Skip to contentSkip to conversation

Exotic architecture fixes

PATCH New Geert Uytterhoeven 4 participants 9 messages May 23, 2013 – Jun 4, 2013

lore
Share & tools
lore
#0 lore raw

I created a branch where I collected several exotic architecture fixes (and a few
other trivial fixes):

git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k.git
exotic-arch-fixes

Several of these have been ack'ed (cfr. *-indicators below). If you reject any of the commits, please scream. If you want to take any of them yourself, please do so.

I intend to merge this into my for-next branch after the release of v3.10-rc4, and ask Linus to pull after the release of v3.10-rc5.
This branch may be rebased.

Thanks!

Geert Uytterhoeven (32):
* avr32: atmel_default_console_device depends on CONFIG_SERIAL_ATMEL
* c6x: Wire up asm-generic/xor.h
cris/kgdb: Properly split long lines in asm
cris/kgdb: Remove unused static int do_printk cris/kgdb: Kill forward declarations for static functions cris/kgdb: Use #ifdef PROCESS_SUPPORT where needed cris/kgb: Make symbols used from asm global
cris/kgdb: Remove obsolete USED*() macros
cris/kgdb: Fix buffer overflow in getpacket() cris/kgdb: Remove sections protected by #ifdef PROCESS_SUPPORT
cris: Provide <asm/kvm_para.h>
cris: Provide inb_p() and outb_p()
* cris: Switch to asm-generic/linkage.h
* cris: Wire up asm-generic/xor.h
cris: Switch cris to drivers/Kconfig
console: Disable VGA text console support on cris
parport: disable PC-style parallel port support on cris
cris: Wire up asm-generic/vga.h
frv: head.S - Remove commented-out initialization code
h8300: Hardcode symbol prefixes in asm sources
* h8300: Fix <asm/tlb.h>
h8300: Fill the system call table using a CALL() macro
h8300: Wire up asm-generic/xor.h
h8300: Limit timer channel ranges in Kconfig
h8300: Switch h8300 to drivers/Kconfig
h8300: Mark H83002 and H83048 CPU support broken
openrisc: Wire up asm-generic/xor.h
score: Remove unneeded <asm/dma-mapping.h>
score: Wire up asm-generic/xor.h
* console/font: Refactor font support code selection logic
* ntp: Remove unused variable flags in __hardpps
* input: cros_ec_keyb_clear_keyboard() depends on CONFIG_PM_SLEEP

Jiang Liu (1):
h8300: add missing definition for read_barries_depends()

 arch/avr32/mach-at32ap/at32ap700x.c              |    2 +
 arch/c6x/include/asm/Kbuild                      |    1 +
 arch/cris/Kconfig                                |   32 +-
 arch/cris/arch-v10/kernel/kgdb.c                 |  870 +++++++---------------
 arch/cris/arch-v32/drivers/Kconfig               |    7 -
 arch/cris/include/asm/Kbuild                     |    4 +
 arch/cris/include/asm/io.h                       |    3 +
 arch/cris/include/asm/linkage.h                  |    6 -
 arch/frv/kernel/head.S                           |    5 -
 arch/h8300/Kconfig                               |  118 +---
 arch/h8300/Kconfig.cpu                           |    4 +
 arch/h8300/include/asm/Kbuild                    |    2 +
 arch/h8300/include/asm/barrier.h                 |    2 +
 arch/h8300/include/asm/linkage.h                 |    6 -
 arch/h8300/include/asm/tlb.h                     |   15 -
 arch/h8300/kernel/entry.S                        |  118 ++--
 arch/h8300/kernel/syscalls.S                     |  648 ++++++++--------
 arch/h8300/lib/abs.S                             |    4 +-
 arch/h8300/lib/memcpy.S                          |    4 +-
 arch/h8300/lib/memset.S                          |    4 +-
 arch/h8300/platform/h8300h/aki3068net/crt0_ram.S |   16 +-
 arch/h8300/platform/h8300h/generic/crt0_ram.S    |   14 +-
 arch/h8300/platform/h8300h/generic/crt0_rom.S    |   14 +-
 arch/h8300/platform/h8300h/h8max/crt0_ram.S      |   16 +-
 arch/h8300/platform/h8s/edosk2674/crt0_ram.S     |   16 +-
 arch/h8300/platform/h8s/edosk2674/crt0_rom.S     |   14 +-
 arch/h8300/platform/h8s/generic/crt0_ram.S       |   16 +-
 arch/h8300/platform/h8s/generic/crt0_rom.S       |   12 +-
 arch/openrisc/include/asm/Kbuild                 |    1 +
 arch/score/include/asm/Kbuild                    |    1 +
 arch/score/include/asm/dma-mapping.h             |    6 -
 drivers/input/keyboard/cros_ec_keyb.c            |   54 +-
 drivers/media/platform/Kconfig                   |    2 +-
 drivers/parport/Kconfig                          |    4 +-
 drivers/staging/media/solo6x10/Kconfig           |    2 +
 drivers/usb/misc/sisusbvga/Kconfig               |    1 +
 drivers/video/console/Kconfig                    |   16 +-
 drivers/video/console/Makefile                   |   14 +-
 kernel/time/ntp.c                                |    1 -
 39 files changed, 789 insertions(+), 1286 deletions(-)
 delete mode 100644 arch/cris/include/asm/linkage.h
 delete mode 100644 arch/h8300/include/asm/linkage.h
 delete mode 100644 arch/score/include/asm/dma-mapping.h

