[valgrind] [Bug 417238] Test memcheck/tests/vbit-test fails on mips64 BE
https://bugs.kde.org/show_bug.cgi?id=417238 --- Comment #3 from Aleksandar Rikalo --- In fact, in most cases of the Ity_I1 operations, the u32 field is used (e.g., vbits.c: 51). The u1 field is used only in few places (removed by the patch). That isn't a problem on LE since unit8_t overlaps with LSB of uint32_t, but it is a problem on BE. Ity_I8, ..., Ity_I64 operations don't require any overlaps. The other possible way to fix the test is to keep u1 field and replace all usages of u32 for Ity_I1 purpose into u1. It seems like it would be more complicated, especially because valgrind_[set|get]_vbits() functions (see valgrind.c) have to be modified. I'm looking at vbits.c and I'm not sure how much logic would be simplified by eliminating the u8, ..., u32 fields (e.g., concat_vbits() function). I would put that option aside. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 415516] Can't cross compile on openwrt
https://bugs.kde.org/show_bug.cgi?id=415516 Aleksandar Rikalo changed: What|Removed |Added CC||aleksandar.rik...@rt-rk.com --- Comment #3 from Aleksandar Rikalo --- Created attachment 126374 --> https://bugs.kde.org/attachment.cgi?id=126374=edit Fix OpenWRT build [WIP] Hi, this patch fixes the "msa" issue with "mipsel-openwrt-linux-uclibc-gcc.bin (OpenWrt/Linaro GCC 4.8-2014.04 r47072) 4.8.3" Please let me know is it helpful. However, there is another problem during linkage of vgdb: /work/toolchain/iDetect-839d28effb806ce37c6c85b0da3de22bd6255450/OpenWrt-Toolchain-ramips-mt7688_gcc-4.8-linaro_uClibc-0.9.33.2.Linux-x86_64/toolchain-mipsel_24kec+dsp_gcc-4.8-linaro_uClibc-0.9.33.2/bin/../lib/gcc/mipsel-openwrt-linux-uclibc/4.8.3/../../../../mipsel-openwrt-linux-uclibc/lib/libpthread.so.0: undefined reference to `dlopen' /work/toolchain/iDetect-839d28effb806ce37c6c85b0da3de22bd6255450/OpenWrt-Toolchain-ramips-mt7688_gcc-4.8-linaro_uClibc-0.9.33.2.Linux-x86_64/toolchain-mipsel_24kec+dsp_gcc-4.8-linaro_uClibc-0.9.33.2/bin/../lib/gcc/mipsel-openwrt-linux-uclibc/4.8.3/../../../../mipsel-openwrt-linux-uclibc/lib/libpthread.so.0: undefined reference to `dlsym' collect2: error: ld returned 1 exit status Makefile:2883: recipe for target 'vgdb' failed It seems that dlopen/dlsym functions are missing in provided version of uClibc. Have you ever encountered this situation (using this particular version of toolchain) ? Best wishes, Aleksandar -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 417238] Test memcheck/tests/vbit-test fails on mips64 BE
https://bugs.kde.org/show_bug.cgi?id=417238 Aleksandar Rikalo changed: What|Removed |Added CC||aleksandar.rik...@rt-rk.com ||, m...@klomp.org -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400162] Patch: Guard against __GLIBC_PREREQ for musl libc
https://bugs.kde.org/show_bug.cgi?id=400162 Aleksandar Rikalo changed: What|Removed |Added CC||mips3...@gmail.com -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400162] Patch: Guard against __GLIBC_PREREQ for musl libc
https://bugs.kde.org/show_bug.cgi?id=400162 --- Comment #4 from Aleksandar Rikalo --- Created attachment 124625 --> https://bugs.kde.org/attachment.cgi?id=124625=edit Fix none/tests/pth_atfork1 musl compilation error Tested on nanoMIPS/musl/linux and x86/glibc. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400162] Patch: Guard against __GLIBC_PREREQ for musl libc
https://bugs.kde.org/show_bug.cgi?id=400162 Aleksandar Rikalo changed: What|Removed |Added CC||aleksandar.rik...@rt-rk.com --- Comment #3 from Aleksandar Rikalo --- (In reply to Randy MacLeod from comment #2) > Created attachment 115831 [details] > This is the correct patch for this bug Hi Randy, There is no on musl, struct statx is in which is already included. So, you need to remove the last #else branch. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added Attachment #122264|0 |1 is obsolete|| --- Comment #20 from Aleksandar Rikalo --- Created attachment 124469 --> https://bugs.kde.org/attachment.cgi?id=124469=edit Add tests for nanoMIPS instruction set -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added Attachment #122263|0 |1 is obsolete|| --- Comment #19 from Aleksandar Rikalo --- Created attachment 124468 --> https://bugs.kde.org/attachment.cgi?id=124468=edit Fix some tests so they can be compiled for nanoMIPS -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added Attachment #122262|0 |1 is obsolete|| --- Comment #18 from Aleksandar Rikalo --- Created attachment 124467 --> https://bugs.kde.org/attachment.cgi?id=124467=edit Add nanoMIPS support to Valgrind 4/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added Attachment #122259|0 |1 is obsolete|| Attachment #122260|0 |1 is obsolete|| Attachment #122261|0 |1 is obsolete|| --- Comment #17 from Aleksandar Rikalo --- Created attachment 124466 --> https://bugs.kde.org/attachment.cgi?id=124466=edit Add nanoMIPS support to Valgrind 3/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added Attachment #116211|0 |1 is obsolete|| --- Comment #15 from Aleksandar Rikalo --- Created attachment 122264 --> https://bugs.kde.org/attachment.cgi?id=122264=edit Add tests for nanoMIPS instruction set Re-based onto current master branch. There are no regressions on x86 and MIPS platforms. