gmane.linux.ports.openrisc

Re: [PATCH 13/21] openrisc: delay: remove inline marking of EXPORT_SYMBOL functions

Will Deacon — 2013-05-09T10:45:44

Acked-by: Will Deacon <will.deacon< at >arm.com>

Will

[PATCH] serio/opencores_ps2: include platform_device.h

Stefan Kristiansson — 2013-05-02T18:10:52

Fixes build issues with 'make ARCH=i386 allyesconfig'

Signed-off-by: Stefan Kristiansson <stefan.kristiansson< at >saunalahti.fi>
---
 drivers/input/serio/opencores_ps2.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/input/serio/opencores_ps2.c b/drivers/input/serio/opencores_ps2.c
index 2dca9fc..e7e24c8 100644
--- a/drivers/input/serio/opencores_ps2.c
+++ b/drivers/input/serio/opencores_ps2.c
< at >< at > -12,6 +12,7 < at >< at >
 #include <linux/module.h>
 #include <linux/serio.h>
 #include <linux/interrupt.h>
+#include <linux/platform_device.h>
 #include <linux/errno.h>
 #include <linux/slab.h>
 #include <linux/init.h>

[PATCH v2] openrisc: call do_notify_resume() withinterrupts enabled

Stefan Kristiansson — 2013-04-29T10:40:32

A signal delivered through do_notify_resume() would cause the
irqs_disabled() check in _local_bh_enable_ip() to be triggered.

Enable interrupts before calling do_notify_resume(), but read
the current_thread_info->flags before doing so.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson< at >saunalahti.fi>
---
 arch/openrisc/kernel/entry.S  | 5 +++--
 arch/openrisc/kernel/signal.c | 7 ++++---
 2 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/arch/openrisc/kernel/entry.S b/arch/openrisc/kernel/entry.S
index d8a455e..7564784 100644
--- a/arch/openrisc/kernel/entry.S
+++ b/arch/openrisc/kernel/entry.S
< at >< at > -858,8 +858,8 < at >< at > _work_pending:
  * if (current_thread_info->flags & _TIF_NEED_RESCHED)
  *     schedule();
  */
-l.lwz   r5,TI_FLAGS(r10)
-l.andir3,r5,_TIF_NEED_RESCHED
+l.lwz   r4,TI_FLAGS(r10)
+l.andir3,r4,_TIF_NEED_RESCHED
 l.sfnei r3,0
 l.bnf   _work_notifysig
  l.nop
< at >< at > -874,6 +874,7 < at >< at > _work_pending:
  * must be set so that the syscall restart functionality works.
  */
 _work_n

Re: [PATCH] openrisc: call do_notify_resume() withinterrupts enabled

Stefan Kristiansson — 2013-04-29T09:50:48

I just realised that this one-liner isn't quite enough,
the flags that are used in do_notify_resume() has to be read with
interrupts off.
Revised patch coming up.

Stefan

[PATCH] openrisc: call do_notify_resume() with interruptsenabled

Stefan Kristiansson — 2013-04-29T07:12:37

A signal delivered through do_notify_resume() would cause the
irqs_disabled() check in _local_bh_enable_ip() to be triggered.

Enable interrupts before calling do_notify_resume().

Signed-off-by: Stefan Kristiansson <stefan.kristiansson< at >saunalahti.fi>
---
 arch/openrisc/kernel/entry.S | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/openrisc/kernel/entry.S b/arch/openrisc/kernel/entry.S
index d8a455e..55ffc97 100644
--- a/arch/openrisc/kernel/entry.S
+++ b/arch/openrisc/kernel/entry.S
< at >< at > -874,6 +874,7 < at >< at > _work_pending:
  * must be set so that the syscall restart functionality works.
  */
 _work_notifysig:
+ENABLE_INTERRUPTS(r29)
 l.jaldo_notify_resume
  l.orir3,r1,0  /* pt_regs */

