Fix a few warnings that bug me

Quoted 7 lines · On Thursday 05 January 2012 22:54:49 Simon Glass wrote:

? ? ? len = min_t(int, len, leni);
- ? ? if (data != databuf.ptr)
- ? ? ? ? ? ? memcpy(data, databuf.ptr, len);
+ ? ? if (dataptr)
+ ? ? ? ? ? ? memcpy(data, dataptr, len);
+ ? ? else
+ ? ? ? ? ? ? memcpy(data, &databuf, len);

this doesn't seem to be equivalent ?
-mike

This fixes the following warnings in dlmalloc seen with my gcc 4.6.

dlmalloc.c: In function 'malloc_bin_reloc':
dlmalloc.c:1493: warning: dereferencing pointer 'p' does break strict-aliasing rules dlmalloc.c:1493: warning: dereferencing pointer 'p' does break strict-aliasing rules dlmalloc.c:1490: note: initialized from here
dlmalloc.c:1493: note: initialized from here

Dear Simon Glass,

Quoted 1 line · In message <[email protected]> you wrote:

This fixes the following warnings in dlmalloc seen with my gcc 4.6.

Which exact version is this?

Quoted 5 lines

dlmalloc.c: In function 'malloc_bin_reloc':
dlmalloc.c:1493: warning: dereferencing pointer 'p' does break strict-aliasing rules dlmalloc.c:1493: warning: dereferencing pointer 'p' does break strict-aliasing rules dlmalloc.c:1490: note: initialized from here
dlmalloc.c:1493: note: initialized from here

Can you please tell me at least one board configuration that throws this error (preferrably ppc, arm or mips) ?

Quoted 3 lines

Can you please tell me at least one board configuration that throws this error (preferrably ppc, arm or mips) ?

I can't see one with those architectures. It does happen for others though:

Quoted 3 lines

Also, at least one for the "mpc5xxx: Fix strict-aliasing warnings in usb_ohci.c" and "ppc4xx: Fix strict-aliasing warnings in usb_ohci.c" patches?

Dear Simon,

Quoted 5 lines · In message <CAPnjgZ1OxJJsSEodSM-fqTN791mW-UezqWjFLQBAS-4RbFHJkQ@mail.gmail.com>…

Quoted 3 lines

Can you please tell me at least one board configuration that throws this error (preferrably ppc, arm or mips) ?

I can't see one with those architectures. It does happen for others though:

...

Strange. ?Both MIPS and PPC build without a single warning for me.

I cannot test avr32 or m68k here.

Quoted 3 lines

Quoted 3 lines

Also, at least one for the "mpc5xxx: Fix strict-aliasing warnings in usb_ohci.c" and "ppc4xx: Fix strict-aliasing warnings in usb_ohci.c" patches?

None here.

I am tempted to blame your tool chain for the problems.

Is anybody else seing these warnings? With which tool chain versions?

Best regards,

Wolfgang Denk

--
DENX Software Engineering GmbH, ? ? MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de If you fail to plan, plan to fail.

Quoted 12 lines · Wolfgang Denk

Quoted 1 line · Simon Glass

This fixes the following warnings in dlmalloc seen with my gcc 4.6.

Which exact version is this?

Quoted 6 lines

dlmalloc.c: In function 'malloc_bin_reloc':
dlmalloc.c:1493: warning: dereferencing pointer 'p' does break strict-aliasing rules dlmalloc.c:1493: warning: dereferencing pointer 'p' does break strict-aliasing rules dlmalloc.c:1490: note: initialized from here
dlmalloc.c:1493: note: initialized from here

Quoted 2 lines

Can you please tell me at least one board configuration that throws this error (preferrably ppc, arm or mips) ?

I can't see one with those architectures. It does happen for others though:

This fixes the following warnings in dlmalloc seen with my gcc 4.6.

dlmalloc.c: In function 'malloc_bin_reloc':
dlmalloc.c:1493: warning: dereferencing pointer 'p' does break strict-aliasing rules dlmalloc.c:1493: warning: dereferencing pointer 'p' does break strict-aliasing rules dlmalloc.c:1490: note: initialized from here
dlmalloc.c:1493: note: initialized from here

Signed-off-by Simon Glass <[email protected]>

---
 common/dlmalloc.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)

common/dlmalloc.c +0 -0

index c645d73..0a719b4 100644
@@ -1487,11 +1487,11 @@ static mbinptr av_[NAV * 2 + 2] = {
 #ifdef CONFIG_NEEDS_MANUAL_RELOC
 void malloc_bin_reloc (void)

Dear Simon Glass,

Quoted 7 lines · Simon Glass

This fixes the following warnings in dlmalloc seen with my gcc 4.6.

dlmalloc.c: In function 'malloc_bin_reloc':
dlmalloc.c:1493: warning: dereferencing pointer 'p' does break strict-aliasing rules dlmalloc.c:1493: warning: dereferencing pointer 'p' does break strict-aliasing rules dlmalloc.c:1490: note: initialized from here
dlmalloc.c:1493: note: initialized from here

This is really annoying, but ...

