|
From: Julian S. <js...@ac...> - 2010-10-15 10:06:57
|
I just created the stable branch for the 3.6.x line. Please try it out on platforms and configurations that are important to you. Unless there are reports of serious breakage, I will release 3.6.0 final on Monday. To check out the branch, do svn co svn://svn.valgrind.org/valgrind/branches/VALGRIND_3_6_BRANCH branch36 or alternatively a test tarball of it is available at http://www.valgrind.org/downloads/valgrind-3.6.0-test2.tar.bz2 It should be buildable and usable on the following targets: {x86,amd64,ppc32,ppc64,arm}-linux, {x86,amd64}-darwin. J |
|
From: Kirill B. <bat...@is...> - 2010-10-15 13:48:38
Attachments:
vfp.stdout.diff
|
Hello, Julian.
I tried to build a 3.6 branch from SVN and run ARM-specific tests from none
testsuite on ARMv7 target. Looks like vfp.stdout.exp file is not correct.
VNMLA/VMLS/VNMLS produce the same result with and without Valgrind on my
hardware but it differs from the one specified in .exp file. I checked
manually some of the tests and it turns out that .exp file is not
correct. For example
VMLS <some number>, -Inf, NaN should produce -NaN as a result, not +NaN.
VMLS <some very little number>, 456.2489562, -7.2945676 should produce
3328.138916, not -3328.138916
----
Kirill
On Fri, 15 Oct 2010, Julian Seward wrote:
>
> I just created the stable branch for the 3.6.x line. Please try it
> out on platforms and configurations that are important to you.
> Unless there are reports of serious breakage, I will release 3.6.0
> final on Monday.
>
> To check out the branch, do
>
> svn co svn://svn.valgrind.org/valgrind/branches/VALGRIND_3_6_BRANCH branch36
>
> or alternatively a test tarball of it is available at
>
> http://www.valgrind.org/downloads/valgrind-3.6.0-test2.tar.bz2
>
> It should be buildable and usable on the following targets:
>
> {x86,amd64,ppc32,ppc64,arm}-linux, {x86,amd64}-darwin.
>
> J
>
> ------------------------------------------------------------------------------
> Download new Adobe(R) Flash(R) Builder(TM) 4
> The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly
> Flex(R) Builder(TM)) enable the development of rich applications that run
> across multiple browsers and platforms. Download your free trials today!
> http://p.sf.net/sfu/adobe-dev2dev
> _______________________________________________
> Valgrind-developers mailing list
> Val...@li...
> https://lists.sourceforge.net/lists/listinfo/valgrind-developers
> |
|
From: Julian S. <js...@ac...> - 2010-10-15 14:18:46
|
Hmm, I wonder if this is a C library issue. For me, on Ubuntu 10.04 it works just fine: $ perl tests/vg_regtest none/tests/arm -- Running tests in none/tests/arm ------------------------------------ neon128: valgrind -q ./neon128 neon64: valgrind -q ./neon64 v6intARM: valgrind -q ./v6intARM v6intThumb: valgrind -q ./v6intThumb v6media: valgrind -q ./v6media vfp: valgrind -q ./vfp -- Finished tests in none/tests/arm ------------------------------------ == 6 tests, 0 stderr failures, 0 stdout failures, 0 post failures == If the programs produce the same result with and without Valgrind then Valgrind is not a problem, but rather it is with libc printing differences. What libc and distro are you using? J On Friday, October 15, 2010, Kirill Batuzov wrote: > Hello, Julian. > > I tried to build a 3.6 branch from SVN and run ARM-specific tests from none > testsuite on ARMv7 target. Looks like vfp.stdout.exp file is not correct. > VNMLA/VMLS/VNMLS produce the same result with and without Valgrind on my > hardware but it differs from the one specified in .exp file. I checked > manually some of the tests and it turns out that .exp file is not > correct. For example > VMLS <some number>, -Inf, NaN should produce -NaN as a result, not +NaN. > VMLS <some very little number>, 456.2489562, -7.2945676 should produce > 3328.138916, not -3328.138916 > > ---- > Kirill > > On Fri, 15 Oct 