Skip to contentSkip to conversation

The cache flush using coprocessor must be in lib_arm/cache-cp15.c

PATCH New Kim 3 participants 10 messages May 28, 2009 – May 29, 2009

lore
Share & tools
lore

Kim

#0 lore raw

PATCH 1/1 The cache flush using coprocessor must be in lib_arm/cache-cp15.c

Hi? I'm about to suggest one more thing related to "change cpu.c under cpu/arm_cortexa8 dir to common code.".

asm ("mcr p15, 0, %0, c7, c5, 0": :"r" (0));

It's the common feature all over the arm core, not only arm cortex A8. The common cache function is defined in the lib_arm/cache-cp15.c as you know. So, It's seems the better method that cache_flush() is moved to lib_arm/cache-cp15.c

---
 cpu/arm_cortexa8/cpu.c |    4 ----
 lib_arm/cache-cp15.c   |    5 +++++
 2 files changed, 5 insertions(+), 4 deletions(-)
cpu/arm_cortexa8/cpu.c +0 -4
index 3e1780b..329febb 100644
@@ -154,7 +154,3 @@ void l2cache_disable()
 	}
 }
 
-static void cache_flush(void)
-{
-	asm ("mcr p15, 0, %0, c7, c5, 0": :"r" (0));
-}
lib_arm/cache-cp15.c +5 -0
index 62ed54f..face9b2 100644
@@ -53,6 +53,11 @@ static void cache_disable(uint32_t cache_bit)
 	cp_delay();
 	set_cr(reg & ~cache_bit);
 }
+
+static void cache_flush(void)
+{
+	asm ("mcr p15, 0, %0, c7, c5, 0": :"r" (0));
+}
 #endif
 
 #ifdef CONFIG_SYS_NO_ICACHE
Signature · 1 line
1.5.6.3
#1 lore raw

PATCH 1/1 The cache flush using coprocessor must be in lib_arm/cache-cp15.c

Quoted 11 lines · Heung Jun

Hi? I'm about to suggest one more thing related to "change cpu.c under cpu/arm_cortexa8 dir to common code.".

asm ("mcr p15, 0, %0, c7, c5, 0": :"r" (0));

It's the common feature all over the arm core, not only arm cortex A8. The common cache function is defined in the lib_arm/cache-cp15.c as you know. So, It's seems the better method that cache_flush() is moved to lib_arm/cache-cp15.c

Yes, this is fine. Thanks.

Do you like to update

http://lists.denx.de/pipermail/u-boot/2009-May/053386.html

regarding this and rename the functions there to e.g. cortexa8_l2cache_enable()/disable() (or soc_* or cpu_*)?

Best regards

Dirk

Quoted 33 lines
---
 cpu/arm_cortexa8/cpu.c |    4 ----
 lib_arm/cache-cp15.c   |    5 +++++
 2 files changed, 5 insertions(+), 4 deletions(-)
cpu/arm_cortexa8/cpu.c +0 -4
index 3e1780b..329febb 100644
@@ -154,7 +154,3 @@ void l2cache_disable()
 	}
 }
 
-static void cache_flush(void)
-{
-	asm ("mcr p15, 0, %0, c7, c5, 0": :"r" (0));
-}
lib_arm/cache-cp15.c +5 -0
index 62ed54f..face9b2 100644
@@ -53,6 +53,11 @@ static void cache_disable(uint32_t cache_bit)
 	cp_delay();
 	set_cr(reg & ~cache_bit);
 }
+
+static void cache_flush(void)
+{
+	asm ("mcr p15, 0, %0, c7, c5, 0": :"r" (0));
+}
 #endif
 
 #ifdef CONFIG_SYS_NO_ICACHE
#2 lore raw

PATCH 1/1 The cache flush using coprocessor must be in lib_arm/cache-cp15.c

In reply to Kim
Quoted 9 lines · Heung Jun

Hi? I'm about to suggest one more thing related to "change cpu.c under cpu/arm_cortexa8 dir to common code.".

asm ("mcr p15, 0, %0, c7, c5, 0": :"r" (0));

It's the common feature all over the arm core, not only arm cortex A8. The common cache function is defined in the lib_arm/cache-cp15.c as you know. So, It's seems the better method that cache_flush() is moved to lib_arm/cache-cp15.c

NACK

to boot linux we are suppose to clean both cache I and DCache

the I/DCache flush is armv and arch dependant (example ARM7TDMI, armv4wt, armv3, armv7, etc...)

so for now please let it in cpu/$(ARCH)/cpu.c

Best Regards,
J.