[PATCH] openrisc: add cache way information to cpuinfo

Stefan Kristiansson — 2013-04-27T18:02:32

Motivation for this is to be able to print the way information
properly in print_cpuinfo(), instead of hardcoding it to one.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson< at >saunalahti.fi>
---
 arch/openrisc/include/asm/cpuinfo.h |  2 ++
 arch/openrisc/kernel/setup.c        | 20 +++++++++++++-------
 2 files changed, 15 insertions(+), 7 deletions(-)

diff --git a/arch/openrisc/include/asm/cpuinfo.h b/arch/openrisc/include/asm/cpuinfo.h
index 917318b..ec10679 100644
--- a/arch/openrisc/include/asm/cpuinfo.h
+++ b/arch/openrisc/include/asm/cpuinfo.h
< at >< at > -24,9 +24,11 < at >< at > struct cpuinfo {
 
 u32 icache_size;
 u32 icache_block_size;
+u32 icache_ways;
 
 u32 dcache_size;
 u32 dcache_block_size;
+u32 dcache_ways;
 };
 
 extern struct cpuinfo cpuinfo;
diff --git a/arch/openrisc/kernel/setup.c b/arch/openrisc/kernel/setup.c
index f4d5bed..6b0e2c4 100644
--- a/arch/openrisc/kernel/setup.c
+++ b/arch/openrisc/kernel/setup.c
< at >< at > -129,13 +129,15 < at >< at > static void print_cpuinfo(void)
 if (upr & SPR_UPR_DCP)
 pr

Re: KALLSYMS

Marek Czerski — 2013-04-18T06:54:13

ok, i lied a little bit about not knowing where to go :P
after some digging i came up with:

diff --git a/arch/openrisc/kernel/vmlinux.lds.S
b/arch/openrisc/kernel/vmlinux.lds.S
index 2d69a85..77d609d 100644
--- a/arch/openrisc/kernel/vmlinux.lds.S
+++ b/arch/openrisc/kernel/vmlinux.lds.S
< at >< at > -41,6 +41,7 < at >< at > SECTIONS
  /* _s_kernel_ro must be page aligned */
  . = ALIGN(PAGE_SIZE);
  _s_kernel_ro = .;
+ _text = .;

         .text                   : AT(ADDR(.text) - LOAD_OFFSET)
  {

it seems to solve the problem, kernel compiles and i can get symbol
description during kernel oops. other architectures do it in the similar
way.

cheers,
Marek


2013/4/17 Marek Czerski <ma.czerski< at >gmail.com>

KALLSYMS

Marek Czerski — 2013-04-17T21:55:36

Hello,

i'm using openrisc linux from git://openrisc.net/jonas/linux and i'm trying
to enable CONFIG_KALLSYMS. Unfortunately i'm getting error at the end of
the build:

  LINK    vmlinux
  LD      vmlinux.o
  MODPOST vmlinux.o
  GEN     .version
  CHK     include/generated/compile.h
  UPD     include/generated/compile.h
  CC      init/version.o
  LD      init/built-in.o
  KSYM    .tmp_kallsyms1.o
.tmp_kallsyms1.o: In function `kallsyms_addresses':
(.rodata+0x0): undefined reference to `_text'
.tmp_kallsyms1.o: In function `kallsyms_addresses':
(.rodata+0x4): undefined reference to `_text'
.tmp_kallsyms1.o: In function `kallsyms_addresses':
(.rodata+0x8): undefined reference to `_text'
.tmp_kallsyms1.o: In function `kallsyms_addresses':
(.rodata+0xc): undefined reference to `_text'
.tmp_kallsyms1.o: In function `kallsyms_addresses':
(.rodata+0x10): undefined reference to `_text'
.tmp_kallsyms1.o:(.rodata+0x14): more undefined references to `_text' follow
make: *** [vmlinux] Błąd 1

don't know where to go fr

Re: [PATCH v3, part4 26/39] mm/openrisc: prepare for removing num_physpages and simplify mem_init()

Jonas Bonn — 2013-04-02T05:31:44

Tested and works fine on OpenRISC.