2010, Julian Seward wrote: > > I just created the stable branch for the 3.6.x line. Please try it > > out on platforms and configurations that are important to you. > > Unless there are reports of serious breakage, I will release 3.6.0 > > final on Monday. > > > > To check out the branch, do > > > > svn co svn://svn.valgrind.org/valgrind/branches/VALGRIND_3_6_BRANCH > > branch36 > > > > or alternatively a test tarball of it is available at > > > > http://www.valgrind.org/downloads/valgrind-3.6.0-test2.tar.bz2 > > > > It should be buildable and usable on the following targets: > > {x86,amd64,ppc32,ppc64,arm}-linux, {x86,amd64}-darwin. > > > > J > > > > ------------------------------------------------------------------------- > > ----- Download new Adobe(R) Flash(R) Builder(TM) 4 > > The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly > > Flex(R) Builder(TM)) enable the development of rich applications that run > > across multiple browsers and platforms. Download your free trials today! > > http://p.sf.net/sfu/adobe-dev2dev > > _______________________________________________ > > Valgrind-developers mailing list > > Val...@li... > > https://lists.sourceforge.net/lists/listinfo/valgrind-developers |
|
From: Dmitry Z. <zh...@is...> - 2010-10-15 18:59:56
|
On 10/15/2010 06:15 PM, Julian Seward wrote:
> Hmm, I wonder if this is a C library issue. For me, on Ubuntu 10.04
> it works just fine:
>
> $ perl tests/vg_regtest none/tests/arm
> -- Running tests in none/tests/arm ------------------------------------
> neon128: valgrind -q ./neon128
> neon64: valgrind -q ./neon64
> v6intARM: valgrind -q ./v6intARM
> v6intThumb: valgrind -q ./v6intThumb
> v6media: valgrind -q ./v6media
> vfp: valgrind -q ./vfp
> -- Finished tests in none/tests/arm ------------------------------------
>
> == 6 tests, 0 stderr failures, 0 stdout failures, 0 post failures ==
>
> If the programs produce the same result with and without Valgrind
> then Valgrind is not a problem, but rather it is with libc printing
> differences. What libc and distro are you using?
>
We are using a soon-to-be LiMo distro with 2.8 Glibc. Personally I
doubt that it is connected with a different printing - the output values
are taken from the memory and are different sometimes in just several
bits. This is probably to different compilation or assembly generation
(I just checked and I am getting the same output for most VMLS
instruction tests as you if I substitute VMLS with VMLA; and I am
puzzled by this) - can you please send us your compiled test?
Regards,
Dmitry
> J
>
> On Friday, October 15, 2010, Kirill Batuzov wrote:
>
>> Hello, Julian.
>>
>> I tried to build a 3.6 branch from SVN and run ARM-specific tests from none
>> testsuite on ARMv7 target. Looks like vfp.stdout.exp file is not correct.
>> VNMLA/VMLS/VNMLS produce the same result with and without Valgrind on my
>> hardware but it differs from the one specified in .exp file. I checked
>> manually some of the tests and it turns out that .exp file is not
>> correct. For example
>> VMLS<some number>, -Inf, NaN should produce -NaN as a result, not +NaN.
>> VMLS<some very little number>, 456.2489562, -7.2945676 should produce
>> 3328.138916, not -3328.138916
>>
>> ----
>> Kirill
>>
>> On Fri, 15 Oct 2010, Julian Seward wrote:
>>
>>> I just created the stable branch for the 3.6.x line. Please try it
>>> out on platforms and configurations that are important to you.
>>> Unless there are reports of serious breakage, I will release 3.6.0
>>> final on Monday.
>>>
>>> To check out the branch, do
>>>
>>> svn co svn://svn.valgrind.org/valgrind/branches/VALGRIND_3_6_BRANCH
>>> branch36
>>>
>>> or alternatively a test tarball of it is available at
>>>
>>> http://www.valgrind.org/downloads/valgrind-3.6.0-test2.tar.bz2
>>>
>>> It should be buildable and usable on the following targets:
>>> {x86,amd64,ppc32,ppc64,arm}-linux, {x86,amd64}-darwin.