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

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

#1 lore raw
Quoted 6 lines · Geert Uytterhoeven

I created a branch where I collected several exotic architecture fixes (and a few
other trivial fixes):

git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k.git exotic-arch-fixes

Quoted 7 lines

Several of these have been ack'ed (cfr. *-indicators below). If you reject any of the commits, please scream. If you want to take any of them yourself, please do so.

I intend to merge this into my for-next branch after the release of v3.10-rc4, and ask Linus to pull after the release of v3.10-rc5.
This branch may be rebased.

Rebased against v3.10-rc3.

Quoted 2 lines

Geert Uytterhoeven (32):
cris: Switch cris to drivers/Kconfig

JFYI, this has a trivial conflict with "CRIS: drop unused Kconfig symbols" in -next.

Quoted 1 line

* console/font: Refactor font support code selection logic

I inserted a revert of Arnd's 8c090cfbf980581454ae4caae731574fedd7dce8 ("staging/solo6x10: depend on CONFIG_FONTS") before this one.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

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

#2 lore raw
Quoted 6 lines · Geert Uytterhoeven
Quoted 4 lines · Geert Uytterhoeven

I created a branch where I collected several exotic architecture fixes (and a few
other trivial fixes):

Quoted 8 lines
Quoted 8 lines

git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k.git exotic-arch-fixes

Several of these have been ack'ed (cfr. *-indicators below). If you reject any of the commits, please scream. If you want to take any of them yourself, please do so.

I intend to merge this into my for-next branch after the release of v3.10-rc4, and ask Linus to pull after the release of v3.10-rc5.

I merged this in the for-next branch of my repo.

Quoted 5 lines
Quoted 2 lines

Geert Uytterhoeven (32):
cris: Switch cris to drivers/Kconfig

JFYI, this has a trivial conflict with "CRIS: drop unused Kconfig symbols" in -next.

Stephen: cris/for-next and m68k/for-next removed two different hunks from arch/cris/arch-v32/drivers/Kconfig. To resolve the conflict, just remove both. Thanks!

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

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

Hi Geert,

[Chen Liqin <[email protected]> removed from cc as that domain appears to not work :-(]

Quoted 11 lines · On Sun, 2 Jun 2013 21:39:38 +0200 Geert Uytterhoeven <[email protected]>…
Quoted 8 lines · Geert Uytterhoeven
Quoted 6 lines · Geert Uytterhoeven

I created a branch where I collected several exotic architecture fixes (and a few
other trivial fixes):

git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k.git exotic-arch-fixes

Quoted 8 lines
Quoted 6 lines
Quoted 6 lines

Several of these have been ack'ed (cfr. *-indicators below). If you reject any of the commits, please scream. If you want to take any of them yourself, please do so.

I intend to merge this into my for-next branch after the release of v3.10-rc4, and ask Linus to pull after the release of v3.10-rc5.

I merged this in the for-next branch of my repo.

Surely, your for-linus branch would have been a better choice. And then maybe merge that into your for-next branch if you really needed these fixes to continue with m68k work?

Quoted 9 lines
Quoted 5 lines
Quoted 2 lines

Geert Uytterhoeven (32):
cris: Switch cris to drivers/Kconfig

JFYI, this has a trivial conflict with "CRIS: drop unused Kconfig symbols" in -next.

Stephen: cris/for-next and m68k/for-next removed two different hunks from arch/cris/arch-v32/drivers/Kconfig. To resolve the conflict, just remove both. Thanks!

Thanks, fixed it up.

Signature · 2 lines
Cheers,
Stephen Rothwell                    [email protected]
#4 lore raw
Quoted 15 lines · Stephen Rothwell
Quoted 14 lines · On Sun, 2 Jun 2013 21:39:38 +0200 Geert Uytterhoeven <[email protected]>…
Quoted 12 lines · Geert Uytterhoeven
Quoted 10 lines · Geert Uytterhoeven

I created a branch where I collected several exotic architecture fixes (and a few
other trivial fixes):

git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k.git exotic-arch-fixes

Several of these have been ack'ed (cfr. *-indicators below). If you reject any of the commits, please scream. If you want to take any of them yourself, please do so.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines
Quoted 2 lines

I intend to merge this into my for-next branch after the release of v3.10-rc4, and ask Linus to pull after the release of v3.10-rc5.