Acked-by: Jonas Bonn <jonas< at >southpole.se>

/Jonas

Re: [PATCH 1/1] openrisc: really pass correct arg toschedule_tail

Jonas Bonn — 2013-02-26T07:12:04

Thanks, guys!  Pushed to master.
/Jonas

head.S: Error: unresolved expression and PATCH

Jack Thomasson — 2013-02-21T00:47:32

i am comparing

http://opencores.org/or1k/OpenRISC_GNU_tool_chain#Installation_of_development_versions

against

http://opencores.org/or1k/Ubuntu_VirtualBox-image_updates_and_information

and trying to build the kernel.  the new toolchain rejects UART_DLL and 
UART_DLM in arch/openrisc/kernel/head.S with "Error: unresolved expression 
that must be resolved".  i searched the linux sources for where these are 
defined, many places do, and think linux/serial_reg.h is the most sensible to 
include.  the following patch does quash the error but i submit for folks more 
knowledgable than i to decide if it is really correct.

--- arch/openrisc/kernel/head.S~        2013-02-20 17:32:11.132491912 -0700
+++ arch/openrisc/kernel/head.S 2013-02-20 17:32:50.824316622 -0700
< at >< at > -26,6 +26,7 < at >< at >
 #include <asm/cache.h>
 #include <asm/spr_defs.h>
 #include <asm/asm-offsets.h>
+#include <linux/serial_reg.h>
 #include <linux/of_fdt.h>
 
 #define tophys(rd,rs)                          \

:{)}

Re: [PATCH 1/1] openrisc: really pass correct arg toschedule_tail

Christian Svensson — 2013-02-22T15:51:33

Confirmed working

Regards,
Christian Svensson



On Fri, Feb 22, 2013 at 3:52 PM, Stefan Kristiansson
<stefan.kristiansson< at >saunalahti.fi> wrote:

Re: [PATCH 1/1] openrisc: really pass correct arg to schedule_tail

Stefan Kristiansson — 2013-02-22T14:52:10

I think the patch below should fix that.

Stefan

-->8--
diff --git a/arch/openrisc/kernel/head.S b/arch/openrisc/kernel/head.S
index 1088b5f..2745196 100644
--- a/arch/openrisc/kernel/head.S
+++ b/arch/openrisc/kernel/head.S
< at >< at > -19,6 +19,7 < at >< at >
 #include <linux/threads.h>
 #include <linux/errno.h>
 #include <linux/init.h>
+#include <linux/serial_reg.h>
 #include <asm/processor.h>
 #include <asm/page.h>
 #include <asm/mmu.h>

Re: [PATCH 1/1] openrisc: really pass correct arg toschedule_tail

Christian Svensson — 2013-02-22T11:50:08

It seems to fix that issue - great!

It did however uncover these two:
arch/openrisc/kernel/head.S:1570: Error: unresolved expression that
must be resolved
arch/openrisc/kernel/head.S:1572: Error: unresolved expression that
must be resolved
1565-
1566-l.lbzr5,3(r3)
1567-l.orir4,r5,0x80
1568-l.sb0x3(r3),r4
1569-l.addir4,r0,((UART_DIVISOR>>8) & 0x000000ff)
1570:l.sbUART_DLM(r3),r4
1571-l.addi  r4,r0,((UART_DIVISOR) & 0x000000ff)
1572-l.sbUART_DLL(r3),r4
1573-l.sb0x3(r3),r5
1574-
1575-l.jrr9

AFAIK those errors exist in current mainline as well. When compiling
3.6.10 I simply added #if 0 to hack around it.
Removing those lines and modifying vmlinux.lds.S to use or1k instead
of or32 successfully compiles the kernel with the new toolchain, so
it's probably the last of these errors :-).

Regards,

Christian Svensson

Re: [PATCH 1/1] remove __devinitdata in ocfb.

Jonas Bonn — 2013-02-18T06:33:11

Thanks.  Applied to master.
/Jonas

[PATCH 1/1] remove __devinitdata in ocfb.

