restart handler to rwdt_ops, and also makes sure we keep its priority to a medium level, in order to not override other more effective handlers.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180212/9ad7f346/attachment.sig>

#37 lore raw

PATCH v5 14/26 ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN

In reply to Fabrizio Castro

R-Car Gen2 and RZ/G1 platforms come with a watchdog IP, therefore enable its driver by default.

---
v4->v5:
* no change

 arch/arm/configs/shmobile_defconfig | 1 +
 1 file changed, 1 insertion(+)
arch/arm/configs/shmobile_defconfig +1 -0
index 578434c..d5cdad8 100644
@@ -132,6 +132,7 @@ CONFIG_CPU_THERMAL=y
 CONFIG_RCAR_THERMAL=y
 CONFIG_WATCHDOG=y
 CONFIG_DA9063_WATCHDOG=y
+CONFIG_RENESAS_WDT=y
 CONFIG_MFD_AS3711=y
 CONFIG_MFD_DA9063=y
 CONFIG_REGULATOR_FIXED_VOLTAGE=y
Signature · 1 line
2.7.4
#38 lore raw

PATCH v5 14/26 ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN

Hello Simon,

what do you think about enabling the watchdog in multi_v7_defconfig as a module?

Thanks,
Fab

Quoted 28 lines

Subject: [PATCH v5 14/26] ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN

R-Car Gen2 and RZ/G1 platforms come with a watchdog IP, therefore enable its driver by default.

---
v4->v5:
* no change

 arch/arm/configs/shmobile_defconfig | 1 +
 1 file changed, 1 insertion(+)
arch/arm/configs/shmobile_defconfig +1 -1
index 578434c..d5cdad8 100644
@@ -132,6 +132,7 @@ CONFIG_CPU_THERMAL=y
 CONFIG_RCAR_THERMAL=y
 CONFIG_WATCHDOG=y
 CONFIG_DA9063_WATCHDOG=y
+CONFIG_RENESAS_WDT=y
 CONFIG_MFD_AS3711=y
 CONFIG_MFD_DA9063=y
 CONFIG_REGULATOR_FIXED_VOLTAGE=y
--
Quoted 1 line

2.7.4

Renesas Electronics Europe Ltd, Dukes Meadow, Millboard Road, Bourne End, Buckinghamshire, SL8 5FH, UK. Registered in England & Wales under Registered No. 04586709.

#39 lore raw

PATCH v5 14/26 ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN

Quoted 3 lines · Fabrizio Castro

Hello Simon,

what do you think about enabling the watchdog in multi_v7_defconfig as a module?

I think its a good idea.

#41 lore raw

PATCH v5 15/26 clk: renesas: r8a7743: Add rwdt clock

In reply to Fabrizio Castro

Add "rwdt" clock to r8a7743_mod_clks. Also, since we may need to access the watchdog registers at any time, declare the clock as critical.

---
v4->v5:
* no change

 drivers/clk/renesas/r8a7743-cpg-mssr.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/clk/renesas/r8a7743-cpg-mssr.c +2 -0
index 6dc0b30..d3c8b1e 100644
@@ -117,6 +117,7 @@ static const struct mssr_mod_clk r8a7743_mod_clks[] __initconst = {
 	DEF_MOD("cmt1",			 329,	R8A7743_CLK_R),
 	DEF_MOD("usbhs-dmac0",		 330,	R8A7743_CLK_HP),
 	DEF_MOD("usbhs-dmac1",		 331,	R8A7743_CLK_HP),
+	DEF_MOD("rwdt",			 402,	R8A7743_CLK_R),
 	DEF_MOD("irqc",			 407,	R8A7743_CLK_CP),
 	DEF_MOD("intc-sys",		 408,	R8A7743_CLK_ZS),
 	DEF_MOD("audio-dmac1",		 501,	R8A7743_CLK_HP),
@@ -195,6 +196,7 @@ static const struct mssr_mod_clk r8a7743_mod_clks[] __initconst = {
 };
 
 static const unsigned int r8a7743_crit_mod_clks[] __initconst = {
+	MOD_CLK_ID(402),	/* RWDT */
 	MOD_CLK_ID(408),	/* INTC-SYS (GIC) */
 };
 
Signature · 1 line
2.7.4
#42 lore raw

PATCH v5 16/26 clk: renesas: r8a7745: Add rwdt clock

In reply to Fabrizio Castro

Add "rwdt" clock to r8a7745_mod_clks. Also, since we may need to access the watchdog registers at any time, declare the clock as critical.

---
v4->v5:
* no change

 drivers/clk/renesas/r8a7745-cpg-mssr.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/clk/renesas/r8a7745-cpg-mssr.c +2 -0
index 2859504..87f5a36 100644
@@ -114,6 +114,7 @@ static const struct mssr_mod_clk r8a7745_mod_clks[] __initconst = {
 	DEF_MOD("cmt1",			 329,	R8A7745_CLK_R),
 	DEF_MOD("usbhs-dmac0",		 330,	R8A7745_CLK_HP),
 	DEF_MOD("usbhs-dmac1",		 331,	R8A7745_CLK_HP),
+	DEF_MOD("rwdt",			 402,	R8A7745_CLK_R),
 	DEF_MOD("irqc",			 407,	R8A7745_CLK_CP),
 	DEF_MOD("intc-sys",		 408,	R8A7745_CLK_ZS),
 	DEF_MOD("audio-dmac0",		 502,	R8A7745_CLK_HP),
@@ -180,6 +181,7 @@ static const struct mssr_mod_clk r8a7745_mod_clks[] __initconst = {
 };
 
 static const unsigned int r8a7745_crit_mod_clks[] __initconst = {
+	MOD_CLK_ID(402),	/* RWDT */
 	MOD_CLK_ID(408),	/* INTC-SYS (GIC) */
 };
 
Signature · 1 line
2.7.4
#43 lore raw

PATCH v5 17/26 clk: renesas: r8a7790: Add rwdt clock

In reply to Fabrizio Castro

Add "rwdt" clock to r8a7790_mod_clks. Also, since we may need to access the watchdog registers at any time, declare the clock as critical.

---
v4->v5:
* no change

 drivers/clk/renesas/r8a7790-cpg-mssr.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/clk/renesas/r8a7790-cpg-mssr.c +2 -0
index 46bb55b..f936cb7 100644
@@ -140,6 +140,7 @@ static const struct mssr_mod_clk r8a7790_mod_clks[] __initconst = {
 	DEF_MOD("cmt1",			 329,	R8A7790_CLK_R),
 	DEF_MOD("usbhs-dmac0",		 330,	R8A7790_CLK_HP),
 	DEF_MOD("usbhs-dmac1",		 331,	R8A7790_CLK_HP),
+	DEF_MOD("rwdt",			 402,	R8A7790_CLK_R),
 	DEF_MOD("irqc",			 407,	R8A7790_CLK_CP),
 	DEF_MOD("intc-sys",		 408,	R8A7790_CLK_ZS),
 	DEF_MOD("audio-dmac1",		 501,	R8A7790_CLK_HP),
@@ -211,6 +212,7 @@ static const struct mssr_mod_clk r8a7790_mod_clks[] __initconst = {
 };
 
 static const unsigned int r8a7790_crit_mod_clks[] __initconst = {
+	MOD_CLK_ID(402),	/* RWDT */
 	MOD_CLK_ID(408),	/* INTC-SYS (GIC) */
 };
 
Signature · 1 line
2.7.4
#44 lore raw

PATCH v5 18/26 clk: renesas: r8a7791/r8a7793: Add rwdt clock

In reply to Fabrizio Castro

Add "rwdt" clock to r8a7791_mod_clks. Also, since we may need to access the watchdog registers at any time, declare the clock as critical.

