| Message ID | 20210115210616.404156-1-ndesaulniers@google.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | Kbuild: DWARF v5 support | expand |
On Fri, Jan 15, 2021 at 10:06 PM Nick Desaulniers <ndesaulniers@google.com> wrote: > > DWARF v5 is the latest standard of the DWARF debug info format. > > DWARF5 wins significantly in terms of size when mixed with compression > (CONFIG_DEBUG_INFO_COMPRESSED). > > Link: http://www.dwarfstd.org/doc/DWARF5.pdf > > Patch 1 is a cleanup from Masahiro and isn't DWARF v5 specific. > Patch 2 is a cleanup that lays the ground work and isn't DWARF > v5 specific. > Patch 3 implements Kconfig and Kbuild support for DWARFv5. > > Changes from v4: > * drop set -e from script as per Nathan. > * add dependency on !CONFIG_DEBUG_INFO_BTF for DWARF v5 as per Sedat. > * Move LLVM_IAS=1 complexity from patch 2 to patch 3 as per Arvind and > Masahiro. Sorry it took me a few tries to understand the point (I > might still not), but it looks much cleaner this way. Sorry Nathan, I > did not carry forward your previous reviews as a result, but I would > appreciate if you could look again. > * Add Nathan's reviewed by tag to patch 1. > * Reword commit message for patch 3 to mention LLVM_IAS=1 and -gdwarf-5 > binutils addition later, and BTF issue. > * I still happen to see a pahole related error spew for the combination > of: > * LLVM=1 > * LLVM_IAS=1 > * CONFIG_DEBUG_INFO_DWARF4 > * CONFIG_DEBUG_INFO_BTF > Though they're non-fatal to the build. I'm not sure yet why removing > any one of the above prevents the warning spew. Maybe we'll need a v6. > En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and CONFIG_DEBUG_INFO_DWARF4. So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. I had some other small nits commented in the single patches. As requested in your previous patch-series, feel free to add my: Tested-by: Sedat Dilek <sedat.dilek@gmail.com> - Sedat - > Changes from v3: > > Changes as per Arvind: > * only add -Wa,-gdwarf-5 for (LLVM=1|CC=clang)+LLVM_IAS=0 builds. > * add -gdwarf-5 to Kconfig shell script. > * only run Kconfig shell script for Clang. > > Apologies to Sedat and Nathan; I appreciate previous testing/review, but > I did no carry forward your Tested-by and Reviewed-by tags, as the > patches have changed too much IMO. > > Changes from v2: > * Drop two of the earlier patches that have been accepted already. > * Add measurements with GCC 10.2 to commit message. > * Update help text as per Arvind with help from Caroline. > * Improve case/wording between DWARF Versions as per Masahiro. > > Changes from the RFC: > * split patch in 3 patch series, include Fangrui's patch, too. > * prefer `DWARF vX` format, as per Fangrui. > * use spaces between assignment in Makefile as per Masahiro. > * simplify setting dwarf-version-y as per Masahiro. > * indent `prompt` in Kconfig change as per Masahiro. > * remove explicit default in Kconfig as per Masahiro. > * add comments to test_dwarf5_support.sh. > * change echo in test_dwarf5_support.sh as per Masahiro. > * remove -u from test_dwarf5_support.sh as per Masahiro. > * add a -gdwarf-5 cc-option check to Kconfig as per Jakub. > > *** BLURB HERE *** > > Masahiro Yamada (1): > Remove $(cc-option,-gdwarf-4) dependency from CONFIG_DEBUG_INFO_DWARF4 > > Nick Desaulniers (2): > Kbuild: make DWARF version a choice > Kbuild: implement support for DWARF v5 > > Makefile | 13 +++++++--- > include/asm-generic/vmlinux.lds.h | 6 ++++- > lib/Kconfig.debug | 42 +++++++++++++++++++++++++------ > scripts/test_dwarf5_support.sh | 8 ++++++ > 4 files changed, 57 insertions(+), 12 deletions(-) > create mode 100755 scripts/test_dwarf5_support.sh > > -- > 2.30.0.284.gd98b1dd5eaa7-goog >