Quoted 15 lines

Signed-off-by Simon Glass <[email protected]>

---
  common/dlmalloc.c |    8 ++++----
  1 files changed, 4 insertions(+), 4 deletions(-)

common/dlmalloc.c +0 -1

index c645d73..0a719b4 100644
@@ -1487,11 +1487,11 @@ static mbinptr av_[NAV * 2 + 2] = {
  #ifdef CONFIG_NEEDS_MANUAL_RELOC
  void malloc_bin_reloc (void)
  {
-	unsigned long *p = (unsigned long *)(&av_[2]);

common/dlmalloc.c static mbinptr av_[NAV * 2 + 2] = { · line 1494

Quoted 8 lines

common/dlmalloc.c +4 -3

+	mbinptr *p = &av_[2];
  	int i;
-	for (i=2; i<(sizeof(av_)/sizeof(mbinptr)); ++i) {
-		*p++ += gd->reloc_off;
-	}
+
+	for (i = 2; i < ARRAY_SIZE(av_); ++i)
+		*p = (mbinptr)((ulong)*p + gd->reloc_off);

your patch breaks avr32 boards at runtime:

This fixes warnings seen with my gcc 4.6.

usb_ohci.c: In function 'submit_control_msg':
usb_ohci.c:1046: warning: dereferencing pointer 'data_buf.70' does break strict-aliasing rules
usb_ohci.c:1046: note: initialized from here usb_ohci.c:1048: warning: dereferencing pointer 'data_buf.70' does break strict-aliasing rules
usb_ohci.c:1048: note: initialized from here usb_ohci.c:1050: warning: dereferencing pointer 'data_buf.70' does break strict-aliasing rules

Dear Simon Glass,

Quoted 11 lines

This fixes warnings seen with my gcc 4.6.

usb_ohci.c: In function 'submit_control_msg':
usb_ohci.c:1046: warning: dereferencing pointer 'data_buf.70' does break strict-aliasing rules
usb_ohci.c:1046: note: initialized from here usb_ohci.c:1048: warning: dereferencing pointer 'data_buf.70' does break strict-aliasing rules
usb_ohci.c:1048: note: initialized from here usb_ohci.c:1050: warning: dereferencing pointer 'data_buf.70' does break strict-aliasing rules

[...]

Are we still seeing these?

Best regards,
Marek Vasut

arch/sandbox/cpu/os.c +2 -0

 
+#include <errno.h>
 #include <fcntl.h>
+#include <linux/types.h>
 #include <stdlib.h>
 #include <termios.h>

Quoted 15 lines · On Thursday 05 January 2012 22:54:53 Simon Glass wrote:

arch/sandbox/cpu/os.c +1 -0

 
+#include <errno.h>

?#include <fcntl.h>
+#include <linux/types.h>
?#include <stdlib.h>
?#include <termios.h>
-#include <unistd.h>
?#include <time.h>

arch/sandbox/cpu/os.c +1 -3

-#include <errno.h>
-#include <sys/types.h>
-#include <sys/stat.h>
+#include <unistd.h>

?#include <sys/mman.h>

arch/sandbox/cpu/os.c line 31

Quoted 3 lines

arch/sandbox/cpu/os.c +2 -1

-#include <linux/types.h>
+#include <sys/stat.h>
+#include <sys/types.h>

the order should generally be:
? ? ? ?- stuff in include/
? ? ? ?- stuff in sys/
? ? ? ?- stuff in linux/
? ? ? ?- stuff in asm/
each sub-region can be sorted, but i don't think we should go against the subdir rule

-mike

Quoted 9 lines · On Thursday 05 January 2012 22:54:54 Simon Glass wrote:

arch/sandbox/cpu/os.c +0 -0

?#include <unistd.h>
?#include <sys/mman.h>
?#include <sys/stat.h>
+#include <sys/time.h>
?#include <sys/types.h>

Quoted 1 line

?#include <os.h>

guess this will need a rebase once 5/9 is rewritten

-mike

This feature is not available on ARM, so it is an error to define it.

Signed-off-by Simon Glass <[email protected]>

---
 include/configs/snapper9260.h |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

include/configs/snapper9260.h +0 -0

index cb3c674..cee65d1 100644

This feature is not available on ARM, so it is an error to define it.