>>>
>>> J
>>>
>>> -------------------------------------------------------------------------
>>> ----- Download new Adobe(R) Flash(R) Builder(TM) 4
>>> The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly
>>> Flex(R) Builder(TM)) enable the development of rich applications that run
>>> across multiple browsers and platforms. Download your free trials today!
>>> http://p.sf.net/sfu/adobe-dev2dev
>>> _______________________________________________
>>> Valgrind-developers mailing list
>>> Val...@li...
>>> https://lists.sourceforge.net/lists/listinfo/valgrind-developers
>>>
>
> ------------------------------------------------------------------------------
> Download new Adobe(R) Flash(R) Builder(TM) 4
> The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly
> Flex(R) Builder(TM)) enable the development of rich applications that run
> across multiple browsers and platforms. Download your free trials today!
> http://p.sf.net/sfu/adobe-dev2dev
> _______________________________________________
> Valgrind-developers mailing list
> Val...@li...
> https://lists.sourceforge.net/lists/listinfo/valgrind-developers
>
>
|
|
From: Maynard J. <may...@us...> - 2010-10-15 15:15:28
|
Julian Seward wrote: > > I just created the stable branch for the 3.6.x line. Please try it > out on platforms and configurations that are important to you. > Unless there are reports of serious breakage, I will release 3.6.0 > final on Monday. > > To check out the branch, do > > svn co svn://svn.valgrind.org/valgrind/branches/VALGRIND_3_6_BRANCH branch36 > > or alternatively a test tarball of it is available at > > http://www.valgrind.org/downloads/valgrind-3.6.0-test2.tar.bz2 > > It should be buildable and usable on the following targets: > > {x86,amd64,ppc32,ppc64,arm}-linux, {x86,amd64}-darwin. I tested this branch on POWER6 SLES 10 SP2, SLES 11 SP1, and RHEL 6 beta. Test results all seem pretty reasonable, aside from some known issues and a few (apparently) distro-specific issues that I'll need to investigate -- probably not worth holding up a release for. There is the problem with the ppc64 VMX test failure (bug 250038) that I really would like to fix before the release. But I doubt I'll have a breakthrough on debugging this before Monday. -Maynard > > J > > ------------------------------------------------------------------------------ > Download new Adobe(R) Flash(R) Builder(TM) 4 > The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly > Flex(R) Builder(TM)) enable the development of rich applications that run > across multiple browsers and platforms. Download your free trials today! > http://p.sf.net/sfu/adobe-dev2dev > _______________________________________________ > Valgrind-users mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-users |
|
From: Vince W. <vi...@cs...> - 2010-10-15 19:17:30
|
On Fri, 15 Oct 2010, Julian Seward wrote:
>
> It should be buildable and usable on the following targets:
>
> {x86,amd64,ppc32,ppc64,arm}-linux, {x86,amd64}-darwin.
compiling on an armv7l system I get the following compile failure:
This is a Beagle Board running Debian Sid, gcc 4.4.5, binutils 2.20.1
make[3]: Entering directory `/arm/vince/valgrind/coregrind'
gcc -DHAVE_CONFIG_H -I. -I.. -I.. -I../include -I../VEX/pub -DVGA_arm=1 -DVGO_linux=1 -DVGP_arm_linux=1 -I../coregrind -DVG_LIBDIR="\"/usr/local/lib/valgrind"\" -DVG_PLATFORM="\"arm-linux\"" -I.. -I../include -I../VEX/pub -DVGA_arm=1 -DVGO_linux=1 -DVGP_arm_linux=1 -I../coregrind -DVG_LIBDIR="\"/usr/local/lib/valgrind"\" -DVG_PLATFORM="\"arm-linux\"" -g -Wno-long-long -MT libcoregrind_arm_linux_a-dispatch-arm-linux.o -MD -MP -MF .deps/libcoregrind_arm_linux_a-dispatch-arm-linux.Tpo -c -o libcoregrind_arm_linux_a-dispatch-arm-linux.o `test -f 'm_dispatch/dispatch-arm-linux.S' || echo './'`m_dispatch/dispatch-arm-linux.S
m_dispatch/dispatch-arm-linux.S: Assembler messages:
m_dispatch/dispatch-arm-linux.S:66: Error: internal_relocation (type: OFFSET_IMM) not fixed up
m_dispatch/dispatch-arm-linux.S:90: Error: internal_relocation (type: OFFSET_IMM) not fixed up
m_dispatch/dispatch-arm-linux.S:135: Error: internal_relocation (type: OFFSET_IMM) not fixed up
m_dispatch/dispatch-arm-linux.S:184: Error: internal_relocation (type: OFFSET_IMM) not fixed up
make[3]: *** [libcoregrind_arm_linux_a-dispatch-arm-linux.o] Error 1
make[3]: Leaving directory `/arm/vince/valgrind/coregrind'
make[2]: *** [all] Error 2
make[2]: Leaving directory `/arm/vince/valgrind/coregrind'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/arm/vince/valgrind'
make: *** [all] Error 2
|
|
From: Julian S. <js...@ac...> - 2010-10-15 23:28:18
|
I notice that the compile command doesn't have -marm in it, and
it should. What happens if you add -marm to AM_CCASFLAGS_ARM_LINUX
at Makefile.all.am:157, do make distclean, and rebuild?