---
v4->v5:
* no change

 drivers/clk/renesas/r8a7791-cpg-mssr.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/clk/renesas/r8a7791-cpg-mssr.c +2 -0
index c0b51f9..820b220 100644
@@ -128,6 +128,7 @@ static const struct mssr_mod_clk r8a7791_mod_clks[] __initconst = {
 	DEF_MOD("cmt1",			 329,	R8A7791_CLK_R),
 	DEF_MOD("usbhs-dmac0",		 330,	R8A7791_CLK_HP),
 	DEF_MOD("usbhs-dmac1",		 331,	R8A7791_CLK_HP),
+	DEF_MOD("rwdt",			 402,	R8A7791_CLK_R),
 	DEF_MOD("irqc",			 407,	R8A7791_CLK_CP),
 	DEF_MOD("intc-sys",		 408,	R8A7791_CLK_ZS),
 	DEF_MOD("audio-dmac1",		 501,	R8A7791_CLK_HP),
@@ -209,6 +210,7 @@ static const struct mssr_mod_clk r8a7791_mod_clks[] __initconst = {
 };
 
 static const unsigned int r8a7791_crit_mod_clks[] __initconst = {
+	MOD_CLK_ID(402),	/* RWDT */
 	MOD_CLK_ID(408),	/* INTC-SYS (GIC) */
 };
 
Signature · 1 line
2.7.4
#45 lore raw

PATCH v5 19/26 clk: renesas: r8a7794: Add rwdt clock

In reply to Fabrizio Castro

Add "rwdt" clock to r8a7794_mod_clks. Also, since we may need to access the watchdog registers at any time, declare the clock as critical.

---
v4->v5:
* no change

 drivers/clk/renesas/r8a7794-cpg-mssr.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/clk/renesas/r8a7794-cpg-mssr.c +2 -0
index ec091a4..2a40bbe 100644
@@ -121,6 +121,7 @@ static const struct mssr_mod_clk r8a7794_mod_clks[] __initconst = {
 	DEF_MOD("cmt1",			 329,	R8A7794_CLK_R),
 	DEF_MOD("usbhs-dmac0",		 330,	R8A7794_CLK_HP),
 	DEF_MOD("usbhs-dmac1",		 331,	R8A7794_CLK_HP),
+	DEF_MOD("rwdt",			 402,	R8A7794_CLK_R),
 	DEF_MOD("irqc",			 407,	R8A7794_CLK_CP),
 	DEF_MOD("intc-sys",		 408,	R8A7794_CLK_ZS),
 	DEF_MOD("audio-dmac0",		 502,	R8A7794_CLK_HP),
@@ -190,6 +191,7 @@ static const struct mssr_mod_clk r8a7794_mod_clks[] __initconst = {
 };
 
 static const unsigned int r8a7794_crit_mod_clks[] __initconst = {
+	MOD_CLK_ID(402),	/* RWDT */
 	MOD_CLK_ID(408),	/* INTC-SYS (GIC) */
 };
 
Signature · 1 line
2.7.4
#46 lore raw

PATCH v5 20/26 ARM: dts: r8a7743: Add watchdog support to SoC dtsi

In reply to Fabrizio Castro

This patch adds watchdog support to the r8a7743 SoC dtsi.

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7743.dtsi | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/arm/boot/dts/r8a7743.dtsi +10 -0
index c2943d7..1d4505f 100644
@@ -368,6 +368,16 @@
 			reg = <0 0xe6160000 0 0x100>;
 		};
 
+		rwdt: watchdog at e6020000 {
+			compatible = "renesas,r8a7743-wdt",
+				     "renesas,rcar-gen2-wdt";
+			reg = <0 0xe6020000 0 0x0c>;
+			clocks = <&cpg CPG_MOD 402>;
+			power-domains = <&sysc R8A7743_PD_ALWAYS_ON>;
+			resets = <&cpg 402>;
+			status = "disabled";
+		};
+
 		sysc: system-controller at e6180000 {
 			compatible = "renesas,r8a7743-sysc";
 			reg = <0 0xe6180000 0 0x200>;
Signature · 1 line
2.7.4
#47 lore raw

PATCH v5 21/26 ARM: dts: r8a7745: Add watchdog support to SoC dtsi

In reply to Fabrizio Castro

This patch adds watchdog support to the r8a7745 SoC dtsi.

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7745.dtsi | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/arm/boot/dts/r8a7745.dtsi +10 -0
index 6bc1cf4..5d1b8f2 100644
@@ -321,6 +321,16 @@
 			reg = <0 0xe6160000 0 0x100>;
 		};
 
+		rwdt: watchdog at e6020000 {
+			compatible = "renesas,r8a7745-wdt",
+				     "renesas,rcar-gen2-wdt";
+			reg = <0 0xe6020000 0 0x0c>;
+			clocks = <&cpg CPG_MOD 402>;
+			power-domains = <&sysc R8A7745_PD_ALWAYS_ON>;
+			resets = <&cpg 402>;
+			status = "disabled";
+		};
+
 		sysc: system-controller at e6180000 {
 			compatible = "renesas,r8a7745-sysc";
 			reg = <0 0xe6180000 0 0x200>;
Signature · 1 line
2.7.4
#48 lore raw

PATCH v5 22/26 ARM: dts: r8a7790: Add watchdog support to SoC dtsi

In reply to Fabrizio Castro

This commit adds watchdog support to the r8a7790 dtsi.

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7790.dtsi | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/arm/boot/dts/r8a7790.dtsi +10 -0
index a84bf35..83d6704 100644
@@ -218,6 +218,16 @@
 		#size-cells = <2>;
 		ranges;
 
+		rwdt: watchdog at e6020000 {
+			compatible = "renesas,r8a7790-wdt",
+				     "renesas,rcar-gen2-wdt";
+			reg = <0 0xe6020000 0 0x0c>;
+			clocks = <&cpg CPG_MOD 402>;
+			power-domains = <&sysc R8A7790_PD_ALWAYS_ON>;
+			resets = <&cpg 402>;
+			status = "disabled";
+		};
+
 		gpio0: gpio at e6050000 {
 			compatible = "renesas,gpio-r8a7790",
 				     "renesas,rcar-gen2-gpio";
Signature · 1 line
2.7.4
#49 lore raw

PATCH v5 23/26 ARM: dts: r8a7791: Add watchdog support to SoC dtsi

In reply to Fabrizio Castro

This commit adds watchdog support to the r8a7791 dtsi.

