JaeSang Yoo

I'm also interested in firefox profiler. Maybe we could do this job together.

As I previously discussed with @dongwooklee96, previous [empty_profile.json.zip](https://github.com/namhyung/uftrace/files/9350074/empty_profile.json.zip) were converted from chrome trace to firefox trace. But, the content of this profile doesn't show any event or function information. It...

I also discovered similar issue with this. (not postgresql, but SEGFAULT on uftrace recording) I've tested uftrace on Asahi linux(the Apple M1-ported version linux). Tried foobar example in the [uftrace.github.io/slide/](https://uftrace.github.io/slide/)...

@honggyukim, Maybe you can reference these steps. ## build This step is same as your build instruction ``` $ git clone https://github.com/postgres/postgres.git $ cd postgres $ ./configure --prefix=`pwd`/build $ make...

@dongwooklee96, I tried build & test postgres, and it seems segment fault not occurs on dynamic tracing. I failed to build postgres with `-pg` or `-finstrument-functions`. Is there any easier...

By running with gdb, it seems SEGFAULT invoked by LD ```text (gdb) r foobar Starting program: /usr/local/bin/uftrace foobar [Thread debugging using libthread_db enabled] Using host libthread_db library "/usr/lib/libthread_db.so.1". [Attaching after...

Maybe this could be issue with paging size. It uses [16K page](https://asahilinux.org/2022/03/asahi-linux-alpha-release/#known-broken-applications) due to M1's IOMMU. Maybe I need to test similar condition with RPi with 16K paging. Whenever you...

I tried by replacing hard-coded `4096` to `getpagesize()` calls, but it still invokes SEGFAULT. As the gdb backtrace says, it invokes on `_start` procedure, as I know, it's early codes...

> Any reason do you think it's related to page size? It'd be nice if you can have symbol in the dynamic loader ('/lib/ld-linux-aarch64.so.1'). Well, page size was known difference...

+1 to this!