[PATCH] m68k: don't use lea on profiling

[PATCH] m68k: don't use lea on profiling

[Sebastian Andrzej Siewior]

this will not work on large binaries:

|bigeasy@ayoka:~/cc$ echo 'int main(void) { return 0; }' > file.c
|bigeasy@ayoka:~/cc$ m68k-linux-gnu-gcc -o file file.c -static -pg
|/tmp/ccw33VYP.o: In function `main':
|file.c:(.text+0x6): relocation truncated to fit: R_68K_PC16 against `.data'
|collect2: ld returned 1 exit status

This is reported as #36047. It appears that the generated label is unused.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 gcc/config/m68k/linux.h |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/gcc/config/m68k/linux.h b/gcc/config/m68k/linux.h
index 113f278..79e9151 100644
--- a/gcc/config/m68k/linux.h
+++ b/gcc/config/m68k/linux.h
@@ -146,7 +146,6 @@ along with GCC; see the file COPYING3.  If not see
 #undef FUNCTION_PROFILER
 #define FUNCTION_PROFILER(FILE, LABELNO) \
 {									\
-  asm_fprintf (FILE, "\tlea (%LLP%d,%Rpc),%Ra1\n", (LABELNO));		\
   if (flag_pic)								\
     fprintf (FILE, "\tbsr.l _mcount@PLTPC\n");				\
   else									\
-- 
1.6.2.5

Re: [PATCH] m68k: don't use lea on profiling

[Maxim Kuvyrkov]

Sebastian Andrzej Siewior wrote:
> this will not work on large binaries:
> 
> |bigeasy@ayoka:~/cc$ echo 'int main(void) { return 0; }' > file.c
> |bigeasy@ayoka:~/cc$ m68k-linux-gnu-gcc -o file file.c -static -pg
> |/tmp/ccw33VYP.o: In function `main':
> |file.c:(.text+0x6): relocation truncated to fit: R_68K_PC16 against `.data'
> |collect2: ld returned 1 exit status
> 
> This is reported as #36047. It appears that the generated label is unused.

As I noted in the PR <http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36047> 
neither GLIBC nor uClibc need the argument so we may simply remove it 
(thus getting rid of the PC-relative relocation that causes the problem).

Andreas, does this look OK to you?

> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
>  gcc/config/m68k/linux.h |    1 -
>  1 files changed, 0 insertions(+), 1 deletions(-)
> 
> diff --git a/gcc/config/m68k/linux.h b/gcc/config/m68k/linux.h
> index 113f278..79e9151 100644
> --- a/gcc/config/m68k/linux.h
> +++ b/gcc/config/m68k/linux.h
> @@ -146,7 +146,6 @@ along with GCC; see the file COPYING3.  If not see
>  #undef FUNCTION_PROFILER
>  #define FUNCTION_PROFILER(FILE, LABELNO) \
>  {									\
> -  asm_fprintf (FILE, "\tlea (%LLP%d,%Rpc),%Ra1\n", (LABELNO));		\
>    if (flag_pic)								\
>      fprintf (FILE, "\tbsr.l _mcount@PLTPC\n");				\
>    else									\


-- 
Maxim Kuvyrkov
CodeSourcery
maxim@codesourcery.com
(650) 331-3385 x724

Re: [PATCH] m68k: don't use lea on profiling

[Jeff Law]

On 11/19/09 06:50, Maxim Kuvyrkov wrote:
> Sebastian Andrzej Siewior wrote:
>> this will not work on large binaries:
>>
>> |bigeasy@ayoka:~/cc$ echo 'int main(void) { return 0; }' > file.c
>> |bigeasy@ayoka:~/cc$ m68k-linux-gnu-gcc -o file file.c -static -pg
>> |/tmp/ccw33VYP.o: In function `main':
>> |file.c:(.text+0x6): relocation truncated to fit: R_68K_PC16 against 
>> `.data'
>> |collect2: ld returned 1 exit status
>>
>> This is reported as #36047. It appears that the generated label is 
>> unused.
>
> As I noted in the PR 
> <http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36047> neither GLIBC nor 
> uClibc need the argument so we may simply remove it (thus getting rid 
> of the PC-relative relocation that causes the problem).
>
> Andreas, does this look OK to you?
My areas of concern would be:

   1. Does anyone hook the mcount call and expect the a0/a1 to have 
reasonable values.  For example, the kernel tracing code?