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7791.dtsi | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/arm/boot/dts/r8a7791.dtsi +10 -0
index d2a4546..80e20ecd 100644
@@ -142,6 +142,16 @@
 		#size-cells = <2>;
 		ranges;
 
+		rwdt: watchdog at e6020000 {
+			compatible = "renesas,r8a7791-wdt",
+				     "renesas,rcar-gen2-wdt";
+			reg = <0 0xe6020000 0 0x0c>;
+			clocks = <&cpg CPG_MOD 402>;
+			power-domains = <&sysc R8A7791_PD_ALWAYS_ON>;
+			resets = <&cpg 402>;
+			status = "disabled";
+		};
+
 		gpio0: gpio at e6050000 {
 			compatible = "renesas,gpio-r8a7791",
 				     "renesas,rcar-gen2-gpio";
Signature · 1 line
2.7.4
#50 lore raw

PATCH v5 24/26 ARM: dts: r8a7794: Add watchdog support to SoC dtsi

In reply to Fabrizio Castro

This commit adds watchdog support to the r8a7794 dtsi.

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7794.dtsi | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/arm/boot/dts/r8a7794.dtsi +10 -0
index e47c5f4..1c9b149 100644
@@ -119,6 +119,16 @@
 		#size-cells = <2>;
 		ranges;
 
+		rwdt: watchdog at e6020000 {
+			compatible = "renesas,r8a7794-wdt",
+				     "renesas,rcar-gen2-wdt";
+			reg = <0 0xe6020000 0 0x0c>;
+			clocks = <&cpg CPG_MOD 402>;
+			power-domains = <&sysc R8A7794_PD_ALWAYS_ON>;
+			resets = <&cpg 402>;
+			status = "disabled";
+		};
+
 		gpio0: gpio at e6050000 {
 			compatible = "renesas,gpio-r8a7794",
 				     "renesas,rcar-gen2-gpio";
Signature · 1 line
2.7.4
#51 lore raw

PATCH v5 25/26 ARM: dts: iwg20m: Add watchdog support to SoM dtsi

In reply to Fabrizio Castro

This patch enables the watchdog from within the iwg20m SoM dtsi.

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7743-iwg20m.dtsi | 5 +++++
 1 file changed, 5 insertions(+)
arch/arm/boot/dts/r8a7743-iwg20m.dtsi +5 -0
index 1d3e950..d364685 100644
@@ -91,6 +91,11 @@
 	};
 };
 
+&rwdt {
+	timeout-sec = <60>;
+	status = "okay";
+};
+
 &sdhi0 {
 	pinctrl-0 = <&sdhi0_pins>;
 	pinctrl-names = "default";
Signature · 1 line
2.7.4
#52 lore raw

PATCH v5 26/26 ARM: dts: iwg22m: Add watchdog support to SoM dtsi

In reply to Fabrizio Castro

This patch enables the watchdog from within the iwg20m SoM dtsi.

---
v4->v5:
* no change

 arch/arm/boot/dts/r8a7745-iwg22m.dtsi | 5 +++++
 1 file changed, 5 insertions(+)
arch/arm/boot/dts/r8a7745-iwg22m.dtsi +5 -0
index 8d0a392b..29b6e10 100644
@@ -91,6 +91,11 @@
 	};
 };
 
+&rwdt {
+	timeout-sec = <60>;
+	status = "okay";
+};
+
 &sdhi1 {
 	pinctrl-0 = <&sdhi1_pins>;
 	pinctrl-names = "default";
Signature · 1 line
2.7.4
#53 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

In reply to Fabrizio Castro
Quoted 62 lines · Fabrizio Castro

Dear All,

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock

Quoted 8 lines

clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

Thanks, I have applied the above "dts" and "defconfig" patches for v4.17.

#54 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

In reply to Fabrizio Castro
Quoted 83 lines · Fabrizio Castro

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock
clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

 .../devicetree/bindings/watchdog/renesas-wdt.txt   | 19 ++++++--
 arch/arm/boot/dts/r8a7743-iwg20m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7743.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7745-iwg22m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7745.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7790.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7791.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7792.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7793.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7794.dtsi                     | 12 ++++-
 arch/arm/configs/shmobile_defconfig                |  1 +
 arch/arm/mach-shmobile/common.h                    |  6 +++
 arch/arm/mach-shmobile/headsmp.S                   | 55 ++++++++++++++++++++++
 arch/arm/mach-shmobile/platsmp-apmu.c              |  1 +
Quoted 8 lines
 arch/arm/mach-shmobile/pm-rcar-gen2.c              | 15 ++++--
 drivers/clk/renesas/r8a7743-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7745-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7790-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7791-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7794-cpg-mssr.c             |  2 +
 drivers/soc/renesas/rcar-rst.c                     | 35 +++++++++++---
 drivers/watchdog/renesas_wdt.c                     | 39 +++++++++++++--

Thanks, I've queued the clock patches in clk-renesas-for-v4.17, as they are a hard dependency for:
(1) The new reset vector in arch/arm/mach-shmobile/
(2) The watchdog driver.

Note that the watchdog driver itself (2) has a hard dependency on the new reset vector (1).

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#55 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

Quoted 94 lines · Geert Uytterhoeven
Quoted 83 lines · Fabrizio Castro

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock
clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

 .../devicetree/bindings/watchdog/renesas-wdt.txt   | 19 ++++++--
 arch/arm/boot/dts/r8a7743-iwg20m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7743.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7745-iwg22m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7745.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7790.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7791.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7792.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7793.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7794.dtsi                     | 12 ++++-
 arch/arm/configs/shmobile_defconfig                |  1 +
 arch/arm/mach-shmobile/common.h                    |  6 +++
 arch/arm/mach-shmobile/headsmp.S                   | 55 ++++++++++++++++++++++
 arch/arm/mach-shmobile/platsmp-apmu.c              |  1 +
Quoted 8 lines
 arch/arm/mach-shmobile/pm-rcar-gen2.c              | 15 ++++--
 drivers/clk/renesas/r8a7743-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7745-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7790-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7791-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7794-cpg-mssr.c             |  2 +
 drivers/soc/renesas/rcar-rst.c                     | 35 +++++++++++---
 drivers/watchdog/renesas_wdt.c                     | 39 +++++++++++++--
Quoted 7 lines

Thanks, I've queued the clock patches in clk-renesas-for-v4.17, as they are a hard dependency for:
(1) The new reset vector in arch/arm/mach-shmobile/
(2) The watchdog driver.

Note that the watchdog driver itself (2) has a hard dependency on the new reset vector (1).

Thanks, I have marked the watchdog driver patch as Deferred. Please resubmit or otherwise ping me once they dependencies are in an rc release.