#3 lore raw

PATCH 1/1 The cache flush using coprocessor must be in lib_arm/cache-cp15.c

Quoted 10 lines · Jean-Christophe PLAGNIOL-VILLARD
Quoted 9 lines · Heung Jun

Hi? I'm about to suggest one more thing related to "change cpu.c under cpu/arm_cortexa8 dir to common code.".

asm ("mcr p15, 0, %0, c7, c5, 0": :"r" (0));

It's the common feature all over the arm core, not only arm cortex A8. The common cache function is defined in the lib_arm/cache-cp15.c as you know. So, It's seems the better method that cache_flush() is moved to lib_arm/cache-cp15.c

Quoted 8 lines

NACK

to boot linux we are suppose to clean both cache I and DCache

the I/DCache flush is armv and arch dependant (example ARM7TDMI, armv4wt, armv3, armv7, etc...)

so for now please let it in cpu/$(ARCH)/cpu.c

Well, in

http://lists.denx.de/pipermail/u-boot/2009-May/053396.html

Kim mentioned that he shuffles around the code due to compile errors while adding his new code. So we have to see if we can stay with the current implementation or have to change something.

Kim: What compile errors do you get for cache_flush()? Could you check if you really don't need it? Why don't you get compile errors if you move it from cpu/arm_cortexa8/cpu.c to lib_arm/cache-cp15.c?

Best regards

Dirk

#4 lore raw

PATCH 1/1 The cache flush using coprocessor must be in lib_arm/cache-cp15.c

Thanks to point. Dirk.

And call me "riverful" next time, plz :)

Quoted 3 lines

Kim mentioned that he shuffles around the code due to compile errors while adding his new code. So we have to see if we can stay with the current implementation or have to change something.

Yes. I shuffles around a little now.

Quoted 3 lines

Kim: What compile errors do you get for cache_flush()? Could you check if you really don't need it? Why don't you get compile errors if you move it from cpu/arm_cortexa8/cpu.c to lib_arm/cache-cp15.c?

The compile errors are two.

1. First error occured cause of omap3 dependent definition CPU_3XX_ES20.

cpu.c:99: warning: function declaration isn't a prototype cpu.c: In function 'l2cache_enable':
cpu.c:104: warning: implicit declaration of function 'get_cpu_rev' cpu.c:104: error: 'CPU_3XX_ES20' undeclared (first use in this function) cpu.c:104: error: (Each undeclared identifier is reported only once cpu.c:104: error: for each function it appears in.)
cpu.c: At top level:
cpu.c:129: warning: function declaration isn't a prototype cpu.c: In function 'l2cache_disable':
cpu.c:134: error: 'CPU_3XX_ES20' undeclared (first use in this function) make[1]: *** [cpu.o] Error 1

2. After upper error was resolved, linking error occured. The definition PRM_RSTCTRL is also omap3 dependent. It is used in the cpu/arm_cortexa8/start.S, and defined in the include/asm-arm/arch-omap3/cpu.h

/arm_cortexa8/s5pc100/libs5pc100.a lib_arm/libarm.a fs/cramfs/libcramfs.a fs/fat/libfat.a fs/fdos/libfdos.a fs/jffs2/libjffs2.a fs/reiserfs/libreiserfs.a fs/ext2/libext2fs.a fs/yaffs2/libyaffs2.a fs/ubifs/libubifs.a net/libnet.a disk/libdisk.a drivers/bios_emulator/libatibiosemu.a drivers/block/libblock.a drivers/dma/libdma.a drivers/fpga/libfpga.a drivers/gpio/libgpio.a drivers/hwmon/libhwmon.a drivers/i2c/libi2c.a drivers/input/libinput.a drivers/misc/libmisc.a drivers/mmc/libmmc.a drivers/mtd/libmtd.a drivers/mtd/nand/libnand.a drivers/mtd/nand_legacy/libnand_legacy.a drivers/mtd/onenand/libonenand.a drivers/mtd/ubi/libubi.a drivers/mtd/spi/libspi_flash.a drivers/net/libnet.a drivers/net/phy/libphy.a drivers/net/sk98lin/libsk98lin.a drivers/pci/libpci.a drivers/pcmcia/libpcmcia.a drivers/spi/libspi.a drivers/rtc/librtc.a drivers/serial/libserial.a drivers/twserial/libtws.a drivers/usb/gadget/libusb_gadget.a drivers/usb/host/libusb_host.a drivers/usb/musb/libusb_musb.a drivers/video/libvideo.a drivers/watchdog/libwatchdog.a common/libcommon.a libfdt/libfdt.a api/libapi.a post/libpost.a board/samsung/tickertape/libtickertape.a --end-group -L /opt/toolchains/arm-2007q3/bin/../lib/gcc/arm-none-linux-gnueabi/4.2.1 -lgcc \
-Map u-boot.map -o u-boot
cpu/arm_cortexa8/start.o: In function `rstctl':
/ubuntu/archive/s5pc100/u-boot-arm/cpu/arm_cortexa8/start.S:514:
undefined reference to `PRM_RSTCTRL'

Thanks one more.