   2. When did glibc change to its new approach?  If it was recent, then 
we probably can't assume the newer implementation.

jeff

Re: [PATCH] m68k: don't use lea on profiling

[Maxim Kuvyrkov]

Jeff Law wrote:
> On 11/19/09 06:50, Maxim Kuvyrkov wrote:
>> Sebastian Andrzej Siewior wrote:
>>> this will not work on large binaries:
>>>
>>> |bigeasy@ayoka:~/cc$ echo 'int main(void) { return 0; }' > file.c
>>> |bigeasy@ayoka:~/cc$ m68k-linux-gnu-gcc -o file file.c -static -pg
>>> |/tmp/ccw33VYP.o: In function `main':
>>> |file.c:(.text+0x6): relocation truncated to fit: R_68K_PC16 against 
>>> `.data'
>>> |collect2: ld returned 1 exit status
>>>
>>> This is reported as #36047. It appears that the generated label is 
>>> unused.
>>
>> As I noted in the PR 
>> <http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36047> neither GLIBC nor 
>> uClibc need the argument so we may simply remove it (thus getting rid 
>> of the PC-relative relocation that causes the problem).
>>
>> Andreas, does this look OK to you?
> My areas of concern would be:
> 
>   1. Does anyone hook the mcount call and expect the a0/a1 to have 
> reasonable values.  For example, the kernel tracing code?

I don't think so.  M68K kernel does not define any target-specific code 
for profiling.

> 
>   2. When did glibc change to its new approach?  If it was recent, then 
> we probably can't assume the newer implementation.

I looked through revision history of glibc-port and didn't find any 
mentions of _mcount implementation in the m68k directories since GLIBC 2.3.

-- 
Maxim Kuvyrkov
CodeSourcery
maxim@codesourcery.com
(650) 331-3385 x724

Re: [PATCH] m68k: don't use lea on profiling

[Andreas Schwab]

Maxim Kuvyrkov <maxim@codesourcery.com> writes:

> I looked through revision history of glibc-port and didn't find any
> mentions of _mcount implementation in the m68k directories since GLIBC
> 2.3.

The libc4 implementation didn't use the profile counter either.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: [PATCH] m68k: don't use lea on profiling

[Sebastian Andrzej Siewior]

* Jeff Law | 2009-11-23 10:30:10 [-0700]:

>>Andreas, does this look OK to you?
>My areas of concern would be:
>
>  1. Does anyone hook the mcount call and expect the a0/a1 to have
>reasonable values.  For example, the kernel tracing code?
I've implemented ftrace support [0] for the nommu variant however it
wasn't merged since they changed the interface a few times and I did not
follow. Back than I had to rip it out in order to compile the kernel
with -pg and it worked with the label.

[0] http://lkml.indiana.edu/hypermail/linux/kernel/0805.1/2054.html

>jeff
>

Sebastian

Re: [PATCH] m68k: don't use lea on profiling

[Andreas Schwab]

Maxim Kuvyrkov <maxim@codesourcery.com> writes:

> Sebastian Andrzej Siewior wrote:
>> this will not work on large binaries:
>>
>> |bigeasy@ayoka:~/cc$ echo 'int main(void) { return 0; }' > file.c
>> |bigeasy@ayoka:~/cc$ m68k-linux-gnu-gcc -o file file.c -static -pg
>> |/tmp/ccw33VYP.o: In function `main':
>> |file.c:(.text+0x6): relocation truncated to fit: R_68K_PC16 against `.data'
>> |collect2: ld returned 1 exit status
>>
>> This is reported as #36047. It appears that the generated label is unused.
>
> As I noted in the PR <http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36047>
> neither GLIBC nor uClibc need the argument so we may simply remove it
> (thus getting rid of the PC-relative relocation that causes the problem).
>
> Andreas, does this look OK to you?

It should also define NO_PROFILE_COUNTERS.  OK with that change and a
proper ChangeLog entry.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

[PATCH v2] m68k: don't use lea on profiling

[Sebastian Andrzej Siewior]

this will not work on large binaries:

|bigeasy@ayoka:~/cc$ echo 'int main(void) { return 0; }' > file.c
|bigeasy@ayoka:~/cc$ m68k-linux-gnu-gcc -o file file.c -static -pg
|/tmp/ccw33VYP.o: In function `main':
|file.c:(.text+0x6): relocation truncated to fit: R_68K_PC16 against `.data'
|collect2: ld returned 1 exit status

This is reported as #36047. It appears that the generated label is unused.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
Andreas, is this what you had in mind?