#56 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

Hi Simon,

Quoted 100 lines · Simon Horman
Quoted 99 lines · Geert Uytterhoeven
Quoted 83 lines · Fabrizio Castro

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock
clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

 .../devicetree/bindings/watchdog/renesas-wdt.txt   | 19 ++++++--
 arch/arm/boot/dts/r8a7743-iwg20m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7743.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7745-iwg22m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7745.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7790.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7791.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7792.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7793.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7794.dtsi                     | 12 ++++-
 arch/arm/configs/shmobile_defconfig                |  1 +
 arch/arm/mach-shmobile/common.h                    |  6 +++
 arch/arm/mach-shmobile/headsmp.S                   | 55 ++++++++++++++++++++++
 arch/arm/mach-shmobile/platsmp-apmu.c              |  1 +
Quoted 8 lines
 arch/arm/mach-shmobile/pm-rcar-gen2.c              | 15 ++++--
 drivers/clk/renesas/r8a7743-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7745-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7790-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7791-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7794-cpg-mssr.c             |  2 +
 drivers/soc/renesas/rcar-rst.c                     | 35 +++++++++++---
 drivers/watchdog/renesas_wdt.c                     | 39 +++++++++++++--

Thanks, I've queued the clock patches in clk-renesas-for-v4.17, as they are a hard dependency for:
(1) The new reset vector in arch/arm/mach-shmobile/
(2) The watchdog driver.

Quoted 6 lines
Quoted 2 lines

Note that the watchdog driver itself (2) has a hard dependency on the new reset vector (1).

Thanks, I have marked the watchdog driver patch as Deferred. Please resubmit or otherwise ping me once they dependencies are in an rc release.

I gave the dependencies a bit more thought.

I think we can gain once release cycle if you would postpone the DTS patches (both SMP routine size adjustments and RWDT device node additions) to v4.18. Then all other parts can be upstreamed in parallel in v4.17, as nothing will be activated before the DTS parts are in.

Does this look sane?
Thanks!

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#57 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

Quoted 110 lines · Geert Uytterhoeven

Hi Simon,

Quoted 100 lines · Simon Horman
Quoted 99 lines · Geert Uytterhoeven
Quoted 83 lines · Fabrizio Castro

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock
clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

 .../devicetree/bindings/watchdog/renesas-wdt.txt   | 19 ++++++--
 arch/arm/boot/dts/r8a7743-iwg20m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7743.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7745-iwg22m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7745.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7790.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7791.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7792.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7793.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7794.dtsi                     | 12 ++++-
 arch/arm/configs/shmobile_defconfig                |  1 +
 arch/arm/mach-shmobile/common.h                    |  6 +++
 arch/arm/mach-shmobile/headsmp.S                   | 55 ++++++++++++++++++++++
 arch/arm/mach-shmobile/platsmp-apmu.c              |  1 +
Quoted 8 lines
 arch/arm/mach-shmobile/pm-rcar-gen2.c              | 15 ++++--
 drivers/clk/renesas/r8a7743-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7745-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7790-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7791-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7794-cpg-mssr.c             |  2 +
 drivers/soc/renesas/rcar-rst.c                     | 35 +++++++++++---
 drivers/watchdog/renesas_wdt.c                     | 39 +++++++++++++--

Thanks, I've queued the clock patches in clk-renesas-for-v4.17, as they are a hard dependency for:
(1) The new reset vector in arch/arm/mach-shmobile/
(2) The watchdog driver.

Quoted 6 lines
Quoted 2 lines

Note that the watchdog driver itself (2) has a hard dependency on the new reset vector (1).

Thanks, I have marked the watchdog driver patch as Deferred. Please resubmit or otherwise ping me once they dependencies are in an rc release.

Quoted 8 lines

I gave the dependencies a bit more thought.

I think we can gain once release cycle if you would postpone the DTS patches (both SMP routine size adjustments and RWDT device node additions) to v4.18. Then all other parts can be upstreamed in parallel in v4.17, as nothing will be activated before the DTS parts are in.

Does this look sane?

What about new DTS on old kernels of a specific vintage?

#58 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

Hi Simon,

Quoted 111 lines · Simon Horman
Quoted 110 lines · Geert Uytterhoeven
Quoted 100 lines · Simon Horman
Quoted 99 lines · Geert Uytterhoeven
Quoted 83 lines · Fabrizio Castro

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock
clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

 .../devicetree/bindings/watchdog/renesas-wdt.txt   | 19 ++++++--
 arch/arm/boot/dts/r8a7743-iwg20m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7743.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7745-iwg22m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7745.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7790.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7791.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7792.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7793.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7794.dtsi                     | 12 ++++-
 arch/arm/configs/shmobile_defconfig                |  1 +
 arch/arm/mach-shmobile/common.h                    |  6 +++
 arch/arm/mach-shmobile/headsmp.S                   | 55 ++++++++++++++++++++++
 arch/arm/mach-shmobile/platsmp-apmu.c              |  1 +
Quoted 8 lines
 arch/arm/mach-shmobile/pm-rcar-gen2.c              | 15 ++++--
 drivers/clk/renesas/r8a7743-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7745-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7790-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7791-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7794-cpg-mssr.c             |  2 +
 drivers/soc/renesas/rcar-rst.c                     | 35 +++++++++++---
 drivers/watchdog/renesas_wdt.c                     | 39 +++++++++++++--

Thanks, I've queued the clock patches in clk-renesas-for-v4.17, as they are a hard dependency for:
(1) The new reset vector in arch/arm/mach-shmobile/
(2) The watchdog driver.

Quoted 6 lines
Quoted 2 lines

Note that the watchdog driver itself (2) has a hard dependency on the new reset vector (1).

Thanks, I have marked the watchdog driver patch as Deferred. Please resubmit or otherwise ping me once they dependencies are in an rc release.

I gave the dependencies a bit more thought.

Quoted 8 lines
Quoted 6 lines

I think we can gain once release cycle if you would postpone the DTS patches (both SMP routine size adjustments and RWDT device node additions) to v4.18. Then all other parts can be upstreamed in parallel in v4.17, as nothing will be activated before the DTS parts are in.

Does this look sane?

What about new DTS on old kernels of a specific vintage?

New DTS would work fine on v4.16 and older, which don't have any of the other patches from this series.
New DTS would work fine on v4.17, assumed all other patches from this series get into v4.17.

The only issue would be a new DTS on a kernel that has some but not all other patches from this series.

If we want to fast-track everything into v4.17, we need a stable branch, to be pulled by all 5 actors (clk-renesas, mach-shmobile, soc-renesas, watchdog, renesas-dts).