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added Attachment #116204|0 |1 is obsolete|| Attachment #116210|0 |1 is obsolete|| --- Comment #14 from Aleksandar Rikalo --- Created attachment 122263 --> https://bugs.kde.org/attachment.cgi?id=122263=edit Fix some tests so they can be compiled for nanoMIPS -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added Attachment #116208|0 |1 is obsolete|| --- Comment #12 from Aleksandar Rikalo --- Created attachment 122261 --> https://bugs.kde.org/attachment.cgi?id=122261=edit Add nanoMIPS support to Valgrind 3/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added Attachment #116209|0 |1 is obsolete|| --- Comment #13 from Aleksandar Rikalo --- Created attachment 122262 --> https://bugs.kde.org/attachment.cgi?id=122262=edit Add nanoMIPS support to Valgrind 4/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added Attachment #116207|0 |1 is obsolete|| --- Comment #11 from Aleksandar Rikalo --- Created attachment 122260 --> https://bugs.kde.org/attachment.cgi?id=122260=edit Add nanoMIPS support to Valgrind 2/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added Attachment #116206|0 |1 is obsolete|| --- Comment #10 from Aleksandar Rikalo --- Created attachment 122259 --> https://bugs.kde.org/attachment.cgi?id=122259=edit Add nanoMIPS support to Valgrind 1/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400593] In Coregrind, use statx for some internal syscalls if [f]stat[64] fail
https://bugs.kde.org/show_bug.cgi?id=400593 Aleksandar Rikalo changed: What|Removed |Added Attachment #116054|0 |1 is obsolete|| --- Comment #3 from Aleksandar Rikalo --- Created attachment 122180 --> https://bugs.kde.org/attachment.cgi?id=122180=edit Use statx in Coregrind The patch is re-based and `Device ID` conversion is fixed. There is no more regressions on x86 and MIPS systems which support statx. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Aleksandar Rikalo changed: What|Removed |Added CC||mips3...@gmail.com -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 --- Comment #7 from Aleksandar Rikalo --- Here are few patches which implement nanoMIPS support: * Mips specific code refactor - there is no functional impact but it is necessary in order apply other patches. * nanoMIPS support to Valgrind */4 - Add full support for nanoMIPS architecture, including VEX, Coregrind, makefiles, VKI, ..., README. These patches can be merged into one. * Fix some tests so they can be compiled for nanoMIPS - improve problematic tests so they can be compiled for nanoMIPS. * Add tests for nanoMIPS instruction set - adds nanoMIPS tests. Regression tests are done on x86, AMD64, mips32 and mips64. This issue also depends on #400593. Regards, Aleksandar -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 --- Comment #6 from Aleksandar Rikalo --- Created attachment 116211 --> https://bugs.kde.org/attachment.cgi?id=116211=edit Add tests for nanoMIPS instruction set -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 --- Comment #5 from Aleksandar Rikalo --- Created attachment 116210 --> https://bugs.kde.org/attachment.cgi?id=116210=edit Fix some tests so they can be compiled for nanoMIPS -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 --- Comment #4 from Aleksandar Rikalo --- Created attachment 116209 --> https://bugs.kde.org/attachment.cgi?id=116209=edit Add nanoMIPS support to Valgrind 4/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 --- Comment #3 from Aleksandar Rikalo --- Created attachment 116208 --> https://bugs.kde.org/attachment.cgi?id=116208=edit Add nanoMIPS support to Valgrind 3/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 --- Comment #2 from Aleksandar Rikalo --- Created attachment 116207 --> https://bugs.kde.org/attachment.cgi?id=116207=edit Add nanoMIPS support to Valgrind 2/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 --- Comment #1 from Aleksandar Rikalo --- Created attachment 116206 --> https://bugs.kde.org/attachment.cgi?id=116206=edit Add nanoMIPS support to Valgrind 1/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400872] New: Add nanoMIPS support to Valgrind
https://bugs.kde.org/show_bug.cgi?id=400872 Bug ID: 400872 Summary: Add nanoMIPS support to Valgrind Product: valgrind Version: unspecified Platform: Other OS: Linux Status: REPORTED Severity: normal Priority: NOR Component: general Assignee: jsew...@acm.org Reporter: aleksandar.rik...@rt-rk.com Target Milestone: --- Created attachment 116204 --> https://bugs.kde.org/attachment.cgi?id=116204=edit Mips specific code refactor Here are few patches which implement nanoMIPS support. This issue depends on #400593. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400593] Statx in Coregrind
https://bugs.kde.org/show_bug.cgi?id=400593 Aleksandar Rikalo changed: What|Removed |Added CC||mips3...@gmail.com, ||philippe.waroquiers@skynet. ||be -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400593] Statx in Coregrind
https://bugs.kde.org/show_bug.cgi?id=400593 Aleksandar Rikalo changed: What|Removed |Added Summary|Statx |Statx in Coregrind -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 400593] New: Statx
https://bugs.kde.org/show_bug.cgi?id=400593
Bug ID: 400593
Summary: Statx
Product: valgrind
Version: unspecified
Platform: Other
OS: Linux
Status: REPORTED
Severity: normal
Priority: NOR
Component: general
Assignee: jsew...@acm.org
Reporter: aleksandar.rik...@rt-rk.com
Target Milestone: ---
Created attachment 116054
--> https://bugs.kde.org/attachment.cgi?id=116054=edit
Use statx in Coregrind
Hello everyone,
We are going to implement support for nanoMIPS architecture - patches will be
ready soon.