On 1/15/21 1:53 PM, Sedat Dilek wrote: > On Fri, Jan 15, 2021 at 10:06 PM Nick Desaulniers > <ndesaulniers@google.com> wrote: >> >> DWARF v5 is the latest standard of the DWARF debug info format. >> >> DWARF5 wins significantly in terms of size when mixed with compression >> (CONFIG_DEBUG_INFO_COMPRESSED). >> >> Link: http://www.dwarfstd.org/doc/DWARF5.pdf >> >> Patch 1 is a cleanup from Masahiro and isn't DWARF v5 specific. >> Patch 2 is a cleanup that lays the ground work and isn't DWARF >> v5 specific. >> Patch 3 implements Kconfig and Kbuild support for DWARFv5. >> >> Changes from v4: >> * drop set -e from script as per Nathan. >> * add dependency on !CONFIG_DEBUG_INFO_BTF for DWARF v5 as per Sedat. >> * Move LLVM_IAS=1 complexity from patch 2 to patch 3 as per Arvind and >> Masahiro. Sorry it took me a few tries to understand the point (I >> might still not), but it looks much cleaner this way. Sorry Nathan, I >> did not carry forward your previous reviews as a result, but I would >> appreciate if you could look again. >> * Add Nathan's reviewed by tag to patch 1. >> * Reword commit message for patch 3 to mention LLVM_IAS=1 and -gdwarf-5 >> binutils addition later, and BTF issue. >> * I still happen to see a pahole related error spew for the combination >> of: >> * LLVM=1 >> * LLVM_IAS=1 >> * CONFIG_DEBUG_INFO_DWARF4 >> * CONFIG_DEBUG_INFO_BTF >> Though they're non-fatal to the build. I'm not sure yet why removing >> any one of the above prevents the warning spew. Maybe we'll need a v6. >> > > En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and > CONFIG_DEBUG_INFO_DWARF4. > So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. I suggested not to add !DEBUG_INFO_BTF to CONFIG_DEBUG_INFO_DWARF4. It is not there before and adding this may suddenly break some users. If certain combination of gcc/llvm does not work for CONFIG_DEBUG_INFO_DWARF4 with pahole, this is a bug bpf community should fix. > > I had some other small nits commented in the single patches. > > As requested in your previous patch-series, feel free to add my: > > Tested-by: Sedat Dilek <sedat.dilek@gmail.com> > > - Sedat - > >> Changes from v3: >> >> Changes as per Arvind: >> * only add -Wa,-gdwarf-5 for (LLVM=1|CC=clang)+LLVM_IAS=0 builds. >> * add -gdwarf-5 to Kconfig shell script. >> * only run Kconfig shell script for Clang. >> >> Apologies to Sedat and Nathan; I appreciate previous testing/review, but >> I did no carry forward your Tested-by and Reviewed-by tags, as the >> patches have changed too much IMO. >> >> Changes from v2: >> * Drop two of the earlier patches that have been accepted already. >> * Add measurements with GCC 10.2 to commit message. >> * Update help text as per Arvind with help from Caroline. >> * Improve case/wording between DWARF Versions as per Masahiro. >> >> Changes from the RFC: >> * split patch in 3 patch series, include Fangrui's patch, too. >> * prefer `DWARF vX` format, as per Fangrui. >> * use spaces between assignment in Makefile as per Masahiro. >> * simplify setting dwarf-version-y as per Masahiro. >> * indent `prompt` in Kconfig change as per Masahiro. >> * remove explicit default in Kconfig as per Masahiro. >> * add comments to test_dwarf5_support.sh. >> * change echo in test_dwarf5_support.sh as per Masahiro. >> * remove -u from test_dwarf5_support.sh as per Masahiro. >> * add a -gdwarf-5 cc-option check to Kconfig as per Jakub. >> >> *** BLURB HERE *** >> >> Masahiro Yamada (1): >> Remove $(cc-option,-gdwarf-4) dependency from CONFIG_DEBUG_INFO_DWARF4 >> >> Nick Desaulniers (2): >> Kbuild: make DWARF version a choice >> Kbuild: implement support for DWARF v5 >> >> Makefile | 13 +++++++--- >> include/asm-generic/vmlinux.lds.h | 6 ++++- >> lib/Kconfig.debug | 42 +++++++++++++++++++++++++------ >> scripts/test_dwarf5_support.sh | 8 ++++++ >> 4 files changed, 57 insertions(+), 12 deletions(-) >> create mode 100755 scripts/test_dwarf5_support.sh >> >> -- >> 2.30.0.284.gd98b1dd5eaa7-goog >>
On Fri, Jan 15, 2021 at 3:24 PM Yonghong Song <yhs@fb.com> wrote: > > > > On 1/15/21 1:53 PM, Sedat Dilek wrote: > > En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and > > CONFIG_DEBUG_INFO_DWARF4. > > So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. Can you privately send me your configs that repro? Maybe I can isolate it to a set of configs? > > I suggested not to add !DEBUG_INFO_BTF to CONFIG_DEBUG_INFO_DWARF4. > It is not there before and adding this may suddenly break some users. > > If certain combination of gcc/llvm does not work for > CONFIG_DEBUG_INFO_DWARF4 with pahole, this is a bug bpf community > should fix. Is there a place I should report bugs? > > > > > I had some other small nits commented in the single patches. > > > > As requested in your previous patch-series, feel free to add my: > > > > Tested-by: Sedat Dilek <sedat.dilek@gmail.com> Yeah, I'll keep it if v6 is just commit message changes.