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#59 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

Quoted 123 lines · Geert Uytterhoeven

Hi Simon,

Quoted 111 lines · Simon Horman
Quoted 110 lines · Geert Uytterhoeven
Quoted 100 lines · Simon Horman
Quoted 99 lines · Geert Uytterhoeven
Quoted 83 lines · Fabrizio Castro

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock
clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

 .../devicetree/bindings/watchdog/renesas-wdt.txt   | 19 ++++++--
 arch/arm/boot/dts/r8a7743-iwg20m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7743.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7745-iwg22m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7745.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7790.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7791.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7792.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7793.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7794.dtsi                     | 12 ++++-
 arch/arm/configs/shmobile_defconfig                |  1 +
 arch/arm/mach-shmobile/common.h                    |  6 +++
 arch/arm/mach-shmobile/headsmp.S                   | 55 ++++++++++++++++++++++
 arch/arm/mach-shmobile/platsmp-apmu.c              |  1 +
Quoted 8 lines
 arch/arm/mach-shmobile/pm-rcar-gen2.c              | 15 ++++--
 drivers/clk/renesas/r8a7743-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7745-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7790-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7791-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7794-cpg-mssr.c             |  2 +
 drivers/soc/renesas/rcar-rst.c                     | 35 +++++++++++---
 drivers/watchdog/renesas_wdt.c                     | 39 +++++++++++++--

Thanks, I've queued the clock patches in clk-renesas-for-v4.17, as they are a hard dependency for:
(1) The new reset vector in arch/arm/mach-shmobile/
(2) The watchdog driver.

Quoted 6 lines
Quoted 2 lines

Note that the watchdog driver itself (2) has a hard dependency on the new reset vector (1).

Thanks, I have marked the watchdog driver patch as Deferred. Please resubmit or otherwise ping me once they dependencies are in an rc release.

I gave the dependencies a bit more thought.

Quoted 8 lines
Quoted 6 lines

I think we can gain once release cycle if you would postpone the DTS patches (both SMP routine size adjustments and RWDT device node additions) to v4.18. Then all other parts can be upstreamed in parallel in v4.17, as nothing will be activated before the DTS parts are in.

Does this look sane?

What about new DTS on old kernels of a specific vintage?

Quoted 7 lines

New DTS would work fine on v4.16 and older, which don't have any of the other patches from this series.
New DTS would work fine on v4.17, assumed all other patches from this series get into v4.17.

The only issue would be a new DTS on a kernel that has some but not all other patches from this series.

Thanks for the analysis. I think this is the key point. What is the risk of that happening?

Quoted 3 lines

If we want to fast-track everything into v4.17, we need a stable branch, to be pulled by all 5 actors (clk-renesas, mach-shmobile, soc-renesas, watchdog, renesas-dts).

It seems that watchdog is the only item on that list that neither you nor I sign off on. So I conclude that if the watchdog patch goes in early enough in the cycle then we ought to be in good shape.

None the less, is it worth the risk?

#60 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

Hi Simon,

Quoted 127 lines · Simon Horman
Quoted 126 lines · Geert Uytterhoeven
Quoted 111 lines · Simon Horman
Quoted 110 lines · Geert Uytterhoeven
Quoted 100 lines · Simon Horman
Quoted 99 lines · Geert Uytterhoeven
Quoted 83 lines · Fabrizio Castro

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock
clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

 .../devicetree/bindings/watchdog/renesas-wdt.txt   | 19 ++++++--
 arch/arm/boot/dts/r8a7743-iwg20m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7743.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7745-iwg22m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7745.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7790.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7791.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7792.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7793.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7794.dtsi                     | 12 ++++-
 arch/arm/configs/shmobile_defconfig                |  1 +
 arch/arm/mach-shmobile/common.h                    |  6 +++
 arch/arm/mach-shmobile/headsmp.S                   | 55 ++++++++++++++++++++++
 arch/arm/mach-shmobile/platsmp-apmu.c              |  1 +
Quoted 8 lines
 arch/arm/mach-shmobile/pm-rcar-gen2.c              | 15 ++++--
 drivers/clk/renesas/r8a7743-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7745-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7790-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7791-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7794-cpg-mssr.c             |  2 +
 drivers/soc/renesas/rcar-rst.c                     | 35 +++++++++++---
 drivers/watchdog/renesas_wdt.c                     | 39 +++++++++++++--

Thanks, I've queued the clock patches in clk-renesas-for-v4.17, as they are a hard dependency for:
(1) The new reset vector in arch/arm/mach-shmobile/
(2) The watchdog driver.

Quoted 6 lines
Quoted 2 lines

Note that the watchdog driver itself (2) has a hard dependency on the new reset vector (1).

Thanks, I have marked the watchdog driver patch as Deferred. Please resubmit or otherwise ping me once they dependencies are in an rc release.

I gave the dependencies a bit more thought.

Quoted 8 lines
Quoted 6 lines

I think we can gain once release cycle if you would postpone the DTS patches (both SMP routine size adjustments and RWDT device node additions) to v4.18. Then all other parts can be upstreamed in parallel in v4.17, as nothing will be activated before the DTS parts are in.

Does this look sane?

What about new DTS on old kernels of a specific vintage?

New DTS would work fine on v4.16 and older, which don't have any of the other patches from this series.
New DTS would work fine on v4.17, assumed all other patches from this series get into v4.17.

Quoted 5 lines
Quoted 2 lines

The only issue would be a new DTS on a kernel that has some but not all other patches from this series.

Thanks for the analysis. I think this is the key point. What is the risk of that happening?

I think the risk is fairly low.
You control when the DTS changes go in, and also when they are backported to LTSI.

Quoted 9 lines
Quoted 3 lines

If we want to fast-track everything into v4.17, we need a stable branch, to be pulled by all 5 actors (clk-renesas, mach-shmobile, soc-renesas, watchdog, renesas-dts).

It seems that watchdog is the only item on that list that neither you nor I sign off on. So I conclude that if the watchdog patch goes in early enough in the cycle then we ought to be in good shape.

None the less, is it worth the risk?

Do we want the functionality early, or can we wait until v4.19?

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#61 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

Quoted 144 lines · Geert Uytterhoeven

Hi Simon,

Quoted 127 lines · Simon Horman
Quoted 126 lines · Geert Uytterhoeven
Quoted 111 lines · Simon Horman
Quoted 110 lines · Geert Uytterhoeven
Quoted 100 lines · Simon Horman
Quoted 99 lines · Geert Uytterhoeven
Quoted 83 lines · Fabrizio Castro

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock
clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

 .../devicetree/bindings/watchdog/renesas-wdt.txt   | 19 ++++++--
 arch/arm/boot/dts/r8a7743-iwg20m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7743.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7745-iwg22m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7745.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7790.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7791.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7792.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7793.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7794.dtsi                     | 12 ++++-
 arch/arm/configs/shmobile_defconfig                |  1 +
 arch/arm/mach-shmobile/common.h                    |  6 +++
 arch/arm/mach-shmobile/headsmp.S                   | 55 ++++++++++++++++++++++
 arch/arm/mach-shmobile/platsmp-apmu.c              |  1 +
Quoted 8 lines
 arch/arm/mach-shmobile/pm-rcar-gen2.c              | 15 ++++--
 drivers/clk/renesas/r8a7743-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7745-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7790-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7791-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7794-cpg-mssr.c             |  2 +
 drivers/soc/renesas/rcar-rst.c                     | 35 +++++++++++---
 drivers/watchdog/renesas_wdt.c                     | 39 +++++++++++++--

Thanks, I've queued the clock patches in clk-renesas-for-v4.17, as they are a hard dependency for:
(1) The new reset vector in arch/arm/mach-shmobile/
(2) The watchdog driver.