I merged this in the for-next branch of my repo.

Surely, your for-linus branch would have been a better choice. And then maybe merge that into your for-next branch if you really needed these fixes to continue with m68k work?

An updated version is now in for-linus, as Linus has pulled the m68k content on the
old for-linus branch.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

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

#5 lore raw

PATCH 1/1 arch/cris: old patch to use hex2bin

In reply to Geert Uytterhoeven

Geert, would you like to add following patch to your tree as well?

Andy Shevchenko (1):
cris: kgdb: use native hex2bin

 arch/cris/arch-v10/kernel/kgdb.c | 50 ++++++----------------------------------
 1 file changed, 7 insertions(+), 43 deletions(-)
Signature · 1 line
1.8.2.rc0.22.gb3600c3
#6 lore raw

PATCH 1/1 cris: kgdb: use native hex2bin

---
 arch/cris/arch-v10/kernel/kgdb.c | 50 ++++++----------------------------------
 1 file changed, 7 insertions(+), 43 deletions(-)
arch/cris/arch-v10/kernel/kgdb.c +7 -43
index 37e6d2c..07bccd8 100644
@@ -278,19 +278,11 @@ void putDebugChar (int val);
 
 void enableDebugIRQ (void);
 
-/* Returns the integer equivalent of a hexadecimal character. */
-static int hex (char ch);
-
 /* Convert the memory, pointed to by mem into hexadecimal representation.
    Put the result in buf, and return a pointer to the last character
    in buf (null). */
 static char *mem2hex (char *buf, unsigned char *mem, int count);
 
-/* Convert the array, in hexadecimal representation, pointed to by buf into
-   binary representation. Put the result in mem, and return a pointer to
-   the character after the last byte written. */
-static unsigned char *hex2mem (unsigned char *mem, char *buf, int count);
-
 /* Put the content of the array, in binary representation, pointed to by buf
    into memory pointed to by mem, and return a pointer to
    the character after the last byte written. */
@@ -560,7 +552,7 @@ write_register (int regno, char *val)
 
         if (regno >= R0 && regno <= PC) {
 		/* 32-bit register with simple offset. */
-		hex2mem ((unsigned char *)current_reg + regno * sizeof(unsigned int),
+		hex2bin((unsigned char *)current_reg + regno * sizeof(unsigned int),
 			 val, sizeof(unsigned int));
 	}
         else if (regno == P0 || regno == VR || regno == P4 || regno == P8) {
@@ -570,12 +562,12 @@ write_register (int regno, char *val)
         else if (regno == CCR) {
 		/* 16 bit register with complex offset. (P4 is read-only, P6 is not implemented, 
                    and P7 (MOF) is 32 bits in ETRAX 100LX. */
-		hex2mem ((unsigned char *)&(current_reg->ccr) + (regno-CCR) * sizeof(unsigned short),
+		hex2bin((unsigned char *)&(current_reg->ccr) + (regno-CCR) * sizeof(unsigned short),
 			 val, sizeof(unsigned short));
 	}
 	else if (regno >= MOF && regno <= USP) {
 		/* 32 bit register with complex offset.  (P8 has been taken care of.) */
-		hex2mem ((unsigned char *)&(current_reg->ibr) + (regno-IBR) * sizeof(unsigned int),
+		hex2bin((unsigned char *)&(current_reg->ibr) + (regno-IBR) * sizeof(unsigned int),
 			 val, sizeof(unsigned int));
 	} 
         else {
@@ -595,7 +587,7 @@ write_stack_register (int thread_id, int regno, char *valptr)
 	stack_registers *d = (stack_registers *)stack_list[thread_id];
 	unsigned int val;
 	
-	hex2mem ((unsigned char *)&val, valptr, sizeof(unsigned int));
+	hex2bin((unsigned char *)&val, valptr, sizeof(unsigned int));
 	if (regno >= R0 && regno < SP) {
 		d->r[regno] = val;
 	}
@@ -659,18 +651,6 @@ read_register (char regno, unsigned int *valptr)
 }
 
 /********************************** Packet I/O ******************************/
-/* Returns the integer equivalent of a hexadecimal character. */
-static int
-hex (char ch)
-{
-	if ((ch >= 'a') && (ch <= 'f'))
-		return (ch - 'a' + 10);
-	if ((ch >= '0') && (ch <= '9'))
-		return (ch - '0');
-	if ((ch >= 'A') && (ch <= 'F'))
-		return (ch - 'A' + 10);
-	return (-1);
-}
 
 /* Convert the memory, pointed to by mem into hexadecimal representation.
    Put the result in buf, and return a pointer to the last character
@@ -703,22 +683,6 @@ mem2hex(char *buf, unsigned char *mem, int count)
 	return (buf);
 }
 