On Fri, Jan 15, 2021 at 3:34 PM Nick Desaulniers <ndesaulniers@google.com> wrote: > > On Fri, Jan 15, 2021 at 3:24 PM Yonghong Song <yhs@fb.com> wrote: > > > > > > > > On 1/15/21 1:53 PM, Sedat Dilek wrote: > > > En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and > > > CONFIG_DEBUG_INFO_DWARF4. > > > So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. > > Can you privately send me your configs that repro? Maybe I can isolate > it to a set of configs? Why privately? To reproduce and fix the issue we (BPF and pahole community) would need the config as well. > > > > > I suggested not to add !DEBUG_INFO_BTF to CONFIG_DEBUG_INFO_DWARF4. > > It is not there before and adding this may suddenly break some users. > > > > If certain combination of gcc/llvm does not work for > > CONFIG_DEBUG_INFO_DWARF4 with pahole, this is a bug bpf community > > should fix. > > Is there a place I should report bugs? bpf@vger.kernel.org (BPF in general) and dwarves@vger.kernel.org (pahole, which seems to be emitting these warnings and having problems with DWARF5). > > > > > > > > > I had some other small nits commented in the single patches. > > > > > > As requested in your previous patch-series, feel free to add my: > > > > > > Tested-by: Sedat Dilek <sedat.dilek@gmail.com> > > Yeah, I'll keep it if v6 is just commit message changes. > > -- > Thanks, > ~Nick Desaulniers
On 1/15/21 3:34 PM, Nick Desaulniers wrote: > On Fri, Jan 15, 2021 at 3:24 PM Yonghong Song <yhs@fb.com> wrote: >> >> >> >> On 1/15/21 1:53 PM, Sedat Dilek wrote: >>> En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and >>> CONFIG_DEBUG_INFO_DWARF4. >>> So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. > > Can you privately send me your configs that repro? Maybe I can isolate > it to a set of configs? > >> >> I suggested not to add !DEBUG_INFO_BTF to CONFIG_DEBUG_INFO_DWARF4. >> It is not there before and adding this may suddenly break some users. >> >> If certain combination of gcc/llvm does not work for >> CONFIG_DEBUG_INFO_DWARF4 with pahole, this is a bug bpf community >> should fix. > > Is there a place I should report bugs? You can send bug report to Arnaldo Carvalho de Melo <acme@kernel.org>, dwarves@vger.kernel.org and bpf@vger.kernel.org. > >> >>> >>> I had some other small nits commented in the single patches. >>> >>> As requested in your previous patch-series, feel free to add my: >>> >>> Tested-by: Sedat Dilek <sedat.dilek@gmail.com> > > Yeah, I'll keep it if v6 is just commit message changes. >
Em Fri, Jan 15, 2021 at 03:43:06PM -0800, Yonghong Song escreveu: > > > On 1/15/21 3:34 PM, Nick Desaulniers wrote: > > On Fri, Jan 15, 2021 at 3:24 PM Yonghong Song <yhs@fb.com> wrote: > > > > > > > > > > > > On 1/15/21 1:53 PM, Sedat Dilek wrote: > > > > En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and > > > > CONFIG_DEBUG_INFO_DWARF4. > > > > So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. > > > > Can you privately send me your configs that repro? Maybe I can isolate > > it to a set of configs? > > > > > > > > I suggested not to add !DEBUG_INFO_BTF to CONFIG_DEBUG_INFO_DWARF4. > > > It is not there before and adding this may suddenly break some users. > > > > > > If certain combination of gcc/llvm does not work for > > > CONFIG_DEBUG_INFO_DWARF4 with pahole, this is a bug bpf community > > > should fix. > > > > Is there a place I should report bugs? > > You can send bug report to Arnaldo Carvalho de Melo <acme@kernel.org>, > dwarves@vger.kernel.org and bpf@vger.kernel.org. I'm coming back from vacation, will try to read the messages and see if I can fix this. Thanks, - Arnaldo > > > > > > > > > > > > > I had some other small nits commented in the single patches. > > > > > > > > As requested in your previous patch-series, feel free to add my: > > > > > > > > Tested-by: Sedat Dilek <sedat.dilek@gmail.com> > > > > Yeah, I'll keep it if v6 is just commit message changes. > >
