[llvm-dev] some questions on libunwind

Schrodinger ZHU Yifan via llvm-dev llvm-dev at lists.llvm.org
Tue Dec 21 00:38:43 PST 2021


Thanks so much for your answers and suggestions! There is no much comment in libunwind's header. Maybe it would be better to state some assumptions and explanations in the src comments or associated docs?

Schrodinger ZHU Yifan
School of Data Science, CUHK(SZ)
Github: SchrodingerZhu
Twitter: ZhuSchrodinger

Sent with ProtonMail Secure Email.

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐

On Monday, December 20th, 2021 at 6:53 PM, PoYao Chang <rzhboyao at gmail.com> wrote:

> Hi,
> 

> For your first question, the installation target was added last week:https://github.com/llvm/llvm-project/commit/1c4867e
> 

> On Mon, Dec 20, 2021 at 12:37 PM Schrodinger ZHU Yifan via llvm-dev <llvm-dev at lists.llvm.org> wrote:
> 

> > Hi,
> > 

> > These questions were originally posted on Discord but I have got no reply at all so I am writing this email to post the questions to the developer list.
> > 

> > Background
> > 

> > 1. https://github.com/jemalloc/jemalloc/issues/585
> > 2. https://reviews.llvm.org/D9744
> > 3. My team was implementing a database and we planned to add a debug tool for memory profiling when jemalloc is enabled. However, nongnu libunwind caused deadlock from time to time when used together with jeprof.
> > 4. I am in charge of migrating my team's toolchain from GCC to LLVM; besides clang, we want to make all of our runtimes (except libc, for the time being) in LLVM, including libcxx, omp, compiler-rt, libunwind
> > 

> > My problems:
> > 1. libunwind.h is provided in the LLVM repo, but I am curious that why it is not an installation target. Currently, my solution to vendor a LLVM's header within my team's repo.
> > 2. nongnu implementation exposes a specialized unw_init_local2 for stack capturing initiated at signal frame, but this function is not provided at LLVM's world. According to the test cases in LLVM's repo, I inferred that LLVM does not need user to specify the signal frame argument. Is this true?
> > 3. Is there any previous discussion in LLVM side on the deadlock situation of jemalloc and libunwind?
> > 

> > Best wishes.
> > 

> > Schrodinger ZHU Yifan
> > School of Data Science, CUHK(SZ)
> > Website: https://github.com/schrodingerzhu
> > Github: SchrodingerZhu
> > Twitter: ZhuSchrodinger
> > 

> > Sent with ProtonMail Secure Email.
> > 

> > _______________________________________________
> > 

> > LLVM Developers mailing list
> > 

> > llvm-dev at lists.llvm.org
> > 

> > https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20211221/212fcd8f/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: publickey - i at zhuyi.fan - 0xE405E45C.asc
Type: application/pgp-keys
Size: 3237 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20211221/212fcd8f/attachment.key>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 855 bytes
Desc: OpenPGP digital signature
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20211221/212fcd8f/attachment.sig>


More information about the llvm-dev mailing list