nanoMIPS has no support for [f]stat[64] syscalls, there is only statx and
support for Linux 4.11+.
In order to support nanoMIPS, we need Valgrind (Coregrind) to use statx (as one
of candidates) for internal purposes.
I'm suggesting this patch for the purpose, but it needs to be tested on all
architectures/platforms.
It is tested on AMD64/Linux and MIPS{32,64}/Linux.
Aleksandar
--
You are receiving this mail because:
You are watching all bug changes.
[valgrind] [Bug 396905] compile tests failure on mips32-linux: Error: opcode not supported on this processor: mips1 (mips1)
https://bugs.kde.org/show_bug.cgi?id=396905 Aleksandar Rikalo changed: What|Removed |Added CC||aleksandar.rik...@rt-rk.com --- Comment #2 from Aleksandar Rikalo --- Hi, Unfortunately, this kind of fix makes troubles in case of specific CFLAGS are passed. However, build whole Valgrind for the "mips1" isn't good idea at all. >From `README.mips`: >> The build system expects that native GCC is configured correctly and >> optimized for the platform. Yet, this may not be the case with some Debian distributions which configure GCC to compile to "mips1" by default. Depending on a target platform, using CFLAGS="-mips32r2", CFLAGS="-mips32" or CFLAGS="-mips64" or CFLAGS="-mips64 -mabi=64" will do the trick and compile Valgrind correctly. << Please let us know if ./configure CFLAGS="-mips32" doesn't work for you. Regards, Aleksandar -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345763] MIPS N32 ABI support
https://bugs.kde.org/show_bug.cgi?id=345763 Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> changed: What|Removed |Added Attachment #112104|0 |1 is obsolete|| --- Comment #37 from Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> --- Created attachment 112117 --> https://bugs.kde.org/attachment.cgi?id=112117=edit Fix tests -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345763] MIPS N32 ABI support
https://bugs.kde.org/show_bug.cgi?id=345763 Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> changed: What|Removed |Added Attachment #112103|0 |1 is obsolete|| --- Comment #36 from Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> --- Created attachment 112116 --> https://bugs.kde.org/attachment.cgi?id=112116=edit Add mips N32 ABI support -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345763] MIPS N32 ABI support
https://bugs.kde.org/show_bug.cgi?id=345763 Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> changed: What|Removed |Added Attachment #111846|0 |1 is obsolete|| --- Comment #35 from Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> --- Created attachment 112104 --> https://bugs.kde.org/attachment.cgi?id=112104=edit Fix tests Re-based & arranged patch. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345763] MIPS N32 ABI support
https://bugs.kde.org/show_bug.cgi?id=345763 Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> changed: What|Removed |Added Attachment #111845|0 |1 is obsolete|| --- Comment #34 from Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> --- Created attachment 112103 --> https://bugs.kde.org/attachment.cgi?id=112103=edit Add mips N32 ABI suupport Re-based & arranged patch. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345763] MIPS N32 ABI support
https://bugs.kde.org/show_bug.cgi?id=345763 --- Comment #33 from Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> --- Created attachment 111846 --> https://bugs.kde.org/attachment.cgi?id=111846=edit Fix tests With the last three changes, there are no failing tests on mips N32. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345763] MIPS N32 ABI support
https://bugs.kde.org/show_bug.cgi?id=345763 Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> changed: What|Removed |Added Attachment #92187|0 |1 is obsolete|| Attachment #95761|0 |1 is obsolete|| Attachment #95762|0 |1 is obsolete|| Attachment #106892|0 |1 is obsolete|| Attachment #106893|0 |1 is obsolete|| Attachment #106894|0 |1 is obsolete|| Attachment #106895|0 |1 is obsolete|| --- Comment #32 from Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> --- Created attachment 111845 --> https://bugs.kde.org/attachment.cgi?id=111845=edit Add mips N32 ABI suupport -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345763] MIPS N32 ABI support
https://bugs.kde.org/show_bug.cgi?id=345763 Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> changed: What|Removed |Added CC||aleksandar.rik...@rt-rk.com --- Comment #31 from Aleksandar Rikalo <aleksandar.rik...@rt-rk.com> --- Created attachment 111844 --> https://bugs.kde.org/attachment.cgi?id=111844=edit Introduce RegWord type On all other architectures size of long matches register width. On mips n32 size of long is 32 bits and register width is 64 bits. Valgrind is written with assumption that long size matches register width. This is reason why both UWord for valgrind and HWord for VEX match size of long. Long size differs from register size on mips n32 ABI. This causes problems in several cases: - For do_syscall arguments is used UWord type. This may cause that 64-bit values lost 32 higher bits. - In several places where Valgrind needs to check size of registers, using long based types, would give incorrect value. There are two possible solutions for these problems: - Change size of HWord or UWord for mips n32 abi- Problem with this solution is that they are already used on great number of places where it is assumed that they match long size. Changing their size would require changes in great number of places, and require new type which matches size of long. Change of this proportion would potentially introduce a great deal of new bugs. - Introduce a new type which will match size of registers on all platforms and use it where necessary, which is implemented in provided patch. This patch has no effect on Valgrind's behavior. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 382563] MIPS MSA ASE support