Signed-off-by Simon Glass <[email protected]>

---
 include/configs/snapper9260.h |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

There doesn't seem to be any reason for using uchar here, so change it to char. This fixes a warning:

pointer targets in passing argument 1 of 'sprintf' differ in signedness

Signed-off-by Simon Glass <[email protected]>

---

This seems to be unsigned char for no good reason. Tidy this up and remove the casts.

Signed-off-by Simon Glass <[email protected]>

---

This series fixes some warnings which seem to still be left over from all the warning-squashing efforts for 2011.12. If patches exist for some of these already then please ignore.

The alias problem in USB is a bit odd, since I thought it was already fixed. But perhaps it has come back.

Dear Simon Glass,

Quoted 4 lines · In message <[email protected]> you wrote:

This series fixes some warnings which seem to still be left over from all the warning-squashing efforts for 2011.12. If patches exist for some of these already then please ignore.

Quoted 2 lines

The alias problem in USB is a bit odd, since I thought it was already fixed. But perhaps it has come back.

Can you please let me know which tool chain is throwing these errors? I built "master" yesterday, and before adding the commit that introduced the ``"_LINUX_CONFIG_H" redefined'' issue the PPC tree was built clean - with ELDK 4.2 (gcc 4.2.2), ELDK 5.1 (gcc 4.6.1) and with ELDK 5.2 (gcc 4.6.3).

Best regards,

Wolfgang Denk

--
DENX Software Engineering GmbH, ? ? MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de Documentation is like sex: when it is good, it is ?very, ?very ?good; and when it is bad, it is better than nothing. ? ? ? ? - Dick Brandon

Hi Wolfgang,

Quoted 7 lines · Wolfgang Denk

Dear Simon Glass,

Quoted 4 lines · In message <[email protected]> you wrote:

This series fixes some warnings which seem to still be left over from all the warning-squashing efforts for 2011.12. If patches exist for some of these already then please ignore.

Quoted 8 lines

Quoted 2 lines

The alias problem in USB is a bit odd, since I thought it was already fixed. But perhaps it has come back.

Can you please let me know which tool chain is throwing these errors? I built "master" yesterday, and before adding the commit that introduced the ``"_LINUX_CONFIG_H" redefined'' issue the PPC tree was built clean - with ELDK 4.2 (gcc 4.2.2), ELDK 5.1 (gcc 4.6.1) and with ELDK 5.2 (gcc 4.6.3).

I just check that - actually my PowerPC one is v4.4.4:

Using built-in specs.
Target: powerpc-linux
Configured with: /home/tony/buildall/src/gcc/configure --target=powerpc-linux --enable-targets=all --prefix=/opt/cross/gcc-4.4.4-nolibc/powerpc-linux/ --enable-languages=c --with-newlib --without-headers --enable-sjlj-exceptions --with-system-libunwind --disable-nls --disable-threads --disable-shared --disable-libmudflap --disable-libssp --disable-libgomp --disable-decimal-float --enable-checking=release --with-mpfr=/home/tony/buildall/src/sys --with-gmp=/home/tony/buildall/src/sys
Thread model: single
gcc version 4.4.4 (GCC)

I got this from Mike F (from the Gentoo site).

Regards,
Simon

Quoted 11 lines

Best regards,

Wolfgang Denk

--
DENX Software Engineering GmbH, ? ? MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de Documentation is like sex: when it is good, it is ?very, ?very ?good; and when it is bad, it is better than nothing. ? ? ? ? - Dick Brandon

Hi Wolfgang / Stefano,

Quoted 16 lines

I just check that - actually my PowerPC one is v4.4.4:

Using built-in specs.
Target: powerpc-linux
Configured with: /home/tony/buildall/src/gcc/configure --target=powerpc-linux --enable-targets=all --prefix=/opt/cross/gcc-4.4.4-nolibc/powerpc-linux/ --enable-languages=c --with-newlib --without-headers --enable-sjlj-exceptions --with-system-libunwind --disable-nls --disable-threads --disable-shared --disable-libmudflap --disable-libssp --disable-libgomp --disable-decimal-float --enable-checking=release --with-mpfr=/home/tony/buildall/src/sys --with-gmp=/home/tony/buildall/src/sys
Thread model: single
gcc version 4.4.4 (GCC)

Quoted 2 lines

I got this from Mike F (from the Gentoo site).

What should we do about this series? I still see these warnings with the toolchains I use [1]. Does anyone else?

I will try myself with mips, as this is one of the architectures reporting warnings.