On Sun, Jan 17, 2021 at 12:14 PM Arnaldo Carvalho de Melo <acme@kernel.org> wrote: > > Em Fri, Jan 15, 2021 at 03:43:06PM -0800, Yonghong Song escreveu: > > > > > > On 1/15/21 3:34 PM, Nick Desaulniers wrote: > > > On Fri, Jan 15, 2021 at 3:24 PM Yonghong Song <yhs@fb.com> wrote: > > > > > > > > > > > > > > > > On 1/15/21 1:53 PM, Sedat Dilek wrote: > > > > > En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and > > > > > CONFIG_DEBUG_INFO_DWARF4. > > > > > So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. > > > > > > Can you privately send me your configs that repro? Maybe I can isolate > > > it to a set of configs? > > > > > > > > > > > I suggested not to add !DEBUG_INFO_BTF to CONFIG_DEBUG_INFO_DWARF4. > > > > It is not there before and adding this may suddenly break some users. > > > > > > > > If certain combination of gcc/llvm does not work for > > > > CONFIG_DEBUG_INFO_DWARF4 with pahole, this is a bug bpf community > > > > should fix. > > > > > > Is there a place I should report bugs? > > > > You can send bug report to Arnaldo Carvalho de Melo <acme@kernel.org>, > > dwarves@vger.kernel.org and bpf@vger.kernel.org. > > I'm coming back from vacation, will try to read the messages and see if > I can fix this. IDK about DWARF v4; that seems to work for me. I was previously observing https://bugzilla.redhat.com/show_bug.cgi?id=1922698 with DWARF v5. I just re-pulled the latest pahole, rebuilt, and no longer see that warning. I now observe a different set. I plan on attending "BPF office hours tomorrow morning," but if anyone wants a sneak peak of the errors and how to reproduce: https://gist.github.com/nickdesaulniers/ae8c9efbe4da69b1cf0dce138c1d2781 (FWIW: some other folks are hitting issues now with kernel's lack of DWARF v5 support: https://bugzilla.redhat.com/show_bug.cgi?id=1922707)
On Wed, Feb 3, 2021 at 5:31 PM Nick Desaulniers <ndesaulniers@google.com> wrote: > > On Sun, Jan 17, 2021 at 12:14 PM Arnaldo Carvalho de Melo > <acme@kernel.org> wrote: > > > > Em Fri, Jan 15, 2021 at 03:43:06PM -0800, Yonghong Song escreveu: > > > > > > > > > On 1/15/21 3:34 PM, Nick Desaulniers wrote: > > > > On Fri, Jan 15, 2021 at 3:24 PM Yonghong Song <yhs@fb.com> wrote: > > > > > > > > > > > > > > > > > > > > On 1/15/21 1:53 PM, Sedat Dilek wrote: > > > > > > En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and > > > > > > CONFIG_DEBUG_INFO_DWARF4. > > > > > > So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. > > > > > > > > Can you privately send me your configs that repro? Maybe I can isolate > > > > it to a set of configs? > > > > > > > > > > > > > > I suggested not to add !DEBUG_INFO_BTF to CONFIG_DEBUG_INFO_DWARF4. > > > > > It is not there before and adding this may suddenly break some users. > > > > > > > > > > If certain combination of gcc/llvm does not work for > > > > > CONFIG_DEBUG_INFO_DWARF4 with pahole, this is a bug bpf community > > > > > should fix. > > > > > > > > Is there a place I should report bugs? > > > > > > You can send bug report to Arnaldo Carvalho de Melo <acme@kernel.org>, > > > dwarves@vger.kernel.org and bpf@vger.kernel.org. > > > > I'm coming back from vacation, will try to read the messages and see if > > I can fix this. > > IDK about DWARF v4; that seems to work for me. I was previously observing > https://bugzilla.redhat.com/show_bug.cgi?id=1922698 > with DWARF v5. I just re-pulled the latest pahole, rebuilt, and no > longer see that warning. > > I now observe a different set. I plan on attending "BPF office hours > tomorrow morning," but if anyone wants a sneak peak of the errors and > how to reproduce: > https://gist.github.com/nickdesaulniers/ae8c9efbe4da69b1cf0dce138c1d2781 > Is there another (easy) way to get your patch set without the b4 tool? Is your patch set present in some patchworks instance, so that I can download it in mbox format, for example? > > (FWIW: some other folks are hitting issues now with kernel's lack of > DWARF v5 support: https://bugzilla.redhat.com/show_bug.cgi?id=1922707) > -- > Thanks, > ~Nick Desaulniers
