On 2021-02-06 15:51 +0100, Frans de Boer wrote:
> On 06/02/2021 15:46, Xi Ruoyao wrote:
> > On 2021-02-06 14:56 +0100, Frans de Boer wrote:
> > > On 06/02/2021 14:34, Pierre Labastie wrote:
> > > > On Sat, 2021-02-06 at 13:43 +0100, Frans de Boer wrote:
> > > > > On 05/02/2021 21:35, Pierre Labastie wrote:
> > > > > > On Fri, 2021-02-05 at 20:54 +0100, Frans de Boer wrote:
> > > > > > > On 05/02/2021 20:16, Frans de Boer wrote:
> > > > > > >
> > > > > > > >
> > > > > > > > On 05/02/2021 16:25, xry...@mengyan1223.wang wrote:
> > > > > > > >
> > > > > > > > >
> > > > > > > > > Hi Frans,
> > > > > > > > >
> > > > > > > > > Could you send the result of
> > > > > > > > >
> > > > > > > > > $LFS_TGT-readelf -a $LFS/lib/libc.so.6 | grep ISA
> > > > > > > > >
> > > > > > > The line '$LFS_TGT-readelf -a $LFS/lib/libc.so.6 | grep ISA'
> > > > > > > returns
> > > > > > > nothing.
> > > > > > You need to use readelf from binutils 2.36. If you have back up to
> > > > > > 2.35.1, it returns nothing.
> > > > > >
> > > > > > Pierre
> > > > > >
> > > > > Okay, I have recompiled ch5, ch6 and ch7 as well as up to and
> > > > > including
> > > > > glibc-2.33 and using binutils-2.36.
> > > > > Attached is the result of that latest config.log (ch8). Other results
> > > > > are found in previous messages.
> > > > >
> > > > > Also, the output of readelf in ch8:
> > > > > Properties: x86 ISA needed: x86-64-baseline, x86-64-v2, x86-64-
> > > > > v3
> > > > > x86 ISA used: x86-64-baseline, x86-64-v2, x86-64-v3, x86-64-v4
> > > > AFAICT, no one among the editors have the first line. Only baseline is
> > > > "needed", not v2 nor v3, although we have more recent CPUs. I cannot
> > > > tell how this "needed" line is determined, but I see in config.log you
> > > > have -march=native. Maybe you could try without any CFLAGS, and check
> > > > that you do not have the v2 and v3 in "needed".
> > > >
> > > > BTW, you say you have the warning, but it does not stop the build does
> > > > it? Everyone seems to have the warning in a couple of tests in binutils
> > > > chapter 8 (ld tests).
> > > > > with as contrast to the ch5 output:
> > > > > Properties: x86 ISA needed: x86-64-baseline
> > > > > x86 ISA used: x86-64-baseline, x86-64-v2, x86-64-v3, x86-64-v4
> > > > >
> > > > > However, still getting the next message during testing.
> > > > > '/sources-base/glibc-2.33/glibc-build/libc.so.6: CPU ISA level is
> > > > > lower
> > > > > than required'
> > > > >
> > > > > After the tests and install, I get the ISA message with everything I
> > > > > do.
> > > > > Maybe the x86_64-baseline is not the original baseline anymore,
> > > > > thereby
> > > > > making my processor and many others obsolete according to the glibc
> > > > > devs?
> > > > > Maybe adding --with-CPU to the configuration might help?
> > > > Well, as said above, try without flags. Then you could add them one by
> > > > one...
> > > >
> > > > Pierre
> > > >
> > > No, the tests do not stop because I use the 'make -k' option. But can't
> > > run anything afterwards because of this error.
> > > BTW, I just tried the --with-cpu=amdfam10 to configure. It halts stating
> > > that this "subspecies" is not supported.
> > >
> > > So, it seems that the glibc dev where playing god and decided that
> > > anything less then the newest processors should have no means to exist
> > > anymore. So, I stick for now with the glib-2.32 version and continue
> > > later.
> > > Maybe someone will figure out how to get rid of this absurd ISA level
> > > check, blocking millions to billions of systems from future updates!!
> > Don't assume the upstream is trying to "fight against you".
> >
> > It's now reported and the upstream is trying to find a solution:
> > https://sourceware.org/bugzilla/show_bug.cgi?id=27318
> >
> > For now the best workaround seems "don't use -march".
> I did not mentioned a fight or anything in that order. Just that this is
> in line with other recent trends of trying to ignore older - but still
> capable - processors and the current worldwide installed base.
You can turn off ISA marker with "libc_cv_include_x86_isa_level=no". Just
append it to glibc configure line, like:
../configure --prefix=/usr \
--disable-werror \
--enable-kernel=3.2 \
--enable-stack-protector=strong \
--with-headers=/usr/include \
libc_cv_slibdir=/lib libc_cv_include_x86_isa_level=no
Add back -march=native and it should be OK.
Upstream will apply a patch like
https://sourceware.org/pipermail/libc-alpha/attachments/20210203/2fcbcba5/attachment-0001.bin
to automatically set libc_cv_include_x86_isa_level=no for -march settings where
the ISA marker does not make sense (like your case).
--
Xi Ruoyao <xry...@mengyan1223.wang>
School of Aerospace Science and Technology, Xidian University
--
http://lists.linuxfromscratch.org/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
Do not top post on this list.
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
http://en.wikipedia.org/wiki/Posting_style