 gcc/ChangeLog           |    7 +++++++
 gcc/config/m68k/linux.h |    3 +--
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index f172477..f0ad5cf 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,10 @@
+2009-12-06  Sebastian Andrzej Siewior <bigeasy@linutronix.de>
+
+	PR target/36047
+
+	* config/m68k/linux.h: Remove LABELNO from the mcount statement. It is
+	not used by glibc/uclibc and does not work with large binaries.
+
 2009-12-06  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
 
 	* doc/install.texi: Document MPC is required.
diff --git a/gcc/config/m68k/linux.h b/gcc/config/m68k/linux.h
index 113f278..1d73ee0 100644
--- a/gcc/config/m68k/linux.h
+++ b/gcc/config/m68k/linux.h
@@ -142,11 +142,10 @@ along with GCC; see the file COPYING3.  If not see
 
 /* Output assembler code to FILE to increment profiler label # LABELNO
    for profiling a function entry.  */
-
+#define NO_PROFILE_COUNTERS 1
 #undef FUNCTION_PROFILER
 #define FUNCTION_PROFILER(FILE, LABELNO) \
 {									\
-  asm_fprintf (FILE, "\tlea (%LLP%d,%Rpc),%Ra1\n", (LABELNO));		\
   if (flag_pic)								\
     fprintf (FILE, "\tbsr.l _mcount@PLTPC\n");				\
   else									\
-- 
1.6.2.5

Re: [PATCH v2] m68k: don't use lea on profiling

[Andreas Schwab]

Sebastian Andrzej Siewior <sebastian@breakpoint.cc> writes:

> this will not work on large binaries:
>
> |bigeasy@ayoka:~/cc$ echo 'int main(void) { return 0; }' > file.c
> |bigeasy@ayoka:~/cc$ m68k-linux-gnu-gcc -o file file.c -static -pg
> |/tmp/ccw33VYP.o: In function `main':
> |file.c:(.text+0x6): relocation truncated to fit: R_68K_PC16 against `.data'
> |collect2: ld returned 1 exit status
>
> This is reported as #36047. It appears that the generated label is unused.
>
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
> Andreas, is this what you had in mind?

Yes, this is ok.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: [PATCH v2] m68k: don't use lea on profiling

[Maxim Kuvyrkov]

On 12/6/09 11:47 PM, Andreas Schwab wrote:
> Sebastian Andrzej Siewior<sebastian@breakpoint.cc>  writes:
>
>> this will not work on large binaries:
>>
>> |bigeasy@ayoka:~/cc$ echo 'int main(void) { return 0; }'>  file.c
>> |bigeasy@ayoka:~/cc$ m68k-linux-gnu-gcc -o file file.c -static -pg
>> |/tmp/ccw33VYP.o: In function `main':
>> |file.c:(.text+0x6): relocation truncated to fit: R_68K_PC16 against `.data'
>> |collect2: ld returned 1 exit status
>>
>> This is reported as #36047. It appears that the generated label is unused.
>>
>> Signed-off-by: Sebastian Andrzej Siewior<bigeasy@linutronix.de>
>> ---
>> Andreas, is this what you had in mind?
>
> Yes, this is ok.

Sebastian,

Do you plan to check in you patch sometime soon?  If you don't have 
write access to the GCC repository I can commit it on your behalf.

Regards,

-- 
Maxim Kuvyrkov
CodeSourcery
maxim@codesourcery.com
(650) 331-3385 x724

Re: [PATCH v2] m68k: don't use lea on profiling

[Sebastian Andrzej Siewior]

* Maxim Kuvyrkov | 2009-12-11 12:45:34 [+0300]:

>>Yes, this is ok.
>
>Sebastian,
>
>Do you plan to check in you patch sometime soon?  If you don't have
>write access to the GCC repository I can commit it on your behalf.
Yes please do it. I don't have write access. I though Andreas takes care
of it because he was fine it and is the m68k maintainer.

>Regards,

Sebastian

Re: [PATCH v2] m68k: don't use lea on profiling

[Maxim Kuvyrkov]

On 12/11/09 12:53 PM, Sebastian Andrzej Siewior wrote:
> * Maxim Kuvyrkov | 2009-12-11 12:45:34 [+0300]:
>
>>> Yes, this is ok.
>>
>> Sebastian,
>>
>> Do you plan to check in you patch sometime soon?  If you don't have
>> write access to the GCC repository I can commit it on your behalf.
> Yes please do it. I don't have write access. I though Andreas takes care
> of it because he was fine it and is the m68k maintainer.

I checked in your patch and closed the PR.

Regards,

-- 
Maxim Kuvyrkov
CodeSourcery
maxim@codesourcery.com
(650) 331-3385 x724