On Wed, Feb 3, 2021 at 6:58 PM Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote: > > On Wed, Feb 3, 2021 at 5:31 PM Nick Desaulniers <ndesaulniers@google.com> wrote: > > > > On Sun, Jan 17, 2021 at 12:14 PM Arnaldo Carvalho de Melo > > <acme@kernel.org> wrote: > > > > > > Em Fri, Jan 15, 2021 at 03:43:06PM -0800, Yonghong Song escreveu: > > > > > > > > > > > > On 1/15/21 3:34 PM, Nick Desaulniers wrote: > > > > > On Fri, Jan 15, 2021 at 3:24 PM Yonghong Song <yhs@fb.com> wrote: > > > > > > > > > > > > > > > > > > > > > > > > On 1/15/21 1:53 PM, Sedat Dilek wrote: > > > > > > > En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and > > > > > > > CONFIG_DEBUG_INFO_DWARF4. > > > > > > > So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. > > > > > > > > > > Can you privately send me your configs that repro? Maybe I can isolate > > > > > it to a set of configs? > > > > > > > > > > > > > > > > > I suggested not to add !DEBUG_INFO_BTF to CONFIG_DEBUG_INFO_DWARF4. > > > > > > It is not there before and adding this may suddenly break some users. > > > > > > > > > > > > If certain combination of gcc/llvm does not work for > > > > > > CONFIG_DEBUG_INFO_DWARF4 with pahole, this is a bug bpf community > > > > > > should fix. > > > > > > > > > > Is there a place I should report bugs? > > > > > > > > You can send bug report to Arnaldo Carvalho de Melo <acme@kernel.org>, > > > > dwarves@vger.kernel.org and bpf@vger.kernel.org. > > > > > > I'm coming back from vacation, will try to read the messages and see if > > > I can fix this. > > > > IDK about DWARF v4; that seems to work for me. I was previously observing > > https://bugzilla.redhat.com/show_bug.cgi?id=1922698 > > with DWARF v5. I just re-pulled the latest pahole, rebuilt, and no > > longer see that warning. > > > > I now observe a different set. I plan on attending "BPF office hours > > tomorrow morning," but if anyone wants a sneak peak of the errors and > > how to reproduce: > > https://gist.github.com/nickdesaulniers/ae8c9efbe4da69b1cf0dce138c1d2781 > > > > Is there another (easy) way to get your patch set without the b4 tool? > Is your patch set present in some patchworks instance, so that I can > download it in mbox format, for example? $ wget https://lore.kernel.org/lkml/20210130004401.2528717-2-ndesaulniers@google.com/raw -O - | git am $ wget https://lore.kernel.org/lkml/20210130004401.2528717-3-ndesaulniers@google.com/raw -O - | git am If you haven't tried b4 yet, it's quite nice. Hard to go back. Lore also has mbox.gz links. Not sure about patchwork. > > > > > (FWIW: some other folks are hitting issues now with kernel's lack of > > DWARF v5 support: https://bugzilla.redhat.com/show_bug.cgi?id=1922707)
On Thu, Feb 4, 2021 at 3:58 AM Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote: ... > > Is there another (easy) way to get your patch set without the b4 tool? > Is your patch set present in some patchworks instance, so that I can > download it in mbox format, for example? > Just to promote the b4 tool - we have some cool wiki in [1]. Personally, I got in touch with b4 when dealing with the ClangBuiltLinux project. Note: Sometimes it takes a bit for the patch(set) to be available from the LORE link. - Sedat - [1] https://github.com/ClangBuiltLinux/linux/wiki/Command-line-tips-and-tricks#fetching-a-single-patch-or-series-from-lkml
On Thu, Feb 4, 2021 at 9:42 AM Sedat Dilek <sedat.dilek@gmail.com> wrote: > > On Thu, Feb 4, 2021 at 3:58 AM Andrii Nakryiko > <andrii.nakryiko@gmail.com> wrote: > ... > > > > Is there another (easy) way to get your patch set without the b4 tool? > > Is your patch set present in some patchworks instance, so that I can > > download it in mbox format, for example? > > > > Just to promote the b4 tool - we have some cool wiki in [1]. > > Personally, I got in touch with b4 when dealing with the > ClangBuiltLinux project. > > Note: Sometimes it takes a bit for the patch(set) to be available from > the LORE link. > > - Sedat - > > [1] https://github.com/ClangBuiltLinux/linux/wiki/Command-line-tips-and-tricks#fetching-a-single-patch-or-series-from-lkml Honestly, when behind a proxy-server I did not find a trick to use b4 here (DNS resolution problems). I tried with proxychains-ng - NOPE. In the meantime I updated the CBL-wiki page and added v8 as an example :-). - Sedat -