Best regards,

HeungJun, Kim(riverful)

#5 lore raw

PATCH 1/1 The cache flush using coprocessor must be in lib_arm/cache-cp15.c

Hi riverful,

Quoted 16 lines · Heung Jun

Thanks to point. Dirk.

And call me "riverful" next time, plz :)

Quoted 3 lines

Kim mentioned that he shuffles around the code due to compile errors while adding his new code. So we have to see if we can stay with the current implementation or have to change something.

Yes. I shuffles around a little now.

Quoted 3 lines

Kim: What compile errors do you get for cache_flush()? Could you check if you really don't need it? Why don't you get compile errors if you move it from cpu/arm_cortexa8/cpu.c to lib_arm/cache-cp15.c?

The compile errors are two.

1. First error occured cause of omap3 dependent definition CPU_3XX_ES20.

Quoted 11 lines

cpu.c:99: warning: function declaration isn't a prototype cpu.c: In function 'l2cache_enable':
cpu.c:104: warning: implicit declaration of function 'get_cpu_rev' cpu.c:104: error: 'CPU_3XX_ES20' undeclared (first use in this function) cpu.c:104: error: (Each undeclared identifier is reported only once cpu.c:104: error: for each function it appears in.)
cpu.c: At top level:
cpu.c:129: warning: function declaration isn't a prototype cpu.c: In function 'l2cache_disable':
cpu.c:134: error: 'CPU_3XX_ES20' undeclared (first use in this function) make[1]: *** [cpu.o] Error 1

Ok, this looks to me that we have to move l2cache_enable() and l2cache_disable() to cpu/arm_cortexa8/omap3/board.c and rename them to cortexa8_l2cache_enable()/disable() as asked in

http://lists.denx.de/pipermail/u-boot/2009-May/053405.html

But it doesn't seem that we have to touch cache_flush()? Then it can stay in cpu/arm_cortexa8/cpu.c as requested by Jean-Christophe?

Quoted 22 lines

2. After upper error was resolved, linking error occured. The definition PRM_RSTCTRL is also omap3 dependent. It is used in the cpu/arm_cortexa8/start.S, and defined in the include/asm-arm/arch-omap3/cpu.h

/arm_cortexa8/s5pc100/libs5pc100.a lib_arm/libarm.a fs/cramfs/libcramfs.a fs/fat/libfat.a fs/fdos/libfdos.a fs/jffs2/libjffs2.a fs/reiserfs/libreiserfs.a fs/ext2/libext2fs.a fs/yaffs2/libyaffs2.a fs/ubifs/libubifs.a net/libnet.a disk/libdisk.a drivers/bios_emulator/libatibiosemu.a drivers/block/libblock.a drivers/dma/libdma.a drivers/fpga/libfpga.a drivers/gpio/libgpio.a drivers/hwmon/libhwmon.a drivers/i2c/libi2c.a drivers/input/libinput.a drivers/misc/libmisc.a drivers/mmc/libmmc.a drivers/mtd/libmtd.a drivers/mtd/nand/libnand.a drivers/mtd/nand_legacy/libnand_legacy.a drivers/mtd/onenand/libonenand.a drivers/mtd/ubi/libubi.a drivers/mtd/spi/libspi_flash.a drivers/net/libnet.a drivers/net/phy/libphy.a drivers/net/sk98lin/libsk98lin.a drivers/pci/libpci.a drivers/pcmcia/libpcmcia.a drivers/spi/libspi.a drivers/rtc/librtc.a drivers/serial/libserial.a drivers/twserial/libtws.a drivers/usb/gadget/libusb_gadget.a drivers/usb/host/libusb_host.a drivers/usb/musb/libusb_musb.a drivers/video/libvideo.a drivers/watchdog/libwatchdog.a

Quoted 8 lines

common/libcommon.a libfdt/libfdt.a api/libapi.a post/libpost.a board/samsung/tickertape/libtickertape.a --end-group -L /opt/toolchains/arm-2007q3/bin/../lib/gcc/arm-none-linux-gnueabi/4.2.1 -lgcc \
-Map u-boot.map -o u-boot
cpu/arm_cortexa8/start.o: In function `rstctl':
/ubuntu/archive/s5pc100/u-boot-arm/cpu/arm_cortexa8/start.S:514:
undefined reference to `PRM_RSTCTRL'

This is new and not discussed anywhere on this list already?