https://bugs.kde.org/show_bug.cgi?id=382563 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added Attachment #108545|0 |1 is obsolete|| --- Comment #18 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 108585 --> https://bugs.kde.org/attachment.cgi?id=108585=edit MIPS MSA 5/5 - MSA regression tests -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 382563] MIPS MSA ASE support
https://bugs.kde.org/show_bug.cgi?id=382563 --- Comment #17 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Hi Petar, Slightly reorganized, refactored and re-based patches are ready. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 382563] MIPS MSA ASE support
https://bugs.kde.org/show_bug.cgi?id=382563 --- Comment #16 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 108545 --> https://bugs.kde.org/attachment.cgi?id=108545=edit MIPS MSA 5/5 - MSA regression tests -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 382563] MIPS MSA ASE support
https://bugs.kde.org/show_bug.cgi?id=382563 --- Comment #15 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 108543 --> https://bugs.kde.org/attachment.cgi?id=108543=edit MIPS MSA 4/5 - Adds support for guest V128 registers on MIPS in Memcheck -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 382563] MIPS MSA ASE support
https://bugs.kde.org/show_bug.cgi?id=382563 --- Comment #14 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 108542 --> https://bugs.kde.org/attachment.cgi?id=108542=edit MIPS MSA 3/5 - Detecting presence of MSA capabilities -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 382563] MIPS MSA ASE support
https://bugs.kde.org/show_bug.cgi?id=382563 --- Comment #13 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 108541 --> https://bugs.kde.org/attachment.cgi?id=108541=edit MIPS MSA 2/5 - Full support of MIPS SIMD Architecture Module (MSA) instruction set for VEX -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 382563] MIPS MSA ASE support
https://bugs.kde.org/show_bug.cgi?id=382563 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added CC||aleksandar.rik...@imgtec.co ||m --- Comment #12 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 108540 --> https://bugs.kde.org/attachment.cgi?id=108540=edit MIPS MSA 1/5 - Introduces new Iops in VEX necessary for implementing some MSA V128 operations -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345307] Warning about "still reachable" memory when using libstdc++ from gcc 5
https://bugs.kde.org/show_bug.cgi?id=345307 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added Attachment #102559|0 |1 is obsolete|| --- Comment #31 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 107894 --> https://bugs.kde.org/attachment.cgi?id=107894=edit leak_cpp_interior additional .exps Re-based patch. Can someone take a look ? -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added Attachment #101431|0 |1 is obsolete|| --- Comment #33 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 105429 --> https://bugs.kde.org/attachment.cgi?id=105429=edit Slightly modified patch for syswrap-generic.c Nothing special here, just use UInt* instead UChar*. Tested on mips32/64-linux and amd64-linux. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #32 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- (In reply to Ivo Raisr from comment #31) > Another patch committed in SVN r16348. Thank You! Please take a look at Tamara's patch for mc_main.c. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added CC||aleksandar.rik...@imgtec.co ||m --- Comment #27 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 105392 --> https://bugs.kde.org/attachment.cgi?id=105392=edit m_libcbase_warnings.diff: Suppress warnings from coregrind/m_libcbase.c (In reply to Ivo Raisr from comment #26) > Aleksandar, do you have another patch ready? Hi Ivo, Acctualy, I have. m_libcbase_warnings.diff: Suppress warnings from coregrind/m_libcbase.c. Function VG_(memset)() - after "while ((!VG_IS_4_ALIGNED(d)))" pointer d is 4-byte aligned, so it is safe to use ASSUME_ALIGNED for UInt* casting. QSORT - According to swaptype variable initialization (m_libcbase.c:817), condition (swaptype <= 1) implies "elements are Word-aligned", so we can use ASSUME_ALIGNED in parts under the above condition ( in BM_SWAP(), BM_PVINIT() and bm_swapfunc() ). This is tested on mips32/64-linux and amd64-linux. Aleksandar -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345307] Warning about "still reachable" memory when using libstdc++ from gcc 5
https://bugs.kde.org/show_bug.cgi?id=345307 --- Comment #30 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Patch from #373069 has been committed. Now we need just https://bugs.kde.org/attachment.cgi?id=102559 to make memcheck/leak_cpp_interior pass with gcc-6. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #25 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- (In reply to Julian Seward from comment #23) > What's the status of this now? Can it be closed? No, it shouldn't. This is just the beginning. I'll prepare next few patches soon. Also, there are patches 2 and 3 which are still uncommitted. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028
--- Comment #19 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> ---
(In reply to Ivo Raisr from comment #17)
> I was also wondering of the impact on other architectures.