On Wed, Feb 3, 2021 at 7:13 PM Nick Desaulniers <ndesaulniers@google.com> wrote: > > On Wed, Feb 3, 2021 at 6:58 PM Andrii Nakryiko > <andrii.nakryiko@gmail.com> wrote: > > > > On Wed, Feb 3, 2021 at 5:31 PM Nick Desaulniers <ndesaulniers@google.com> wrote: > > > > > > On Sun, Jan 17, 2021 at 12:14 PM Arnaldo Carvalho de Melo > > > <acme@kernel.org> wrote: > > > > > > > > Em Fri, Jan 15, 2021 at 03:43:06PM -0800, Yonghong Song escreveu: > > > > > > > > > > > > > > > On 1/15/21 3:34 PM, Nick Desaulniers wrote: > > > > > > On Fri, Jan 15, 2021 at 3:24 PM Yonghong Song <yhs@fb.com> wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > > On 1/15/21 1:53 PM, Sedat Dilek wrote: > > > > > > > > En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and > > > > > > > > CONFIG_DEBUG_INFO_DWARF4. > > > > > > > > So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. > > > > > > > > > > > > Can you privately send me your configs that repro? Maybe I can isolate > > > > > > it to a set of configs? > > > > > > > > > > > > > > > > > > > > I suggested not to add !DEBUG_INFO_BTF to CONFIG_DEBUG_INFO_DWARF4. > > > > > > > It is not there before and adding this may suddenly break some users. > > > > > > > > > > > > > > If certain combination of gcc/llvm does not work for > > > > > > > CONFIG_DEBUG_INFO_DWARF4 with pahole, this is a bug bpf community > > > > > > > should fix. > > > > > > > > > > > > Is there a place I should report bugs? > > > > > > > > > > You can send bug report to Arnaldo Carvalho de Melo <acme@kernel.org>, > > > > > dwarves@vger.kernel.org and bpf@vger.kernel.org. > > > > > > > > I'm coming back from vacation, will try to read the messages and see if > > > > I can fix this. > > > > > > IDK about DWARF v4; that seems to work for me. I was previously observing > > > https://bugzilla.redhat.com/show_bug.cgi?id=1922698 > > > with DWARF v5. I just re-pulled the latest pahole, rebuilt, and no > > > longer see that warning. > > > > > > I now observe a different set. I plan on attending "BPF office hours > > > tomorrow morning," but if anyone wants a sneak peak of the errors and > > > how to reproduce: > > > https://gist.github.com/nickdesaulniers/ae8c9efbe4da69b1cf0dce138c1d2781 > > > > > > > Is there another (easy) way to get your patch set without the b4 tool? > > Is your patch set present in some patchworks instance, so that I can > > download it in mbox format, for example? > > $ wget https://lore.kernel.org/lkml/20210130004401.2528717-2-ndesaulniers@google.com/raw > -O - | git am > $ wget https://lore.kernel.org/lkml/20210130004401.2528717-3-ndesaulniers@google.com/raw > -O - | git am > > If you haven't tried b4 yet, it's quite nice. Hard to go back. Lore > also has mbox.gz links. Not sure about patchwork. > Ok, I managed to apply that on linux-next, but I can't get past this: ld.lld: error: undefined symbol: pa_trampoline_start >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) ld.lld: error: undefined symbol: pa_trampoline_header >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) ld.lld: error: undefined symbol: pa_trampoline_pgd >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) >>> referenced by trampoline_64.S:142 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:142) >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) ld.lld: error: undefined symbol: pa_wakeup_start >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) ld.lld: error: undefined symbol: pa_wakeup_header >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) ld.lld: error: undefined symbol: pa_machine_real_restart_asm >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) ld.lld: error: undefined symbol: pa_startup_32 >>> referenced by trampoline_64.S:77 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:77) >>> arch/x86/realmode/rm/trampoline_64.o:(trampoline_start) ld.lld: error: undefined symbol: pa_tr_flags >>> referenced by trampoline_64.S:124 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:124) >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) ld.lld: error: undefined symbol: pa_tr_cr4 >>> referenced by trampoline_64.S:138 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:138) >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) ld.lld: error: undefined symbol: pa_tr_efer >>> referenced by trampoline_64.S:146 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:146) >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) >>> referenced by trampoline_64.S:147 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:147) >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) ld.lld: error: undefined symbol: pa_startup_64 >>> referenced by trampoline_64.S:161 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:161) >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) ld.lld: error: undefined symbol: pa_tr_gdt >>> referenced by arch/x86/realmode/rm/trampoline_64.o:(tr_gdt) >>> referenced by reboot.S:28 (/data/users/andriin/linux/arch/x86/realmode/rm/reboot.S:28) >>> arch/x86/realmode/rm/reboot.o:(machine_real_restart_asm) ld.lld: error: undefined symbol: pa_machine_real_restart_paging_off >>> referenced by reboot.S:34 (/data/users/andriin/linux/arch/x86/realmode/rm/reboot.S:34) >>> arch/x86/realmode/rm/reboot.o:(machine_real_restart_asm) ld.lld: error: undefined symbol: pa_machine_real_restart_idt >>> referenced by reboot.S:47 (/data/users/andriin/linux/arch/x86/realmode/rm/reboot.S:47) >>> arch/x86/realmode/rm/reboot.o:(machine_real_restart_asm) ld.lld: error: undefined symbol: pa_machine_real_restart_gdt >>> referenced by reboot.S:54 (/data/users/andriin/linux/arch/x86/realmode/rm/reboot.S:54) >>> arch/x86/realmode/rm/reboot.o:(machine_real_restart_asm) >>> referenced by arch/x86/realmode/rm/reboot.o:(machine_real_restart_gdt) ld.lld: error: undefined symbol: pa_wakeup_gdt >>> referenced by arch/x86/realmode/rm/wakeup_asm.o:(wakeup_gdt) CC arch/x86/mm/numa_64.o CC arch/x86/mm/amdtopology.o HOSTCC arch/x86/entry/vdso/vdso2c make[4]: *** [arch/x86/realmode/rm/realmode.elf] Error 1 make[3]: *** [arch/x86/realmode/rm/realmode.bin] Error 2 make[2]: *** [arch/x86/realmode] Error 2 make[2]: *** Waiting for unfinished jobs.... Hopefully Arnaldo will have better luck. > > > > > > > > (FWIW: some other folks are hitting issues now with kernel's lack of > > > DWARF v5 support: https://bugzilla.redhat.com/show_bug.cgi?id=1922707) > > > -- > Thanks, > ~Nick Desaulniers
Moving a bunch of folks + lists to BCC. On Thu, Feb 4, 2021 at 3:54 PM Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote: > > On Wed, Feb 3, 2021 at 7:13 PM Nick Desaulniers <ndesaulniers@google.com> wrote: > > > > On Wed, Feb 3, 2021 at 6:58 PM Andrii Nakryiko > > <andrii.nakryiko@gmail.com> wrote: > > > > > > On Wed, Feb 3, 2021 at 5:31 PM Nick Desaulniers <ndesaulniers@google.com> wrote: > > > > > > > > On Sun, Jan 17, 2021 at 12:14 PM Arnaldo Carvalho de Melo > > > > <acme@kernel.org> wrote: > > > > > > > > > > Em Fri, Jan 15, 2021 at 03:43:06PM -0800, Yonghong Song escreveu: > > > > > > > > > > > > > > > > > > On 1/15/21 3:34 PM, Nick Desaulniers wrote: > > > > > > > On Fri, Jan 15, 2021 at 3:24 PM Yonghong Song <yhs@fb.com> wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > On 1/15/21 1:53 PM, Sedat Dilek wrote: > > > > > > > > > En plus, I encountered breakage with GCC v10.2.1 and LLVM=1 and > > > > > > > > > CONFIG_DEBUG_INFO_DWARF4. > > > > > > > > > So might be good to add a "depends on !DEBUG_INFO_BTF" in this combination. > > > > > > > > > > > > > > Can you privately send me your configs that repro? Maybe I can isolate > > > > > > > it to a set of configs? > > > > > > > > > > > > > > > > > > > > > > > I suggested not to add !DEBUG_INFO_BTF to CONFIG_DEBUG_INFO_DWARF4. > > > > > > > > It is not there before and adding this may suddenly break some users. > > > > > > > > > > > > > > > > If certain combination of gcc/llvm does not work for > > > > > > > > CONFIG_DEBUG_INFO_DWARF4 with pahole, this is a bug bpf community > > > > > > > > should fix. > > > > > > > > > > > > > > Is there a place I should report bugs? > > > > > > > > > > > > You can send bug report to Arnaldo Carvalho de Melo <acme@kernel.org>, > > > > > > dwarves@vger.kernel.org and bpf@vger.kernel.org. > > > > > > > > > > I'm coming back from vacation, will try to read the messages and see if > > > > > I can fix this. > > > > > > > > IDK about DWARF v4; that seems to work for me. I was previously observing > > > > https://bugzilla.redhat.com/show_bug.cgi?id=1922698 > > > > with DWARF v5. I just re-pulled the latest pahole, rebuilt, and no > > > > longer see that warning. > > > > > > > > I now observe a different set. I plan on attending "BPF office hours > > > > tomorrow morning," but if anyone wants a sneak peak of the errors and > > > > how to reproduce: > > > > https://gist.github.com/nickdesaulniers/ae8c9efbe4da69b1cf0dce138c1d2781 > > > > > > > > > > Is there another (easy) way to get your patch set without the b4 tool? > > > Is your patch set present in some patchworks instance, so that I can > > > download it in mbox format, for