So to summarize my understanding:

We have to:

a) move l2cache_enable() and l2cache_disable()
b) keep cache_flush() as it is
c) talk about start.S

Do you agree?

Best regards

Dirk

#6 lore raw

PATCH 1/1 The cache flush using coprocessor must be in lib_arm/cache-cp15.c

Quoted 34 lines · Dirk Behme

Hi riverful,

Quoted 16 lines · Heung Jun

Thanks to point. Dirk.

And call me "riverful" next time, plz :)

Quoted 3 lines

Kim mentioned that he shuffles around the code due to compile errors while adding his new code. So we have to see if we can stay with the current implementation or have to change something.

Yes. I shuffles around a little now.

Quoted 3 lines

Kim: What compile errors do you get for cache_flush()? Could you check if you really don't need it? Why don't you get compile errors if you move it from cpu/arm_cortexa8/cpu.c to lib_arm/cache-cp15.c?

The compile errors are two.

1. First error occured cause of omap3 dependent definition CPU_3XX_ES20.

Quoted 11 lines

cpu.c:99: warning: function declaration isn't a prototype cpu.c: In function 'l2cache_enable':
cpu.c:104: warning: implicit declaration of function 'get_cpu_rev' cpu.c:104: error: 'CPU_3XX_ES20' undeclared (first use in this function) cpu.c:104: error: (Each undeclared identifier is reported only once cpu.c:104: error: for each function it appears in.)
cpu.c: At top level:
cpu.c:129: warning: function declaration isn't a prototype cpu.c: In function 'l2cache_disable':
cpu.c:134: error: 'CPU_3XX_ES20' undeclared (first use in this function) make[1]: *** [cpu.o] Error 1

Ok, this looks to me that we have to move l2cache_enable() and l2cache_disable() to cpu/arm_cortexa8/omap3/board.c and rename them to cortexa8_l2cache_enable()/disable() as asked in

<snip>

Quoted 7 lines

This is new and not discussed anywhere on this list already?

So to summarize my understanding:

We have to:

a) move l2cache_enable() and l2cache_disable()

move I agree rename no

we need to move the l2cache_XXXXX to cpu/$(ARCH)/$(SOC)/cache.c

Quoted 3 lines

b) keep cache_flush() as it is

c) talk about start.S

I'll propose a patch during the WE

Best Regards,
J.

#7 lore raw

PATCH 1/1 The cache flush using coprocessor must be in lib_arm/cache-cp15.c

Quoted 1 line · Jean-Christophe PLAGNIOL-VILLARD

On 09:07 Fri 29 May , Dirk Behme wrote:

...

Quoted 2 lines
Quoted 1 line

c) talk about start.S

I'll propose a patch during the WE

Do you have already any idea about how such a patch could look like? I haven't looked into it, yet.

Best regards

Dirk

#8 lore raw

PATCH 1/1 The cache flush using coprocessor must be in lib_arm/cache-cp15.c

Quoted 8 lines · Dirk Behme
Quoted 1 line · Jean-Christophe PLAGNIOL-VILLARD

On 09:07 Fri 29 May , Dirk Behme wrote:

...

Quoted 2 lines
Quoted 1 line

c) talk about start.S

I'll propose a patch during the WE

Do you have already any idea about how such a patch could look like? I haven't looked into it, yet.

yes I've one, IIRC I've a patch that need to be rebase but it will clean this today I've not the time sorry

Best Regards,
J.

#9 lore raw

PATCH 1/1 The cache flush using coprocessor must be in lib_arm/cache-cp15.c

Quoted 10 lines · Jean-Christophe PLAGNIOL-VILLARD
Quoted 7 lines · Dirk Behme
Quoted 1 line · Jean-Christophe PLAGNIOL-VILLARD

On 09:07 Fri 29 May , Dirk Behme wrote:

...

Quoted 2 lines
Quoted 1 line

c) talk about start.S

I'll propose a patch during the WE

Do you have already any idea about how such a patch could look like? I haven't looked into it, yet.

yes I've one, IIRC I've a patch that need to be rebase but it will clean this today I've not the time sorry

Do you like to send the patch as it is (should just take 1min?)?

Then, riverful or someone else maybe can help to rebase it, test it and we can already review/discuss about it.

Many thanks

Dirk