Quoted 6 lines
Quoted 2 lines

Note that the watchdog driver itself (2) has a hard dependency on the new reset vector (1).

Thanks, I have marked the watchdog driver patch as Deferred. Please resubmit or otherwise ping me once they dependencies are in an rc release.

I gave the dependencies a bit more thought.

Quoted 8 lines
Quoted 6 lines

I think we can gain once release cycle if you would postpone the DTS patches (both SMP routine size adjustments and RWDT device node additions) to v4.18. Then all other parts can be upstreamed in parallel in v4.17, as nothing will be activated before the DTS parts are in.

Does this look sane?

What about new DTS on old kernels of a specific vintage?

New DTS would work fine on v4.16 and older, which don't have any of the other patches from this series.
New DTS would work fine on v4.17, assumed all other patches from this series get into v4.17.

Quoted 5 lines
Quoted 2 lines

The only issue would be a new DTS on a kernel that has some but not all other patches from this series.

Thanks for the analysis. I think this is the key point. What is the risk of that happening?

I think the risk is fairly low.
You control when the DTS changes go in, and also when they are backported to LTSI.

Quoted 4 lines
Quoted 3 lines

If we want to fast-track everything into v4.17, we need a stable branch, to be pulled by all 5 actors (clk-renesas, mach-shmobile, soc-renesas, watchdog, renesas-dts).

Quoted 7 lines
Quoted 5 lines

It seems that watchdog is the only item on that list that neither you nor I sign off on. So I conclude that if the watchdog patch goes in early enough in the cycle then we ought to be in good shape.

None the less, is it worth the risk?

Do we want the functionality early, or can we wait until v4.19?

I understand that the watchdog driver has been accepted and that as things currently stand if it is enabled in the build it will be enabled in DT and a crash will result. From my POV that is a regression so I have decided to adopt the alternate plan proposed by Geert above: in short drop DT patches (to avoid enabling broken driver) and enqueue mach-shmobile patches (which don't make things any worse.

Patch-by-patch things are as follows.

Dropped from "[PATCH v5 00/26] Fix watchdog on Renesas R-Car Gen2 and RZ/G1":

979d28f29742 ARM: dts: r8a7743: Adjust SMP routine size 0b205f679f5d ARM: dts: r8a7745: Adjust SMP routine size 2ad21a1d78b3 ARM: dts: r8a7790: Adjust SMP routine size 82c978459c77 ARM: dts: r8a7791: Adjust SMP routine size d303698e6ff3 ARM: dts: r8a7792: Adjust SMP routine size d8b8b9a1e295 ARM: dts: r8a7793: Adjust SMP routine size 70b3ab369773 ARM: dts: r8a7794: Adjust SMP routine size

ebf26cf1b1de ARM: dts: r8a7743: Add watchdog support to SoC dtsi 5a4566ab3777 ARM: dts: r8a7745: Add watchdog support to SoC dtsi acd58e2bb1e5 ARM: dts: r8a7790: Add watchdog support to SoC dtsi e3f89168e72f ARM: dts: r8a7791: Add watchdog support to SoC dtsi e76a6a69a1fc ARM: dts: r8a7794: Add watchdog support to SoC dtsi 242cc1ab7f7c ARM: dts: iwg20m: Add watchdog support to SoM dtsi 748ed07f6c7c ARM: dts: iwg22m: Add watchdog support to SoM dtsi

Dropped from "[PATCH/RFC 00/11] ARM: dts: rcar-gen2: Enable watchdog support":

c257202482e2 ARM: dts: r8a7792: Add RWDT node 01ec04dec8eb ARM: dts: r8a7793: Add RWDT node 8e6ebe4f2f94 ARM: dts: lager: Enable watchdog support f223ad0198f6 ARM: dts: koelsch: Enable watchdog support 1d14d2bb700d ARM: dts: porter: Enable watchdog support bf71652fc955 ARM: dts: blanche: Enable watchdog support 13a0b10aeea4 ARM: dts: wheat: Enable watchdog support 4e247bf110b8 ARM: dts: gose: Enable watchdog support 2dcbd24bbf5e ARM: dts: alt: Enable watchdog support 7f6844c900b4 ARM: dts: silk: Enable watchdog support

And applied:

[v5] ARM: shmobile: rcar-gen2: Add watchdog support [v5] ARM: shmobile: Add watchdog support

In the case of the last patch above I removed the #ifdef from the header file.

#62 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

Hi Simon,

Quoted 192 lines · Simon Horman
Quoted 142 lines · Geert Uytterhoeven
Quoted 127 lines · Simon Horman
Quoted 126 lines · Geert Uytterhoeven
Quoted 111 lines · Simon Horman
Quoted 110 lines · Geert Uytterhoeven
Quoted 100 lines · Simon Horman
Quoted 99 lines · Geert Uytterhoeven
Quoted 83 lines · Fabrizio Castro

this series has been around for some time as RFC, and it has collected useful comments from the community along the way.
The solution proposed by this patch set works for most R-Car Gen2 and RZ/G1 devices, but not all of them. We now know that for some R-Car Gen2 early revisions there is no proper software fix. Anyway, no product has been built around early revisions, but development boards mounting early revisions (basically prototypes) are still out there. As a result, this series isn't enabling the internal watchdog on R-Car Gen2 boards, developers may enable it in board specific device trees if needed.
This series has been tested by me on the iwg20d, iwg22d, Lager, Alt, and Koelsch boards.

The problem
===========
To deal with SMP on R-Car Gen2 and RZ/G1, we install a reset vector to ICRAM1 and we program the [S]BAR registers so that when we turn ON the non-boot CPUs they are redirected to the reset vector installed by Linux in ICRAM1, and eventually they continue the execution to RAM, where the SMP bring-up code will take care of the rest.
The content of the [S]BAR registers survives a watchdog triggered reset, and as such after the watchdog fires the boot core will try and execute the SMP bring-up code instead of jumping to the bootrom code.

The fix
=======
The main strategy for the solution is to let the reset vector decide if it needs to jump to shmobile_boot_fn or to the bootrom code. In a watchdog triggered reset scenario, since the [S]BAR registers keep their values, the boot CPU will jump into the newly designed reset vector, the assembly routine will eventually test WOVF (a bit in register RWTCSRA that indicates if the watchdog counter has overflown, the value of this bit gets retained in this scenario), and jump to the bootrom code which will in turn load up the bootloader, etc.
When bringing up SMP or using CPU hotplug, the reset vector will jump to shmobile_boot_fn instead.

Thank you All for your help.