example? > > > > $ wget https://lore.kernel.org/lkml/20210130004401.2528717-2-ndesaulniers@google.com/raw > > -O - | git am > > $ wget https://lore.kernel.org/lkml/20210130004401.2528717-3-ndesaulniers@google.com/raw > > -O - | git am > > > > If you haven't tried b4 yet, it's quite nice. Hard to go back. Lore > > also has mbox.gz links. Not sure about patchwork. > > > > Ok, I managed to apply that on linux-next, but I can't get past this: > > ld.lld: error: undefined symbol: pa_trampoline_start > >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) Thanks for testing and the report. Do you have a .config you can send me to reproduce? > > ld.lld: error: undefined symbol: pa_trampoline_header > >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) > > ld.lld: error: undefined symbol: pa_trampoline_pgd > >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) > >>> referenced by trampoline_64.S:142 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:142) > >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) > > ld.lld: error: undefined symbol: pa_wakeup_start > >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) > > ld.lld: error: undefined symbol: pa_wakeup_header > >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) > > ld.lld: error: undefined symbol: pa_machine_real_restart_asm > >>> referenced by arch/x86/realmode/rm/header.o:(real_mode_header) > > ld.lld: error: undefined symbol: pa_startup_32 > >>> referenced by trampoline_64.S:77 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:77) > >>> arch/x86/realmode/rm/trampoline_64.o:(trampoline_start) > > ld.lld: error: undefined symbol: pa_tr_flags > >>> referenced by trampoline_64.S:124 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:124) > >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) > > ld.lld: error: undefined symbol: pa_tr_cr4 > >>> referenced by trampoline_64.S:138 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:138) > >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) > > ld.lld: error: undefined symbol: pa_tr_efer > >>> referenced by trampoline_64.S:146 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:146) > >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) > >>> referenced by trampoline_64.S:147 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:147) > >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) > > ld.lld: error: undefined symbol: pa_startup_64 > >>> referenced by trampoline_64.S:161 (/data/users/andriin/linux/arch/x86/realmode/rm/trampoline_64.S:161) > >>> arch/x86/realmode/rm/trampoline_64.o:(startup_32) > > ld.lld: error: undefined symbol: pa_tr_gdt > >>> referenced by arch/x86/realmode/rm/trampoline_64.o:(tr_gdt) > >>> referenced by reboot.S:28 (/data/users/andriin/linux/arch/x86/realmode/rm/reboot.S:28) > >>> arch/x86/realmode/rm/reboot.o:(machine_real_restart_asm) > > ld.lld: error: undefined symbol: pa_machine_real_restart_paging_off > >>> referenced by reboot.S:34 (/data/users/andriin/linux/arch/x86/realmode/rm/reboot.S:34) > >>> arch/x86/realmode/rm/reboot.o:(machine_real_restart_asm) > > ld.lld: error: undefined symbol: pa_machine_real_restart_idt > >>> referenced by reboot.S:47 (/data/users/andriin/linux/arch/x86/realmode/rm/reboot.S:47) > >>> arch/x86/realmode/rm/reboot.o:(machine_real_restart_asm) > > ld.lld: error: undefined symbol: pa_machine_real_restart_gdt > >>> referenced by reboot.S:54 (/data/users/andriin/linux/arch/x86/realmode/rm/reboot.S:54) > >>> arch/x86/realmode/rm/reboot.o:(machine_real_restart_asm) > >>> referenced by arch/x86/realmode/rm/reboot.o:(machine_real_restart_gdt) > > ld.lld: error: undefined symbol: pa_wakeup_gdt > >>> referenced by arch/x86/realmode/rm/wakeup_asm.o:(wakeup_gdt) > CC arch/x86/mm/numa_64.o > CC arch/x86/mm/amdtopology.o > HOSTCC arch/x86/entry/vdso/vdso2c > make[4]: *** [arch/x86/realmode/rm/realmode.elf] Error 1 > make[3]: *** [arch/x86/realmode/rm/realmode.bin] Error 2 > make[2]: *** [arch/x86/realmode] Error 2 > make[2]: *** Waiting for unfinished jobs.... > > > Hopefully Arnaldo will have better luck. > > > > > > > > > > > > > > (FWIW: some other folks are hitting issues now with kernel's lack of > > > > DWARF v5 support: https://bugzilla.redhat.com/show_bug.cgi?id=1922707) > > > > > > -- > > Thanks, > > ~Nick Desaulniers