> Please could you conduct a quick test to see how the compiler resolves
> ASSUME_ALIGNED on amd64/ppc, for example? And mips/arm for comparison?
I have used following code to see how the GCC resolves ASSUME_ALIGNED on
MIPS32/64, ARM32, S390, PPC32/64 and AMD64:
char *a;
int f1() {
return *ASSUME_ALIGNED(int*, a);
}
int f2() {
return *((int*)a);
}
With -O1 (or higher) generated code for f1() and f2() are identical.
Without optimizations (-O0), f1() has few instructions more than f2(), more
precisely:
- AMD64: 2 additional mov instructions
- PPC32: additional stw/lwz pair
- PPC64: additional std/ld pair
- S390: additional st/l pair
- ARM32: additional str/ldr pair + two instructions for stack arrangement
- MIPS32: additional sw/lw pair
- MIPS64: additional sd/ld pair
--
You are receiving this mail because:
You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added Attachment #101429|0 |1 is obsolete|| --- Comment #18 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 104189 --> https://bugs.kde.org/attachment.cgi?id=104189=edit Patch 1/4 Fixed patch, after comment 17. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #16 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Thank You for review. I agree with changes in comment and also You are right about name. I vote for ASSUME_ALIGNED, in that case, it is clear that there is a assumption. (In reply to Ivo Raisr from comment #15) > I am quite happy with the first patch (1/4). > That particular expression in vki-linux.h has been extensively studied and > found to be benign with regards to potential misaligned access. > > I'd reword the comment for PTR_CAST slightly: > // Some architectures (eg. mips, arm) do not support unaligned memory access > // by hardware, so GCC warn about suspicious situations. This macro could > // be used to avoid these warnings but only after careful examination. > > Also why did you chose name 'PTR_CAST'? Perhaps ALIGN_CAST or ASSUME_ALIGNED > would be more self explanatory? -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #14 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Ivo, Julian, Can we do something about this? It would be nice if we could define the strategy and start to eliminate false warnings. Please take a look on Patch 1/4, if proposed concept is acceptable to you, I will continue to use PTR_CAST. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 375514] valgrind_get_tls_addr() does not work in case of static TLS
https://bugs.kde.org/show_bug.cgi?id=375514 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added Attachment #103619|0 |1 is obsolete|| --- Comment #4 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 103665 --> https://bugs.kde.org/attachment.cgi?id=103665=edit GDB server - valgrind_get_tls_addr() fix for MIPS v2 Suggested comments and dlog(...) are added. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #13 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- (In reply to Julian Seward from comment #9) > This feels to me like hiding misalignment problems. I'd prefer to remove > misaligned > accesses where possible. Building with --enable-usban at least makes it > possible > to see, on any platform, where the run-time misaligned accesses are. Hi Julian, My idea is not to hide all warnings, but to hide "false" ones. There are situations (eg. vki-linux.h:677) where we know (for sure) that address is aligned. Macro introduced in Patch 1/4 should be used to silence compiler in these cases. With -O2, there are no differences in object codes with or without using PTR_CAST macro. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 375514] New: valgrind_get_tls_addr() does not work in case of static TLS
https://bugs.kde.org/show_bug.cgi?id=375514 Bug ID: 375514 Summary: valgrind_get_tls_addr() does not work in case of static TLS Product: valgrind Version: unspecified Platform: Other OS: Linux Status: UNCONFIRMED Severity: normal Priority: NOR Component: general Assignee: jsew...@acm.org Reporter: aleksandar.rik...@imgtec.com CC: mips3...@gmail.com Target Milestone: --- Created attachment 103619 --> https://bugs.kde.org/attachment.cgi?id=103619=edit GDB server - valgrind_get_tls_addr() fix for MIPS This is about valgrind_get_tls_addr() from m_gdbserver/target.c. valgrind_get_tls_addr() fails on MIPS (and maybe on some other platforms) when given module uses static TLS. Proposed patch fixes problem for MIPS. Changes are made according to td_thr_tlsbase() function from GLIBC/nptl_db/td_thr_tlsbase.c. Related test is gdbserver_tests/hgtls. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345307] Warning about "still reachable" memory when using libstdc++ from gcc 5
https://bugs.kde.org/show_bug.cgi?id=345307 --- Comment #29 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Can anyone take a look on this? Thank you in advance! Aleksandar -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #11 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- > Can we simplify the patch #4 just to > # define DO(n) (*fpu)[n] = arch->vex.guest_f##n > ? > Would that work? (Check the attached patch.) Hi Petar, It wouldn't work - (*fpu)[n] are doubles, but arch->vex.guest_f##n are ULongs. In this case we need raw memory move, without long2double conversions. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 345307] Warning about "still reachable" memory when using libstdc++ from gcc 5