Best regards,

Fabrizio Castro (26):
ARM: shmobile: Add watchdog support
ARM: dts: r8a7743: Adjust SMP routine size
ARM: dts: r8a7745: Adjust SMP routine size
ARM: dts: r8a7790: Adjust SMP routine size
ARM: dts: r8a7791: Adjust SMP routine size
ARM: dts: r8a7792: Adjust SMP routine size
ARM: dts: r8a7793: Adjust SMP routine size
ARM: dts: r8a7794: Adjust SMP routine size
soc: renesas: rcar-rst: Enable watchdog as reset trigger for Gen2
ARM: shmobile: rcar-gen2: Add watchdog support
dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add R-Car Gen2 support
watchdog: renesas_wdt: Add restart handler
ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
clk: renesas: r8a7743: Add rwdt clock
clk: renesas: r8a7745: Add rwdt clock
clk: renesas: r8a7790: Add rwdt clock
clk: renesas: r8a7791/r8a7793: Add rwdt clock
clk: renesas: r8a7794: Add rwdt clock
ARM: dts: r8a7743: Add watchdog support to SoC dtsi
ARM: dts: r8a7745: Add watchdog support to SoC dtsi
ARM: dts: r8a7790: Add watchdog support to SoC dtsi
ARM: dts: r8a7791: Add watchdog support to SoC dtsi
ARM: dts: r8a7794: Add watchdog support to SoC dtsi
ARM: dts: iwg20m: Add watchdog support to SoM dtsi
ARM: dts: iwg22m: Add watchdog support to SoM dtsi

 .../devicetree/bindings/watchdog/renesas-wdt.txt   | 19 ++++++--
 arch/arm/boot/dts/r8a7743-iwg20m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7743.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7745-iwg22m.dtsi              |  5 ++
 arch/arm/boot/dts/r8a7745.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7790.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7791.dtsi                     | 12 ++++-
 arch/arm/boot/dts/r8a7792.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7793.dtsi                     |  2 +-
 arch/arm/boot/dts/r8a7794.dtsi                     | 12 ++++-
 arch/arm/configs/shmobile_defconfig                |  1 +
 arch/arm/mach-shmobile/common.h                    |  6 +++
 arch/arm/mach-shmobile/headsmp.S                   | 55 ++++++++++++++++++++++
 arch/arm/mach-shmobile/platsmp-apmu.c              |  1 +
Quoted 8 lines
 arch/arm/mach-shmobile/pm-rcar-gen2.c              | 15 ++++--
 drivers/clk/renesas/r8a7743-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7745-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7790-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7791-cpg-mssr.c             |  2 +
 drivers/clk/renesas/r8a7794-cpg-mssr.c             |  2 +
 drivers/soc/renesas/rcar-rst.c                     | 35 +++++++++++---
 drivers/watchdog/renesas_wdt.c                     | 39 +++++++++++++--

Thanks, I've queued the clock patches in clk-renesas-for-v4.17, as they are a hard dependency for:
(1) The new reset vector in arch/arm/mach-shmobile/
(2) The watchdog driver.

Quoted 6 lines
Quoted 2 lines

Note that the watchdog driver itself (2) has a hard dependency on the new reset vector (1).

Thanks, I have marked the watchdog driver patch as Deferred. Please resubmit or otherwise ping me once they dependencies are in an rc release.

I gave the dependencies a bit more thought.

Quoted 8 lines
Quoted 6 lines

I think we can gain once release cycle if you would postpone the DTS patches (both SMP routine size adjustments and RWDT device node additions) to v4.18. Then all other parts can be upstreamed in parallel in v4.17, as nothing will be activated before the DTS parts are in.

Does this look sane?

What about new DTS on old kernels of a specific vintage?

New DTS would work fine on v4.16 and older, which don't have any of the other patches from this series.
New DTS would work fine on v4.17, assumed all other patches from this series get into v4.17.

Quoted 5 lines
Quoted 2 lines

The only issue would be a new DTS on a kernel that has some but not all other patches from this series.

Thanks for the analysis. I think this is the key point. What is the risk of that happening?

I think the risk is fairly low.
You control when the DTS changes go in, and also when they are backported to LTSI.

Quoted 4 lines
Quoted 3 lines

If we want to fast-track everything into v4.17, we need a stable branch, to be pulled by all 5 actors (clk-renesas, mach-shmobile, soc-renesas, watchdog, renesas-dts).

Quoted 7 lines
Quoted 5 lines

It seems that watchdog is the only item on that list that neither you nor I sign off on. So I conclude that if the watchdog patch goes in early enough in the cycle then we ought to be in good shape.

None the less, is it worth the risk?

Do we want the functionality early, or can we wait until v4.19?

I understand that the watchdog driver has been accepted and that as things currently stand if it is enabled in the build it will be enabled in DT and a crash will result. From my POV that is a regression so I have decided to adopt the alternate plan proposed by Geert above: in short drop DT patches (to avoid enabling broken driver) and enqueue mach-shmobile patches (which don't make things any worse.

Patch-by-patch things are as follows.

Dropped from "[PATCH v5 00/26] Fix watchdog on Renesas R-Car Gen2 and RZ/G1":

979d28f29742 ARM: dts: r8a7743: Adjust SMP routine size 0b205f679f5d ARM: dts: r8a7745: Adjust SMP routine size 2ad21a1d78b3 ARM: dts: r8a7790: Adjust SMP routine size 82c978459c77 ARM: dts: r8a7791: Adjust SMP routine size d303698e6ff3 ARM: dts: r8a7792: Adjust SMP routine size d8b8b9a1e295 ARM: dts: r8a7793: Adjust SMP routine size 70b3ab369773 ARM: dts: r8a7794: Adjust SMP routine size

ebf26cf1b1de ARM: dts: r8a7743: Add watchdog support to SoC dtsi 5a4566ab3777 ARM: dts: r8a7745: Add watchdog support to SoC dtsi acd58e2bb1e5 ARM: dts: r8a7790: Add watchdog support to SoC dtsi e3f89168e72f ARM: dts: r8a7791: Add watchdog support to SoC dtsi e76a6a69a1fc ARM: dts: r8a7794: Add watchdog support to SoC dtsi 242cc1ab7f7c ARM: dts: iwg20m: Add watchdog support to SoM dtsi 748ed07f6c7c ARM: dts: iwg22m: Add watchdog support to SoM dtsi

Dropped from "[PATCH/RFC 00/11] ARM: dts: rcar-gen2: Enable watchdog support":

c257202482e2 ARM: dts: r8a7792: Add RWDT node 01ec04dec8eb ARM: dts: r8a7793: Add RWDT node 8e6ebe4f2f94 ARM: dts: lager: Enable watchdog support f223ad0198f6 ARM: dts: koelsch: Enable watchdog support 1d14d2bb700d ARM: dts: porter: Enable watchdog support bf71652fc955 ARM: dts: blanche: Enable watchdog support 13a0b10aeea4 ARM: dts: wheat: Enable watchdog support 4e247bf110b8 ARM: dts: gose: Enable watchdog support 2dcbd24bbf5e ARM: dts: alt: Enable watchdog support 7f6844c900b4 ARM: dts: silk: Enable watchdog support

Quoted 6 lines

And applied:

[v5] ARM: shmobile: rcar-gen2: Add watchdog support [v5] ARM: shmobile: Add watchdog support

In the case of the last patch above I removed the #ifdef from the header file.

