1 |
On 03/29/15 15:07, Matt Turner wrote: |
2 |
> On Sun, Mar 29, 2015 at 11:58 AM, Matthias Schwarzott <zzam@g.o> wrote: |
3 |
>> Hi there! |
4 |
>> |
5 |
>> I updated my ~amd64 system recently to new hardware (Intel Core i3-4160). |
6 |
>> Since then valgrind did no longer work for 32bit programs because |
7 |
>> "-march=native" did choose instructions that valgrind does not support |
8 |
>> in 32bit mode (even ld.so was unusable). |
9 |
>> |
10 |
>> After some research I put this into make.conf and now it works: |
11 |
>> CFLAGS_x86="${CFLAGS_x86} -mno-avx2 -mno-sse4 -mno-bmi -mno-bmi2" |
12 |
>> CXXFLAGS_x86="${CXXFLAGS_x86} -mno-avx2 -mno-sse4 -mno-bmi -mno-bmi2" |
13 |
>> |
14 |
>> Is this the best solution to the problem? |
15 |
>> If yes, the valgrind ebuild could suggest something like this. |
16 |
>> Either always show it or check cpu-flags first (is this maintainable?). |
17 |
> Valgrind's policy is that they don't implement new instruction sets in 32-bit. |
18 |
> |
19 |
> Doing what you've done is the only option I'm aware of, short of |
20 |
> implementing support in valgrind for these instructions. |
21 |
> |
22 |
|
23 |
This is a known issue. I could add a pkg_postinst() message, but with |
24 |
valgrind, there would be lots of caveats. Maybe open a bug and we'll |
25 |
track the issue in gentoo that way, like the strlen issue and a few |
26 |
other valgrind oldies. |
27 |
|
28 |
-- |
29 |
Anthony G. Basile, Ph.D. |
30 |
Gentoo Linux Developer [Hardened] |
31 |
E-Mail : blueness@g.o |
32 |
GnuPG FP : 1FED FAD9 D82C 52A5 3BAB DC79 9384 FA6E F52D 4BBA |
33 |
GnuPG ID : F52D4BBA |