https://bugs.kde.org/show_bug.cgi?id=345307 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added CC||aleksandar.rik...@imgtec.co ||m --- Comment #28 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 102559 --> https://bugs.kde.org/attachment.cgi?id=102559=edit leak_cpp_interior additional .exps memcheck/tests/leak_cpp_interior fails consistently on systems with gcc 5.0+ and it would be good if we could find some solution for it. This patch adds additional .exp files for gcc6 case. Together with https://bugsfiles.kde.org/attachment.cgi?id=102518 it makes test pass with gcc 6.0+. Please, see: https://bugs.kde.org/show_bug.cgi?id=373069 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 373069] memcheck/tests/leak_cpp_interior fails with GCC 5.1+
https://bugs.kde.org/show_bug.cgi?id=373069 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added CC||mips3...@gmail.com -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 373069] memcheck/tests/leak_cpp_interior fails with GCC 5.1+
https://bugs.kde.org/show_bug.cgi?id=373069 --- Comment #3 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- (In reply to Ivo Raisr from comment #1) > I cannot comment anything about std::string implementation. > However for your second query have a look at bug: > https://bugs.kde.org/show_bug.cgi?id=345307 Thank You for quick response. I've suggested patch for the first problem. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 373069] memcheck/tests/leak_cpp_interior fails with GCC 5.1+
https://bugs.kde.org/show_bug.cgi?id=373069 --- Comment #2 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 102518 --> https://bugs.kde.org/attachment.cgi?id=102518=edit Turn off CXX11_ABI This patch forces leak_cpp_interior to be compiled using old implementation of std::string. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 373069] New: memcheck/tests/leak_cpp_interior fails with GCC 5.1+
https://bugs.kde.org/show_bug.cgi?id=373069 Bug ID: 373069 Summary: memcheck/tests/leak_cpp_interior fails with GCC 5.1+ Product: valgrind Version: unspecified Platform: unspecified OS: Linux Status: UNCONFIRMED Severity: normal Priority: NOR Component: memcheck Assignee: jsew...@acm.org Reporter: aleksandar.rik...@imgtec.com Target Milestone: --- Memcheck's leak_cpp_interior test has purpose to test heuristics for detecting some common cases, where the pointer is pointing in the middle of the block, but the chunk is not lost. One of such cases is old implementation of string that has format: [_Rep] *_M_length * [basic_string]_M_capacity * _M_dataplus _M_refcount * _M_p > unnamed array of char_type New library ABI introduced in GCC 5.1 includes new implementation of std::string. In new implementation pointer is no longer displaced from the beginning of the chunk. This causes failing of leak_cpp_interior test. >From GCC version 5.1.0, allocation of exceptions in libstdc++ is changed. Before GCC version 5.1.0 emergency buffer was statically allocated. In version 5.1.0 it is swapped with object of class poll with dynamically allocated arena. This change also causes failing of leak_cpp_interior test. Memcheck detects dynamically allocated block of 72,704 bytes for 64bit architectures or 18,688 bytes for 32bit architectures, that is not visible when using older versions of libstdc++. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #7 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Patch 1/4 (warn01.diff) removes 98 "cast-align" warnings caused by including vki-linux.h from various sources. PTR_CAST macro from this patch can be used in future, to avoid "cast-align" warnings. Patch 2/4 (warn02.diff) removes few "cast-align" warnings from coregrind/launcher-linux.c. Patch 3/4 (warn03.diff) removes "cast-align" warning from coregrind/m_syswrap/syswrap-generic.c. Patch 4/4 (warn04.diff) is MIPS specific and removes 32 "cast-equal" warnings from coregrind/m_coredump/coredump-elf.c. These numbers are relevant for MIPS build. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added CC||mips3...@gmail.com -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #5 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- I propose these four independent patches that reduce number of warnings by ~50%. Applying patches: patch -p1 < warn0x.diff -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #4 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 101432 --> https://bugs.kde.org/attachment.cgi?id=101432=edit Patch 4/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #3 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 101431 --> https://bugs.kde.org/attachment.cgi?id=101431=edit Patch 3/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #1 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 101429 --> https://bugs.kde.org/attachment.cgi?id=101429=edit Patch 1/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 --- Comment #2 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 101430 --> https://bugs.kde.org/attachment.cgi?id=101430=edit Patch 2/4 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 370028] New: Reduce the number of compiler warnings on MIPS platforms
https://bugs.kde.org/show_bug.cgi?id=370028 Bug ID: 370028 Summary: Reduce the number of compiler warnings on MIPS platforms Product: valgrind Version: unspecified Platform: unspecified OS: Linux Status: UNCONFIRMED Severity: minor Priority: NOR Component: general Assignee: jsew...@acm.org Reporter: aleksandar.rik...@imgtec.com During compilation on MIPS, GCC issues a lot of "cast-align" and other types of warnings. Reproducible: Always Steps to Reproduce: 1. Regular build Actual Results: GCC issues a lot of warnings. Expected Results: Reasonable number of warnings. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 366079] FPXX Support for MIPS32 Valgrind
https://bugs.kde.org/show_bug.cgi?id=366079 --- Comment #6 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 100684 --> https://bugs.kde.org/attachment.cgi?id=100684=edit MIPS64 version of test -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 366079] FPXX Support for MIPS32 Valgrind
