[PATCH] hexagon: include: asm: kgdb: extend DBG_MAX_REG_NUM for "cs0/1"

[PATCH] hexagon: include: asm: kgdb: extend DBG_MAX_REG_NUM for "cs0/1"

[Chen Gang]

Need extend maximized number for "cs0/1", the related warning (with
allmodconfig for v4):

  arch/hexagon/kernel/kgdb.c:79: warning: excess elements in array initializer
  arch/hexagon/kernel/kgdb.c:79: warning: (near initialization for 'dbg_reg_def')
  arch/hexagon/kernel/kgdb.c:80: warning: excess elements in array initializer
  arch/hexagon/kernel/kgdb.c:80: warning: (near initialization for 'dbg_reg_def')


Signed-off-by: Chen Gang <gang.chen@asianux.com>
---
 arch/hexagon/include/asm/kgdb.h |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/arch/hexagon/include/asm/kgdb.h b/arch/hexagon/include/asm/kgdb.h
index 32a6fb6..ccd3ac3 100644
--- a/arch/hexagon/include/asm/kgdb.h
+++ b/arch/hexagon/include/asm/kgdb.h
@@ -34,10 +34,11 @@ static inline void arch_kgdb_breakpoint(void)
  * 32 gpr + sa0/1 + lc0/1 + m0/1 + gp + ugp + pred + pc = 42 total.
  * vm regs = psp+elr+est+badva = 4
  * syscall+restart = 2 more
- * so 48 = 42 +4 + 2
+ * also add cs0/1 = 2
+ * so 48 = 42 + 4 + 2 + 2
  */
 #define DBG_USER_REGS 42
-#define DBG_MAX_REG_NUM (DBG_USER_REGS + 6)
+#define DBG_MAX_REG_NUM (DBG_USER_REGS + 8)
 #define NUMREGBYTES  (DBG_MAX_REG_NUM*4)
 
 #endif /* __HEXAGON_KGDB_H__ */
-- 
1.7.7.6

[Suggestion] are they toolchains' issue?

[Chen Gang]

Hello Maintainers:

When compiling with allmodconfig for v4, an error occurs:

    CC      kernel/fork.o
  /tmp/ccTx69dj.s: Assembler messages:
  /tmp/ccTx69dj.s:7853: Error: invalid instruction `if (tstbit(r4.new,r0))jump:nt .L242'

Also when with allmodconfig (default for v2), it will cause error:

  /usr/local/bin/hexagon-ld: Relocatable linking with relocations from format elf32-littlehexagon (init/.tmp_main.o) to format elf32-littlehexagon (init/main.o) is not supported

Also when with defconfig, it will cause error:

    LD      init/mounts.o
  /usr/local/bin/hexagon-ld: Relocatable linking with relocations from format elf32-littlehexagon (init/do_mounts.o) to format elf32-littlehexagon (init/mounts.o) is not supported


Toolchain Information:

  https://www.codeaurora.org/patches/quic/hexagon/dev/
  gcc-4.4_full_111104.bz2 and gnutools_full_111104.bz2
  
  [root@gchenlinux hexagon]# hexagon-gcc -v
  Using built-in specs.
  Target: hexagon
  Configured with: ../gcc-4.4/configure --without-header --disable-nls --enable-language=c --disable-threads --disable-shared --enable-werror=no target_configargs=enable_vtable_verify=yes --target=hexagon
  Thread model: single
  gcc version 4.4.0 : Hexagon Build Version (UNKNOWN) UNKNOWN (GCC) 
  [root@gchenlinux hexagon]# hexagon-ld -v
  GNU ld (GNU Binutils) 2.20 20091016
  Hexagon Build Version X.Y.ZZ XXXXXXXXXXX_YYYY_ZZZZZ
  [root@gchenlinux hexagon]# hexagon-as -v
  GNU assembler version 2.20 (hexagon) using BFD version (GNU Binutils) 2.20 20091016
  Hexagon Build Version X.Y.ZZ XXXXXXXXXXX_YYYY_ZZZZZ


Please help check, and welcome any suggestions or completions.


Thanks.
--
Chen Gang.

Re: [PATCH] hexagon: include: asm: kgdb: extend DBG_MAX_REG_NUM for "cs0/1"

[Linas Vepstas]

On 1 November 2013 06:55, Chen Gang <gang.chen@asianux.com> wrote:
> Need extend maximized number for "cs0/1", the related warning (with
> allmodconfig for v4):

Looks ok to me.

Acked-by: Linas Vepstas <linasvepstas@gmail.com>

Re: [Suggestion] are they toolchains' issue?