-/* Convert the array, in hexadecimal representation, pointed to by buf into
-   binary representation. Put the result in mem, and return a pointer to
-   the character after the last byte written. */
-static unsigned char*
-hex2mem (unsigned char *mem, char *buf, int count)
-{
-	int i;
-	unsigned char ch;
-	for (i = 0; i < count; i++) {
-		ch = hex (*buf++) << 4;
-		ch = ch + hex (*buf++);
-		*mem++ = ch;
-	}
-	return (mem);
-}
-
 /* Put the content of the array, in binary representation, pointed to by buf
    into memory pointed to by mem, and return a pointer to the character after
    the last byte written.
@@ -964,7 +928,7 @@ handle_exception (int sigval)
 				   Success: OK
 				   Failure: void. */
 #ifdef PROCESS_SUPPORT
-				hex2mem ((unsigned char *)&reg_g, &remcomInBuffer[1], sizeof(registers));
+				hex2bin((unsigned char *)&reg_g, &remcomInBuffer[1], sizeof(registers));
 				if (current_thread_g == executing_task) {
 					copy_registers (&reg, &reg_g, sizeof(registers));
 				}
@@ -972,7 +936,7 @@ handle_exception (int sigval)
 					copy_registers_to_stack(current_thread_g, &reg_g);
 				}
 #else
-				hex2mem((char *)&reg, &remcomInBuffer[1], sizeof(registers));
+				hex2bin((char *)&reg, &remcomInBuffer[1], sizeof(registers));
 #endif
 				gdb_cris_strcpy (remcomOutBuffer, "OK");
 				break;
@@ -1053,7 +1017,7 @@ handle_exception (int sigval)
 					int length = gdb_cris_strtol(lenptr+1, &dataptr, 16);
 					if (*lenptr == ',' && *dataptr == ':') {
 						if (remcomInBuffer[0] == 'M') {
-							hex2mem(addr, dataptr + 1, length);
+							hex2bin(addr, dataptr + 1, length);
 						}
 						else /* X */ {
 							bin2mem(addr, dataptr + 1, length);
Signature · 1 line
1.8.2.rc0.22.gb3600c3
#7 lore raw

Re: [PATCH] cris: kgdb: use native hex2bin

Quoted 11 lines · Andy Shevchenko
arch/cris/arch-v10/kernel/kgdb.c +0 -2
index 37e6d2c..07bccd8 100644
@@ -659,18 +651,6 @@ read_register (char regno, unsigned int *valptr)
 }
 
 /********************************** Packet I/O ******************************/
-/* Returns the integer equivalent of a hexadecimal character. */
-static int

arch/cris/arch-v10/kernel/kgdb.c read_register (char regno, unsigned int *valptr) · line 673

Quoted 10 lines
arch/cris/arch-v10/kernel/kgdb.c +0 -10
-hex (char ch)
-{
-       if ((ch >= 'a') && (ch <= 'f'))
-               return (ch - 'a' + 10);
-       if ((ch >= '0') && (ch <= '9'))
-               return (ch - '0');
-       if ((ch >= 'A') && (ch <= 'F'))
-               return (ch - 'A' + 10);
-       return (-1);
-}

This one handles both upper and lower case hex characters, while hex2bin() handles lower case only. Is that OK, or should hex2bin be extended?

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

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

#8 lore raw

Re: [PATCH] cris: kgdb: use native hex2bin

Quoted 10 lines · Geert Uytterhoeven
Quoted 8 lines · Andy Shevchenko
arch/cris/arch-v10/kernel/kgdb.c +0 -0
index 37e6d2c..07bccd8 100644
@@ -659,18 +651,6 @@ read_register (char regno, unsigned int *valptr)
 }
 
Quoted 16 lines

arch/cris/arch-v10/kernel/kgdb.c read_register (char regno, unsigned int *valptr) · line 673

Quoted 13 lines
arch/cris/arch-v10/kernel/kgdb.c +0 -12
 /********************************** Packet I/O ******************************/
-/* Returns the integer equivalent of a hexadecimal character. */
-static int
-hex (char ch)
-{
-       if ((ch >= 'a') && (ch <= 'f'))
-               return (ch - 'a' + 10);
-       if ((ch >= '0') && (ch <= '9'))
-               return (ch - '0');
-       if ((ch >= 'A') && (ch <= 'F'))
-               return (ch - 'A' + 10);
-       return (-1);
-}

This one handles both upper and lower case hex characters, while hex2bin() handles lower case only. Is that OK, or should hex2bin be extended?

Its implementation is located in lib/hexdump.c
If you look at it carefully you find that hex2bin calls hex_to_bin() which services both cases (via using tolower()).

--
With Best Regards,
Andy Shevchenko