Thank you, sounds good to me.

BTW, who will take "dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support"? G?nter, Wim: will you take this through the watchdog tree, too?

Thanks again!

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#63 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

In reply to Simon Horman

Hi Simon,

Quoted 41 lines · Simon Horman

I understand that the watchdog driver has been accepted and that as things currently stand if it is enabled in the build it will be enabled in DT and a crash will result. From my POV that is a regression so I have decided to adopt the alternate plan proposed by Geert above: in short drop DT patches (to avoid enabling broken driver) and enqueue mach-shmobile patches (which don't make things any worse.

Patch-by-patch things are as follows.

Dropped from "[PATCH v5 00/26] Fix watchdog on Renesas R-Car Gen2 and RZ/G1":

979d28f29742 ARM: dts: r8a7743: Adjust SMP routine size 0b205f679f5d ARM: dts: r8a7745: Adjust SMP routine size 2ad21a1d78b3 ARM: dts: r8a7790: Adjust SMP routine size 82c978459c77 ARM: dts: r8a7791: Adjust SMP routine size d303698e6ff3 ARM: dts: r8a7792: Adjust SMP routine size d8b8b9a1e295 ARM: dts: r8a7793: Adjust SMP routine size 70b3ab369773 ARM: dts: r8a7794: Adjust SMP routine size

ebf26cf1b1de ARM: dts: r8a7743: Add watchdog support to SoC dtsi 5a4566ab3777 ARM: dts: r8a7745: Add watchdog support to SoC dtsi acd58e2bb1e5 ARM: dts: r8a7790: Add watchdog support to SoC dtsi e3f89168e72f ARM: dts: r8a7791: Add watchdog support to SoC dtsi e76a6a69a1fc ARM: dts: r8a7794: Add watchdog support to SoC dtsi 242cc1ab7f7c ARM: dts: iwg20m: Add watchdog support to SoM dtsi 748ed07f6c7c ARM: dts: iwg22m: Add watchdog support to SoM dtsi

Dropped from "[PATCH/RFC 00/11] ARM: dts: rcar-gen2: Enable watchdog support":

c257202482e2 ARM: dts: r8a7792: Add RWDT node 01ec04dec8eb ARM: dts: r8a7793: Add RWDT node 8e6ebe4f2f94 ARM: dts: lager: Enable watchdog support f223ad0198f6 ARM: dts: koelsch: Enable watchdog support 1d14d2bb700d ARM: dts: porter: Enable watchdog support bf71652fc955 ARM: dts: blanche: Enable watchdog support 13a0b10aeea4 ARM: dts: wheat: Enable watchdog support 4e247bf110b8 ARM: dts: gose: Enable watchdog support 2dcbd24bbf5e ARM: dts: alt: Enable watchdog support 7f6844c900b4 ARM: dts: silk: Enable watchdog support

Quoted 6 lines

And applied:

[v5] ARM: shmobile: rcar-gen2: Add watchdog support [v5] ARM: shmobile: Add watchdog support

In the case of the last patch above I removed the #ifdef from the header file.

I guess the time is ripe to apply the DTS patches again, in the following order:
1. SMP routine size adjustments,
2. Watchdog support additions to dtsi,
3. Watchdog enablement for dts?

Thanks!

Gr{oetje,eeting}s,

Geert

Signature · 5 lines
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
#64 lore raw

v5 cover Fix watchdog on Renesas R-Car Gen2 and RZ/G1

Quoted 49 lines · Geert Uytterhoeven

Hi Simon,

Quoted 46 lines · Simon Horman

I understand that the watchdog driver has been accepted and that as things currently stand if it is enabled in the build it will be enabled in DT and a crash will result. From my POV that is a regression so I have decided to adopt the alternate plan proposed by Geert above: in short drop DT patches (to avoid enabling broken driver) and enqueue mach-shmobile patches (which don't make things any worse.

Patch-by-patch things are as follows.

Dropped from "[PATCH v5 00/26] Fix watchdog on Renesas R-Car Gen2 and RZ/G1":

979d28f29742 ARM: dts: r8a7743: Adjust SMP routine size 0b205f679f5d ARM: dts: r8a7745: Adjust SMP routine size 2ad21a1d78b3 ARM: dts: r8a7790: Adjust SMP routine size 82c978459c77 ARM: dts: r8a7791: Adjust SMP routine size d303698e6ff3 ARM: dts: r8a7792: Adjust SMP routine size d8b8b9a1e295 ARM: dts: r8a7793: Adjust SMP routine size 70b3ab369773 ARM: dts: r8a7794: Adjust SMP routine size

ebf26cf1b1de ARM: dts: r8a7743: Add watchdog support to SoC dtsi 5a4566ab3777 ARM: dts: r8a7745: Add watchdog support to SoC dtsi acd58e2bb1e5 ARM: dts: r8a7790: Add watchdog support to SoC dtsi e3f89168e72f ARM: dts: r8a7791: Add watchdog support to SoC dtsi e76a6a69a1fc ARM: dts: r8a7794: Add watchdog support to SoC dtsi 242cc1ab7f7c ARM: dts: iwg20m: Add watchdog support to SoM dtsi 748ed07f6c7c ARM: dts: iwg22m: Add watchdog support to SoM dtsi

Dropped from "[PATCH/RFC 00/11] ARM: dts: rcar-gen2: Enable watchdog support":

c257202482e2 ARM: dts: r8a7792: Add RWDT node 01ec04dec8eb ARM: dts: r8a7793: Add RWDT node 8e6ebe4f2f94 ARM: dts: lager: Enable watchdog support f223ad0198f6 ARM: dts: koelsch: Enable watchdog support 1d14d2bb700d ARM: dts: porter: Enable watchdog support bf71652fc955 ARM: dts: blanche: Enable watchdog support 13a0b10aeea4 ARM: dts: wheat: Enable watchdog support 4e247bf110b8 ARM: dts: gose: Enable watchdog support 2dcbd24bbf5e ARM: dts: alt: Enable watchdog support 7f6844c900b4 ARM: dts: silk: Enable watchdog support

And applied:

[v5] ARM: shmobile: rcar-gen2: Add watchdog support [v5] ARM: shmobile: Add watchdog support

Quoted 8 lines
Quoted 1 line

In the case of the last patch above I removed the #ifdef from the header file.

I guess the time is ripe to apply the DTS patches again, in the following order:
1. SMP routine size adjustments,
2. Watchdog support additions to dtsi,
3. Watchdog enablement for dts?

Thanks!

Thanks, I have tentatively reapplied all of the above.