https://bugs.kde.org/show_bug.cgi?id=366079 --- Comment #5 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 100683 --> https://bugs.kde.org/attachment.cgi?id=100683=edit FP32 support on MIPS64 Support for FP32 emulation on MIPS64 platforms (including prctl(GET/SET_FP_MODE) syscalls). -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 366079] FPXX Support for MIPS32 Valgrind
https://bugs.kde.org/show_bug.cgi?id=366079 --- Comment #2 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 100287 --> https://bugs.kde.org/attachment.cgi?id=100287=edit MIPS32 test for syscalls prctl(GET/SET_FP_MODE) -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 366079] FPXX Support for MIPS32 Valgrind
https://bugs.kde.org/show_bug.cgi?id=366079 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added CC||aleksandar.rik...@imgtec.co ||m --- Comment #1 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 100286 --> https://bugs.kde.org/attachment.cgi?id=100286=edit Support for mips32 FPXX With this patch, MIPS32 Valgrind compiled with -mfpxx can handle all types (regarding FP_ABI flag) of MIPS32 ELFs. - Functions arch_elf_pt_proc() and arch_check_elf() are added to elf reader according to linux/fs/binfmt_elf.c from Linux 4.1; - Processing .MIPS.abiflags section and initializing appropriate FPU mode for MIPS32 are added; - Emulation of prctl(GET/SET_FP_MODE) sys-calls are implemented for MIPS32. Applying the patch: patch -p1 < mips_fpxx_full_support.diff The patch is tested on MIPS32 platforms with Linux 3.13 and 4.1. Also, there are no regressions on MIPS64. Regtests need to be performed on other architectures. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 366079] New: FPXX Support for MIPS32 Valgrind
https://bugs.kde.org/show_bug.cgi?id=366079 Bug ID: 366079 Summary: FPXX Support for MIPS32 Valgrind Product: valgrind Version: unspecified Platform: Other OS: Linux Status: UNCONFIRMED Severity: major Priority: NOR Component: general Assignee: jsew...@acm.org Reporter: aleksandar.rik...@imgtec.com Currently, support for MIPS32 FPXX ABI is incomplete. Programs compiled with options -mfp64 or -mfpxx are not interpreted correctly on Valgrind which is compiled with -mfpxx. Reproducible: Always Steps to Reproduce: 1. Build Valgrind with this configure: ./configure --host=mipsel-linux-gnu CFLAGS="-march=mips32r2 -mabi=32 -mfpxx -EL" 2. Build FP test sample (eg. none/tests/mips32/round.c), with -mfp64 3. Run through the Valgrind on Linux 4.1+. Actual Results: Illegal instruction / incorrect calculations. Expected Results: Correctly executed test. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 360008] Contents of Power vr registers contents is not printed correctly when the --vgdb-shadow-registers=yes option is used.
https://bugs.kde.org/show_bug.cgi?id=360008 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added CC||aleksandar.rik...@imgtec.co ||m --- Comment #13 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- R15864/5 fails to build (tested on MIPS and x86) due to missing files: m_gdbserver/power64-core2-valgrind-s1.xml m_gdbserver/power64-core2-valgrind-s2.xml m_gdbserver/power-vsx-valgrind-s1.xml m_gdbserver/power-vsx-valgrind-s2.xml m_gdbserver/power-vsx.xml gcc -m64 -O2 -g -std=gnu99 -Wall -Wmissing-prototypes -Wshadow -Wpointer-arith -Wstrict-prototypes -Wmissing-declarations -Wcast-align -Wcast-qual -Wwrite-strings -Wempty-body -Wformat -Wformat-security -Wignored-qualifiers -Wmissing-parameter-type -Wold-style-declaration -fno-stack-protector -fno-strict-aliasing -fno-builtin -O -g -fno-omit-frame-pointer -fno-strict-aliasing -fpic -fno-builtin -fno-ipa-icf -nodefaultlibs -shared -Wl,-z,interpose,-z,initfirst -m64 -o vgpreload_core-amd64-linux.so vgpreload_core_amd64_linux_so-vg_preloaded.o make[3]: *** No rule to make target `m_gdbserver/power64-core2-valgrind-s1.xml', needed by `all-am'. Stop. make[3]: Leaving directory `/home/aca/posao/valgrind/coregrind' make[2]: *** [all] Error 2 make[2]: Leaving directory `/home/aca/posao/valgrind/coregrind' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/aca/posao/valgrind' make: *** [all] Error 2 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 303877] valgrind doesn't support compressed debuginfo sections.
https://bugs.kde.org/show_bug.cgi?id=303877 --- Comment #34 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- (In reply to Ivo Raisr from comment #31) Thank You for reviewing. I have fixed all things that you mention. The full patch is attached, and the new version of tests is also attached. > 2) When including "tinfl.c", do we want to define "TINFL_HEADER_FILE_ONLY"? > 20) I don't see any coregrind/Makefile changes to build m_debuginfo/tinfl.c? In case we use #include "tinfl.c" without defining TINFL_HEADER_FILE_ONLY, we don't need tinfl.c in Makefile. If we include header only (by defining TINFL_HEADER_FILE_ONLY before #include) then tinfl.c needs to be compiled separately (this solution is applied in rev2 patch). > 5) CEnt.data has now non-fixed size. Why CACHE_ENTRY_SIZE is still used in > various places > around image.c; for example in alloc_CEnt() and realloc_CEnt()? CACHE_ENTRY_SIZE is still in use as default (and minimal) size of cache entry. Larger entries will be allocated in case size of the uncompressed data is grater than CACHE_ENTRY_SIZE. > However I don't have any system with toolchain supporting '-gz' at hand. > I assume you tested on MIPS. Anyone can test on a different architecture or > distribution? It seems that nobody has GCC which supports -gz, so the test is useless for now. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 303877] valgrind doesn't support compressed debuginfo sections.