J
On Friday, October 15, 2010, Vince Weaver wrote:
> On Fri, 15 Oct 2010, Julian Seward wrote:
> > It should be buildable and usable on the following targets:
> > {x86,amd64,ppc32,ppc64,arm}-linux, {x86,amd64}-darwin.
>
> compiling on an armv7l system I get the following compile failure:
>
> This is a Beagle Board running Debian Sid, gcc 4.4.5, binutils 2.20.1
>
> make[3]: Entering directory `/arm/vince/valgrind/coregrind'
> gcc -DHAVE_CONFIG_H -I. -I.. -I.. -I../include -I../VEX/pub -DVGA_arm=1
> -DVGO_linux=1 -DVGP_arm_linux=1 -I../coregrind
> -DVG_LIBDIR="\"/usr/local/lib/valgrind"\" -DVG_PLATFORM="\"arm-linux\""
> -I.. -I../include -I../VEX/pub -DVGA_arm=1 -DVGO_linux=1 -DVGP_arm_linux=1
> -I../coregrind -DVG_LIBDIR="\"/usr/local/lib/valgrind"\"
> -DVG_PLATFORM="\"arm-linux\"" -g -Wno-long-long -MT
> libcoregrind_arm_linux_a-dispatch-arm-linux.o -MD -MP -MF
> .deps/libcoregrind_arm_linux_a-dispatch-arm-linux.Tpo -c -o
> libcoregrind_arm_linux_a-dispatch-arm-linux.o `test -f
> 'm_dispatch/dispatch-arm-linux.S' || echo
> './'`m_dispatch/dispatch-arm-linux.S m_dispatch/dispatch-arm-linux.S:
> Assembler messages:
> m_dispatch/dispatch-arm-linux.S:66: Error: internal_relocation (type:
> OFFSET_IMM) not fixed up m_dispatch/dispatch-arm-linux.S:90: Error:
> internal_relocation (type: OFFSET_IMM) not fixed up
> m_dispatch/dispatch-arm-linux.S:135: Error: internal_relocation (type:
> OFFSET_IMM) not fixed up m_dispatch/dispatch-arm-linux.S:184: Error:
> internal_relocation (type: OFFSET_IMM) not fixed up make[3]: ***
> [libcoregrind_arm_linux_a-dispatch-arm-linux.o] Error 1 make[3]: Leaving
> directory `/arm/vince/valgrind/coregrind'
> make[2]: *** [all] Error 2
> make[2]: Leaving directory `/arm/vince/valgrind/coregrind'
> make[1]: *** [all-recursive] Error 1
> make[1]: Leaving directory `/arm/vince/valgrind'
> make: *** [all] Error 2
|
|
From: Vince W. <vi...@cs...> - 2010-10-18 15:34:04
|
On Sat, 16 Oct 2010, Julian Seward wrote:
>
> I notice that the compile command doesn't have -marm in it, and
> it should. What happens if you add -marm to AM_CCASFLAGS_ARM_LINUX
> at Makefile.all.am:157, do make distclean, and rebuild?
yes, adding -marm to AM_CCASFLAGS_ARM_LINUX fixes things, and
Valgrind runs properly after the build finishes.