Sebastian Macke — 2013-02-17T00:03:12

__devinitdata seems to be obsolete in Linux 3.8


Signed-off-by: Sebastian Macke <sebastian< at >macke.de>
---
  drivers/video/ocfb.c |    2 +-
  1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/video/ocfb.c b/drivers/video/ocfb.c
index d93c01b..36efdbf 100644
--- a/drivers/video/ocfb.c
+++ b/drivers/video/ocfb.c
< at >< at > -52,7 +52,7 < at >< at >

  #define OCFB_NAME     "OC VGA/LCD"

-static char *mode_option __devinitdata = NULL;
+static char *mode_option = NULL;

  static const struct fb_videomode default_mode = {
      /* 640x480 < at > 60 Hz, 31.5 kHz hsync */

Re: [PATCH 1/1] openrisc: update DTLB-miss handler last

Sebastian Macke — 2013-02-16T20:08:53

Tested-by: Sebastian Macke <sebastian< at >macke.de>

By flushing the tlb before and after the printk commands I could force 
the bug in the old version. Testing it with the new patch no kernel 
panic occurs.

On 13/02/2013 11:25 PM, Jonas Bonn wrote:

Re: [PATCH 1/1] openrisc: really pass correct arg toschedule_tail

Jonas Bonn — 2013-02-16T08:47:51

Pushed to master and for-upstream branches at 
git://openrisc.net/jonas/linux.  This should fix the build issues people 
were reporting.

/Jonas

Re: [PATCH 1/1] openrisc: update DTLB-miss handler last

Julius Baxter — 2013-02-15T17:37:16

Acked-by:  Julius Baxter <juliusbaxter< at >gmail.com>

[PATCH 1/1] openrisc: really pass correct arg toschedule_tail

Jonas Bonn — 2013-02-15T16:48:40

Commit 287ad220cd8b5a9d29f71c78f6e4051093f051fc tried to set up the argument
to schedule_tail, but ended up using TI_STACK which isn't a defined symbol.
Sadly, the old openrisc compiler silently ignores this fact and it was first
discovered now when building with an updated toolchain.

Reported-by: Christian Svensson <blue< at >cmd.nu>
Signed-off-by: Jonas Bonn <jonas< at >southpole.se>
---

Hi Christian,

I think this patch should fix the build issue you were seeing.  It seems the
old toolchain was just silently ignoring the fact that this symbol wasn't
defined...!!!  Anyway, glad to hear that the updated toolchain catches stuff
like this at least.

Thanks for reporting.  Let me know if this works for you (feel free to send
a Tested-by tag, if so).

/Jonas

 arch/openrisc/kernel/entry.S | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/openrisc/kernel/entry.S b/arch/openrisc/kernel/entry.S
index 6b299c4..f6b24dd 100644
--- a/arch/openrisc/kernel/entry.S
+++ b/arch/openrisc/kernel/entry.S
< at >< at > -10

arch/openrisc/kernel/entry.S build error

Christian Svensson — 2013-02-15T09:52:46

Hello.

I have cloned git://openrisc.net/jonas/linux (HEAD is
16c70c64be4ea0f64950f8e13b4b1802dbdef9d7).

When I try to compile this I get the following. A grep for TI_STACK
shows only entry.S.

# ARCH="openrisc" make
  CHK     include/generated/uapi/linux/version.h
  CHK     include/generated/utsrelease.h
  CALL    scripts/checksyscalls.sh
  CHK     include/generated/compile.h
  AS      arch/openrisc/kernel/entry.o
arch/openrisc/kernel/entry.S: Assembler messages:
arch/openrisc/kernel/entry.S:1053: Error: unresolved expression that
must be resolved
make[1]: *** [arch/openrisc/kernel/entry.o] Error 1
make: *** [arch/openrisc/kernel] Error 2

entry.S:
1043-
1044-/* Unwind stack to pre-switch state */
1045-l.addi  r1,r1,(INT_FRAME_SIZE)
1046-
1047-/* Return via the link-register back to where we 'came from', where
1048- * that may be either schedule(), ret_from_fork(), or
1049- * ret_from_kernel_thread().  If we are returning to a new thread,
1050- * we are expected to have set up the arg to schedule_tai