I will try myself with mips, as this is one of the architectures reporting warnings.

Quoted 2 lines · On Tuesday 21 February 2012 05:20:47 Stefano Babic wrote:

I will try myself with mips, as this is one of the architectures reporting warnings.

i don't think Simon did say that mips was throwing these warnings, so testing that arch and looking for these warnings won't help

Quoted 6 lines · Mike Frysinger

Quoted 2 lines · On Tuesday 21 February 2012 05:20:47 Stefano Babic wrote:

I will try myself with mips, as this is one of the architectures reporting warnings.

i don't think Simon did say that mips was throwing these warnings, so testing that arch and looking for these warnings won't help

In the thread I read that Simon got warnings with both mips and powerpc, as well as with avr32 and m68k. But mips and powerpc are free of warnings when I build with gcc 4.6.

Stefano

--
=====================================================================
DENX Software Engineering GmbH, ? ? MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 ?Email: office at denx.de
=====================================================================

68k only:

 arch/m68k/lib/board.c               |    2 +-

- I believe this fix is sensible for the reasons described in the patch, regardless of what the toolchain does.

 arch/powerpc/lib/board.c            |    7 +--

- Similar to the68k patch, again I think it makes sense to apply this. The variable is unsigned char for no good reason.

Sandbox:

 arch/sandbox/cpu/os.c               |   11 +++--

- This should be applied as it is not toolchain dependent (Mike may have picked it up already)

 common/dlmalloc.c                   |    8 ++--

- These happen with MIPS, AVR, 68k for me (with Mike's toolchains).

Yes but not just those. There are detailed notes with each patch, but here is a summary. I think 6 out of 9 patches should be applied regardless of toolchain.

68k only:

 arch/m68k/lib/board.c               |    2 +-

- I believe this fix is sensible for the reasons described in the patch, regardless of what the toolchain does.

 arch/powerpc/lib/board.c            |    7 +--

- Similar to the68k patch, again I think it makes sense to apply this. The variable is unsigned char for no good reason.

Sandbox:

 arch/sandbox/cpu/os.c               |   11 +++--

- This should be applied as it is not toolchain dependent (Mike may have picked it up already)

ARM and PPC:

 drivers/usb/host/ohci-hcd.c         |   26 ++++++------

- These happen with my 4.4.4 compiler, apparently not on 4.6. Drop them if you like.

ARM:

 include/configs/snapper9260.h       |    1 -

- This fixes a real problem regardless of toolchain.

Yes but not just those. There are detailed notes with each patch, but here is a summary. I think 6 out of 9 patches should be applied regardless of toolchain.

PPC only:
arch/powerpc/cpu/mpc5xxx/usb_ohci.c | 79 ++++++++++++++++++++-----------=
---
arch/powerpc/cpu/ppc4xx/usb_ohci.c | 81 ++++++++++++++++++++-----------=
---
- These two seem not to happen with a later toolchain. Please skip these if you like.

 common/dlmalloc.c                   |    8 ++--

- These happen with MIPS, AVR, 68k for me (with Mike's toolchains). Does not happen on 4.6 ARM.

ARM and PPC:

 drivers/usb/host/ohci-hcd.c         |   26 ++++++------

- These happen with my 4.4.4 compiler, apparently not on 4.6. Drop them if you like.

This series fixes some warnings which seem to still be left over from all the warning-squashing efforts for 2011.12. If patches exist for some of these already then please ignore.

The alias problem in USB is a bit odd, since I thought it was already fixed. But perhaps it has come back.

Simon Glass (9):
usb: Fix strict-aliasing warning in host/ohci-hcd.c Fix strict-aliasing warning in dlmalloc
mpc5xxx: Fix strict-aliasing warnings in usb_ohci.c
ppc4xx: Fix strict-aliasing warnings in usb_ohci.c
sandbox: sort header files in os.c
sandbox: Add required header to os.c
Remove CONFIG_SYS_EXTBDINFO from snapper9260.h
m68k: Change memsz to a signed char to avoid warning
ppc: Change memsz variable to signed char

 arch/m68k/lib/board.c               |    2 +-
 arch/powerpc/cpu/mpc5xxx/usb_ohci.c |   79
++++++++++++++++++++-------------- arch/powerpc/cpu/ppc4xx/usb_ohci.c  |  
81 ++++++++++++++++++++-------------- arch/powerpc/lib/board.c           
|    7 +--
 arch/sandbox/cpu/os.c               |   11 +++--
 common/dlmalloc.c                   |    8 ++--
 drivers/usb/host/ohci-hcd.c         |   26 ++++++------
 include/configs/snapper9260.h       |    1 -
 8 files changed, 122 insertions(+), 93 deletions(-)

Dear Simon Glass,

Quoted 24 lines

This series fixes some warnings which seem to still be left over from all the warning-squashing efforts for 2011.12. If patches exist for some of these already then please ignore.