(sorry for the delay in responding, a power outage at work meant I
couldn't access the machine to test this until this morning).
Vince
> On Friday, October 15, 2010, Vince Weaver wrote:
> > On Fri, 15 Oct 2010, Julian Seward wrote:
> > > It should be buildable and usable on the following targets:
> > > {x86,amd64,ppc32,ppc64,arm}-linux, {x86,amd64}-darwin.
> >
> > compiling on an armv7l system I get the following compile failure:
> >
> > This is a Beagle Board running Debian Sid, gcc 4.4.5, binutils 2.20.1
> >
> > make[3]: Entering directory `/arm/vince/valgrind/coregrind'
> > gcc -DHAVE_CONFIG_H -I. -I.. -I.. -I../include -I../VEX/pub -DVGA_arm=1
> > -DVGO_linux=1 -DVGP_arm_linux=1 -I../coregrind
> > -DVG_LIBDIR="\"/usr/local/lib/valgrind"\" -DVG_PLATFORM="\"arm-linux\""
> > -I.. -I../include -I../VEX/pub -DVGA_arm=1 -DVGO_linux=1 -DVGP_arm_linux=1
> > -I../coregrind -DVG_LIBDIR="\"/usr/local/lib/valgrind"\"
> > -DVG_PLATFORM="\"arm-linux\"" -g -Wno-long-long -MT
> > libcoregrind_arm_linux_a-dispatch-arm-linux.o -MD -MP -MF
> > .deps/libcoregrind_arm_linux_a-dispatch-arm-linux.Tpo -c -o
> > libcoregrind_arm_linux_a-dispatch-arm-linux.o `test -f
> > 'm_dispatch/dispatch-arm-linux.S' || echo
> > './'`m_dispatch/dispatch-arm-linux.S m_dispatch/dispatch-arm-linux.S:
> > Assembler messages:
> > m_dispatch/dispatch-arm-linux.S:66: Error: internal_relocation (type:
> > OFFSET_IMM) not fixed up m_dispatch/dispatch-arm-linux.S:90: Error:
> > internal_relocation (type: OFFSET_IMM) not fixed up
> > m_dispatch/dispatch-arm-linux.S:135: Error: internal_relocation (type:
> > OFFSET_IMM) not fixed up m_dispatch/dispatch-arm-linux.S:184: Error:
> > internal_relocation (type: OFFSET_IMM) not fixed up make[3]: ***
> > [libcoregrind_arm_linux_a-dispatch-arm-linux.o] Error 1 make[3]: Leaving
> > directory `/arm/vince/valgrind/coregrind'
> > make[2]: *** [all] Error 2
> > make[2]: Leaving directory `/arm/vince/valgrind/coregrind'
> > make[1]: *** [all-recursive] Error 1
> > make[1]: Leaving directory `/arm/vince/valgrind'
> > make: *** [all] Error 2
>
>
--
/* Vince Weaver vi...@cs... http://csl.cornell.edu/~vince */
main(){char O,o[66]="|\n\\/_ ",*I=o+7,l[]="B!FhhBHCWE9C?cJFKET$+h'Iq*chT"
,i=0,_;while(_=l[i++])for(O=0;O++<_>>5;)*I=*(I++-(_&31));*I=0;puts(o+5);}
|
|
From: Julian S. <js...@ac...> - 2010-10-18 18:12:47
|
Thanks for the confirmation. Committed as r11457.
J
On Monday, October 18, 2010, Vince Weaver wrote:
> On Sat, 16 Oct 2010, Julian Seward wrote:
> > I notice that the compile command doesn't have -marm in it, and
> > it should. What happens if you add -marm to AM_CCASFLAGS_ARM_LINUX
> > at Makefile.all.am:157, do make distclean, and rebuild?
>
> yes, adding -marm to AM_CCASFLAGS_ARM_LINUX fixes things, and
> Valgrind runs properly after the build finishes.
>
> (sorry for the delay in responding, a power outage at work meant I
> couldn't access the machine to test this until this morning).