https://bugs.kde.org/show_bug.cgi?id=303877 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added Attachment #98288|0 |1 is obsolete|| --- Comment #33 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 98375 --> https://bugs.kde.org/attachment.cgi?id=98375=edit Test cases - rev2 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 303877] valgrind doesn't support compressed debuginfo sections.
https://bugs.kde.org/show_bug.cgi?id=303877 --- Comment #30 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Thank you for suggestions. I've prepared few patches: ElfXX_Chdr-build-fixup.diff - tinfl is moved to m_debuginfo, checking for Elf32/64_Chdr structs is placed into "configure" and line endings in tinfl.c are converted to Unix style. Elf32/64_Chdr, SHF_COMPRESSED and ELFCOMPRESS_ZLIB come from libc () and exist only in newer versions of libc. These structures/constants are not present in Kernel, so I didn't move them to vki. We can place them to separate header or pub_tool_libcbase.h, if you think it's more appropriate. zdebug-support.diff - support for .zdebug sections is added to readelf.c. Three kinds of compressed debug sections are supported by Binutils 2.26 (zlib, zlib-gnu and zlib-gabi) and I have tested Valgrind with all of them on MIPS/X86 Linux. According to the manual (https://gcc.gnu.org/onlinedocs/gcc/Debugging-Options.html), GCC should support -gz=zlib and -gz=zlib-gnu options for generating compressed debug sections, but I haven't succeeded to set up the environment in which it works. I've written two simple Memchek tests (compressed_sections_test.diff) which use these options, but they are not tested properly. The order of applying patches (after applying compressed-dwarf-support.diff and tinfl-modifications.diff) : patch -p1 < ElfXX_Chdr-build-fixup.diff patch -p1 < zdebug-support.diff patch -p1 < compressed_sections_test.diff -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 303877] valgrind doesn't support compressed debuginfo sections.
https://bugs.kde.org/show_bug.cgi?id=303877 --- Comment #29 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 98288 --> https://bugs.kde.org/attachment.cgi?id=98288=edit Test cases -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 303877] valgrind doesn't support compressed debuginfo sections.
https://bugs.kde.org/show_bug.cgi?id=303877 --- Comment #28 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 98287 --> https://bugs.kde.org/attachment.cgi?id=98287=edit .zdebug sections support -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 303877] valgrind doesn't support compressed debuginfo sections.
https://bugs.kde.org/show_bug.cgi?id=303877 --- Comment #27 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 98286 --> https://bugs.kde.org/attachment.cgi?id=98286=edit Corrections according to comment #26 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 303877] valgrind doesn't support compressed debuginfo sections.
https://bugs.kde.org/show_bug.cgi?id=303877 --- Comment #23 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- It looks strange... Have you run ./autogen.sh before compiling ? I've tested the same configuration (amd64, linux) with trunk and VALGRIND_3_11_0 revisions. However, the patch is not tested on arm and ppc. (In reply to Quanah Gibson-Mount from comment #22) > These patches don't work with the current release: > > ../coregrind/libcoregrind-amd64-linux.a(libcoregrind_amd64_linux_a-image.o): > In function `get_slowcase': > /tmp/valgrind/valgrind-3.11.0/coregrind/m_debuginfo/image.c:636: undefined > reference to `tinfl_decompress_mem_to_mem' > collect2: error: ld returned 1 exit status > make[3]: *** [memcheck-amd64-linux] Error 1 -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 303877] valgrind doesn't support compressed debuginfo sections.
https://bugs.kde.org/show_bug.cgi?id=303877 --- Comment #21 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 97742 --> https://bugs.kde.org/attachment.cgi?id=97742=edit Avoid warnings from tinfl.c In order to avoid these warnings we need to modify tinfl.c in some way. I suggest the attached solution. MZ_MALLOC/REALLOC/FREE, memcpy, memset, etc. are replaced with appropriate Valgrind functions. Tinfl is also added to Makefile, so it can be used from other parts of project. The patch can be applied after applying compressed-dwarf-support.diff. To apply the patch use command: patch -p1 < tinfl-modifications.diff The alternative solution is to delete "High level decompression functions" from tinfl.c. -- You are receiving this mail because: You are watching all bug changes.
[valgrind] [Bug 303877] valgrind doesn't support compressed debuginfo sections.
https://bugs.kde.org/show_bug.cgi?id=303877 Aleksandar Rikalo <aleksandar.rik...@imgtec.com> changed: What|Removed |Added CC||aleksandar.rik...@imgtec.co ||m --- Comment #19 from Aleksandar Rikalo <aleksandar.rik...@imgtec.com> --- Created attachment 97607 --> https://bugs.kde.org/attachment.cgi?id=97607=edit Compressed debug sections support for Valgrind - rev 1 I suggest the following solution. The patch has been tested on MIPS32/64. Applying the patch: patch -p1 < compressed-dwarf-support.diff Note: tinfl.c from miniz project (https://github.com/richgel999/miniz) is used without modifications. -- You are receiving this mail because: You are watching all bug changes.