[rkuo]

On Fri, Nov 01, 2013 at 08:05:52PM +0800, Chen Gang wrote:
> Hello Maintainers:
> 
> When compiling with allmodconfig for v4, an error occurs:
> 
>     CC      kernel/fork.o
>   /tmp/ccTx69dj.s: Assembler messages:
>   /tmp/ccTx69dj.s:7853: Error: invalid instruction `if (tstbit(r4.new,r0))jump:nt .L242'
> 
> Also when with allmodconfig (default for v2), it will cause error:
> 
>   /usr/local/bin/hexagon-ld: Relocatable linking with relocations from format elf32-littlehexagon (init/.tmp_main.o) to format elf32-littlehexagon (init/main.o) is not supported
> 
> Also when with defconfig, it will cause error:
> 
>     LD      init/mounts.o
>   /usr/local/bin/hexagon-ld: Relocatable linking with relocations from format elf32-littlehexagon (init/do_mounts.o) to format elf32-littlehexagon (init/mounts.o) is not supported
> 

Hi Chen,

Thanks for checking all these.  There are newer Hexagon tools available
here:

https://sourcery.mentor.com/GNUToolchain/release2568?cmpid=7108

You might want to try those out.  The tool prefix is hexagon-linux-, but
aside from that, a v4 defconfig on 3.12-rc7 compiled for me.

'allmodconfig' does not appear to compile due to DMA related issues...

Anyways, would you mind giving the newer tools a try?


Thanks,
Richard Kuo


-- 

Sent by an employee of the Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation

Re: [PATCH] hexagon: include: asm: kgdb: extend DBG_MAX_REG_NUM for "cs0/1"

[Chen Gang]

On 11/02/2013 01:14 AM, Linas Vepstas wrote:
> On 1 November 2013 06:55, Chen Gang <gang.chen@asianux.com> wrote:
>> Need extend maximized number for "cs0/1", the related warning (with
>> allmodconfig for v4):
> 
> Looks ok to me.
> 
> Acked-by: Linas Vepstas <linasvepstas@gmail.com>
> 
> 

Thanks.
-- 
Chen Gang

Re: [Suggestion] are they toolchains' issue?

[Chen Gang]

On 11/02/2013 03:42 AM, rkuo wrote:
> On Fri, Nov 01, 2013 at 08:05:52PM +0800, Chen Gang wrote:
>> Hello Maintainers:
>>
>> When compiling with allmodconfig for v4, an error occurs:
>>
>>     CC      kernel/fork.o
>>   /tmp/ccTx69dj.s: Assembler messages:
>>   /tmp/ccTx69dj.s:7853: Error: invalid instruction `if (tstbit(r4.new,r0))jump:nt .L242'
>>
>> Also when with allmodconfig (default for v2), it will cause error:
>>
>>   /usr/local/bin/hexagon-ld: Relocatable linking with relocations from format elf32-littlehexagon (init/.tmp_main.o) to format elf32-littlehexagon (init/main.o) is not supported
>>
>> Also when with defconfig, it will cause error:
>>
>>     LD      init/mounts.o
>>   /usr/local/bin/hexagon-ld: Relocatable linking with relocations from format elf32-littlehexagon (init/do_mounts.o) to format elf32-littlehexagon (init/mounts.o) is not supported
>>
> 
> Hi Chen,
> 
> Thanks for checking all these.  There are newer Hexagon tools available
> here:
> 
> https://sourcery.mentor.com/GNUToolchain/release2568?cmpid=7108
> 
> You might want to try those out.  The tool prefix is hexagon-linux-, but
> aside from that, a v4 defconfig on 3.12-rc7 compiled for me.
> 

Thank you for your information.

> 'allmodconfig' does not appear to compile due to DMA related issues...
> 
> Anyways, would you mind giving the newer tools a try?
> 
> 

I will/should try. try to let our hexagon pass 'allmodconfig' within
this month (2013-11-30).


Thanks.
-- 
Chen Gang

Re: [Suggestion] are they toolchains' issue?

[Chen Gang]

On 11/02/2013 09:29 PM, Chen Gang wrote:
> On 11/02/2013 03:42 AM, rkuo wrote:
>>
>> Thanks for checking all these.  There are newer Hexagon tools available
>> here:
>>
>> https://sourcery.mentor.com/GNUToolchain/release2568?cmpid=7108
>>
>> You might want to try those out.  The tool prefix is hexagon-linux-, but
>> aside from that, a v4 defconfig on 3.12-rc7 compiled for me.
>>
> 
> Thank you for your information.
> 

I finish building the new version of gcc and binutils for kernel. And
now it can continue -- no related issues which mentioned originally
(should be fixed in this new version).

>> 'allmodconfig' does not appear to compile due to DMA related issues...
>>
>> Anyways, would you mind giving the newer tools a try?
>>
>>
> 
> I will/should try. try to let our hexagon pass 'allmodconfig' within
> this month (2013-11-30).
> 

I am continuing, hope we can finish within this month. :-)

Thanks.
-- 
Chen Gang

Re: [Suggestion] are they toolchains' issue?