>
>
> Vince
>
> > On Friday, October 15, 2010, Vince Weaver wrote:
> > > On Fri, 15 Oct 2010, Julian Seward wrote:
> > > > It should be buildable and usable on the following targets:
> > > > {x86,amd64,ppc32,ppc64,arm}-linux, {x86,amd64}-darwin.
> > >
> > > compiling on an armv7l system I get the following compile failure:
> > >
> > > This is a Beagle Board running Debian Sid, gcc 4.4.5, binutils 2.20.1
> > >
> > > make[3]: Entering directory `/arm/vince/valgrind/coregrind'
> > > gcc -DHAVE_CONFIG_H -I. -I.. -I.. -I../include -I../VEX/pub
> > > -DVGA_arm=1 -DVGO_linux=1 -DVGP_arm_linux=1 -I../coregrind
> > > -DVG_LIBDIR="\"/usr/local/lib/valgrind"\" -DVG_PLATFORM="\"arm-linux\""
> > > -I.. -I../include -I../VEX/pub -DVGA_arm=1 -DVGO_linux=1
> > > -DVGP_arm_linux=1 -I../coregrind
> > > -DVG_LIBDIR="\"/usr/local/lib/valgrind"\"
> > > -DVG_PLATFORM="\"arm-linux\"" -g -Wno-long-long -MT
> > > libcoregrind_arm_linux_a-dispatch-arm-linux.o -MD -MP -MF
> > > .deps/libcoregrind_arm_linux_a-dispatch-arm-linux.Tpo -c -o
> > > libcoregrind_arm_linux_a-dispatch-arm-linux.o `test -f
> > > 'm_dispatch/dispatch-arm-linux.S' || echo
> > > './'`m_dispatch/dispatch-arm-linux.S m_dispatch/dispatch-arm-linux.S:
> > > Assembler messages:
> > > m_dispatch/dispatch-arm-linux.S:66: Error: internal_relocation (type:
> > > OFFSET_IMM) not fixed up m_dispatch/dispatch-arm-linux.S:90: Error:
> > > internal_relocation (type: OFFSET_IMM) not fixed up
> > > m_dispatch/dispatch-arm-linux.S:135: Error: internal_relocation (type:
> > > OFFSET_IMM) not fixed up m_dispatch/dispatch-arm-linux.S:184: Error:
> > > internal_relocation (type: OFFSET_IMM) not fixed up make[3]: ***
> > > [libcoregrind_arm_linux_a-dispatch-arm-linux.o] Error 1 make[3]:
> > > Leaving directory `/arm/vince/valgrind/coregrind'
> > > make[2]: *** [all] Error 2
> > > make[2]: Leaving directory `/arm/vince/valgrind/coregrind'
> > > make[1]: *** [all-recursive] Error 1
> > > make[1]: Leaving directory `/arm/vince/valgrind'
> > > make: *** [all] Error 2
|
|
From: Rich C. <Ric...@me...> - 2010-10-18 16:47:25
|
I ran on openSUSE 11.3 and this the summary results of 'make regtest'. == 562 tests, 74 stderr failures, 54 stdout failures, 3 post failures == Rich On Fri, 15 Oct 2010 12:03:44 +0200 Julian Seward <js...@ac...> wrote: > > I just created the stable branch for the 3.6.x line. Please try it > out on platforms and configurations that are important to you. > Unless there are reports of serious breakage, I will release 3.6.0 > final on Monday. > > To check out the branch, do > > svn co svn://svn.valgrind.org/valgrind/branches/VALGRIND_3_6_BRANCH branch36 > > or alternatively a test tarball of it is available at > > http://www.valgrind.org/downloads/valgrind-3.6.0-test2.tar.bz2 > > It should be buildable and usable on the following targets: > > {x86,amd64,ppc32,ppc64,arm}-linux, {x86,amd64}-darwin. > > J > > ------------------------------------------------------------------------------ > Download new Adobe(R) Flash(R) Builder(TM) 4 > The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly > Flex(R) Builder(TM)) enable the development of rich applications that run > across multiple browsers and platforms. Download your free trials today! > http://p.sf.net/sfu/adobe-dev2dev > _______________________________________________ > Valgrind-developers mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-developers -- |