The alias problem in USB is a bit odd, since I thought it was already fixed. But perhaps it has come back.

Simon Glass (9):
? usb: Fix strict-aliasing warning in host/ohci-hcd.c ? Fix strict-aliasing warning in dlmalloc
? mpc5xxx: Fix strict-aliasing warnings in usb_ohci.c ? ppc4xx: Fix strict-aliasing warnings in usb_ohci.c ? sandbox: sort header files in os.c
? sandbox: Add required header to os.c
? Remove CONFIG_SYS_EXTBDINFO from snapper9260.h ? m68k: Change memsz to a signed char to avoid warning ? ppc: Change memsz variable to signed char

?arch/m68k/lib/board.c ? ? ? ? ? ? ? | ? ?2 +- ?arch/powerpc/cpu/mpc5xxx/usb_ohci.c | ? 79 ++++++++++++++++++++-------------- arch/powerpc/cpu/ppc4xx/usb_ohci.c ?| 81 ++++++++++++++++++++-------------- arch/powerpc/lib/board.c | ? ?7 +--

Quoted 5 lines

?arch/sandbox/cpu/os.c ? ? ? ? ? ? ? | ? 11 +++--
?common/dlmalloc.c ? ? ? ? ? ? ? ? ? | ? ?8 ++--
?drivers/usb/host/ohci-hcd.c ? ? ? ? | ? 26 ++++++------
?include/configs/snapper9260.h ? ? ? | ? ?1 -

?8 files changed, 122 insertions(+), 93 deletions(-)

Can you please split out the USB part and send it to me (I became new usb maint.) please?

Thanks in advance!

Best regards,
Marek Vasut

Quoted 2 lines

Quoted 2 lines

?arch/powerpc/cpu/mpc5xxx/usb_ohci.c | ? 79
++++++++++++++++++++--------------

arch/powerpc/cpu/ppc4xx/usb_ohci.c ?|

Quoted 1 line

Quoted 1 line

81 ++++++++++++++++++++-------------- arch/powerpc/lib/board.c

The two PowerPC ones apparently only happen with the 4.4 toolchain, so I think we are going to drop those and I am going to upgrade my toolchain instead :-)

This has just grabbed my attention with the comment regarding the 4.4 toolchain as I know that MontaVista ship a version of the 4.4.0 toolchain for PowerPC with one of their embedded Linux products. Is there a generic issue with 4.4.x versions or just a rogue board or two?

Hi Marek,

Quoted 34 lines · Marek Vasut

Dear Simon Glass,

Quoted 24 lines

This series fixes some warnings which seem to still be left over from all the warning-squashing efforts for 2011.12. If patches exist for some of these already then please ignore.

The alias problem in USB is a bit odd, since I thought it was already fixed. But perhaps it has come back.

Simon Glass (9):
usb: Fix strict-aliasing warning in host/ohci-hcd.c Fix strict-aliasing warning in dlmalloc
mpc5xxx: Fix strict-aliasing warnings in usb_ohci.c
ppc4xx: Fix strict-aliasing warnings in usb_ohci.c
sandbox: sort header files in os.c
sandbox: Add required header to os.c
Remove CONFIG_SYS_EXTBDINFO from snapper9260.h
m68k: Change memsz to a signed char to avoid warning
ppc: Change memsz variable to signed char

 arch/m68k/lib/board.c               |    2 +-
 arch/powerpc/cpu/mpc5xxx/usb_ohci.c |   79

++++++++++++++++++++-------------- arch/powerpc/cpu/ppc4xx/usb_ohci.c | 81 ++++++++++++++++++++-------------- arch/powerpc/lib/board.c

Quoted 7 lines

| 7 +--

 arch/sandbox/cpu/os.c               |   11 +++--
 common/dlmalloc.c                   |    8 ++--
 drivers/usb/host/ohci-hcd.c         |   26 ++++++------
 include/configs/snapper9260.h       |    1 -
 8 files changed, 122 insertions(+), 93 deletions(-)

Quoted 4 lines

Can you please split out the USB part and send it to me (I became new usb maint.) please?

Thanks in advance!

Of course - I think you want to look at this one:

http://patchwork.ozlabs.org/patch/134600/

The two PowerPC ones apparently only happen with the 4.4 toolchain, so I think we are going to drop those and I am going to upgrade my toolchain instead :-)