[Chen Gang]

Hello Maintainers:

I met an issue for linux next-20131105

The related error:

    CC [M]  fs/ubifs/lpt_commit.o
  fs/ubifs/lpt_commit.c: In function 'next_pnode_to_dirty':
  fs/ubifs/lpt_commit.c:620:1: internal compiler error: in hwloop_optimize, at config/hexagon/hexagon.c:5796
  Please submit a full bug report,
  with preprocessed source if appropriate.
  See <http://gcc.gnu.org/bugs.html> for instructions.

The related information:

  [root@gchenlinux linux-next]# hexagon-linux-gcc -v
  Using built-in specs.
  COLLECT_GCC=hexagon-linux-gcc
  COLLECT_LTO_WRAPPER=/usr/local/libexec/gcc/hexagon-linux/4.6.1/lto-wrapper
  Target: hexagon-linux
  Configured with: ../gcc-hexagon-2012.03/configure --without-header --disable-nls --enable-language=c --disable-threads --disable-shared --enable-werror=no target_configargs=enable_vtable_verify=yes --target=hexagon-linux
  Thread model: single
  gcc version 4.6.1 (GCC) 
  [root@gchenlinux linux-next]# hexagon-linux-ld -v
  GNU ld (GNU Binutils) 2.21.53.20110905
  [root@gchenlinux linux-next]# hexagon-linux-as -v
  GNU assembler version 2.21.53 (hexagon-linux) using BFD version (GNU Binutils) 2.21.53.20110905

Welcome any suggestions or completions.

Thanks.

On 11/06/2013 10:56 AM, Chen Gang wrote:
> On 11/02/2013 09:29 PM, Chen Gang wrote:
>> On 11/02/2013 03:42 AM, rkuo wrote:
>>>
>>> Thanks for checking all these.  There are newer Hexagon tools available
>>> here:
>>>
>>> https://sourcery.mentor.com/GNUToolchain/release2568?cmpid=7108
>>>
>>> You might want to try those out.  The tool prefix is hexagon-linux-, but
>>> aside from that, a v4 defconfig on 3.12-rc7 compiled for me.
>>>
>>
>> Thank you for your information.
>>
> 
> I finish building the new version of gcc and binutils for kernel. And
> now it can continue -- no related issues which mentioned originally
> (should be fixed in this new version).
> 
>>> 'allmodconfig' does not appear to compile due to DMA related issues...
>>>
>>> Anyways, would you mind giving the newer tools a try?
>>>
>>>
>>
>> I will/should try. try to let our hexagon pass 'allmodconfig' within
>> this month (2013-11-30).
>>
> 
> I am continuing, hope we can finish within this month. :-)
> 
> Thanks.
> 


-- 
Chen Gang

Re: [Suggestion] are they toolchains' issue?

[rkuo]

On Fri, Nov 08, 2013 at 05:30:02PM +0800, Chen Gang wrote:
> Hello Maintainers:
> 
> I met an issue for linux next-20131105
> 
> The related error:
> 
>     CC [M]  fs/ubifs/lpt_commit.o
>   fs/ubifs/lpt_commit.c: In function 'next_pnode_to_dirty':
>   fs/ubifs/lpt_commit.c:620:1: internal compiler error: in hwloop_optimize, at config/hexagon/hexagon.c:5796
>   Please submit a full bug report,
>   with preprocessed source if appropriate.
>   See <http://gcc.gnu.org/bugs.html> for instructions.
> 
> The related information:
> 
>   [root@gchenlinux linux-next]# hexagon-linux-gcc -v
>   Using built-in specs.
>   COLLECT_GCC=hexagon-linux-gcc
>   COLLECT_LTO_WRAPPER=/usr/local/libexec/gcc/hexagon-linux/4.6.1/lto-wrapper
>   Target: hexagon-linux
>   Configured with: ../gcc-hexagon-2012.03/configure --without-header --disable-nls --enable-language=c --disable-threads --disable-shared --enable-werror=no target_configargs=enable_vtable_verify=yes --target=hexagon-linux
>   Thread model: single
>   gcc version 4.6.1 (GCC) 
>   [root@gchenlinux linux-next]# hexagon-linux-ld -v
>   GNU ld (GNU Binutils) 2.21.53.20110905
>   [root@gchenlinux linux-next]# hexagon-linux-as -v
>   GNU assembler version 2.21.53 (hexagon-linux) using BFD version (GNU Binutils) 2.21.53.20110905
> 
> Welcome any suggestions or completions.
> 
> Thanks.
> 

That's definitely a bug; I reproduced it over here as well.  I've submitted 
it to the tools team; I will let you know anything else I find.


Thanks for reporting this!

-Richard Kuo


-- 

Sent by an employee of the Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation

Re: [Suggestion] are they toolchains' issue?

[Chen Gang]

On 11/10/2013 08:28 AM, rkuo wrote:
> That's definitely a bug; I reproduced it over here as well.  I've submitted 
> it to the tools team; I will let you know anything else I find.
> 
> 
> Thanks for reporting this!

Thank you too, I will/should continue for allmodconfig (at present, I
will/should by pass this issue, firstly).


Thanks.
-- 
Chen Gang