[PATCH 12/13] dlfcn,elf: implement dlmem() [BZ #11767]

[stsp]

29.03.2023 20:20, Carlos O'Donell пишет:
> It is not a reviewers task to show this to you, it is only their task that
> they tell you that it exists. The onus is strongly on the patch submitter to
> produce evidence to the contrary. The reviewers on the list here are a scarce
> resource, and they offer the experience and technical knowledge to help you.

OK, I'll try to demonstrate the contrary,
but I am not even sure in what way you
want such a demonstration...
But lets try:

$ LD_LIBRARY_PATH=..:. ./tst-dlmem-fdlopen
unaligned buf gives buffer not aligned: Invalid argument
7fb413101000-7fb413102000 r--p 00000000 00:28 17195405 
/home/stas/src/glibc-dev/build/dlfcn/glreflib1.so
7fb413102000-7fb413103000 r-xp 00001000 00:28 17195405 
/home/stas/src/glibc-dev/build/dlfcn/glreflib1.so
7fb413103000-7fb413104000 r--p 00002000 00:28 17195405 
/home/stas/src/glibc-dev/build/dlfcn/glreflib1.so
7fb413104000-7fb413105000 r--p 00002000 00:28 17195405 
/home/stas/src/glibc-dev/build/dlfcn/glreflib1.so
7fb413105000-7fb413106000 rw-p 00003000 00:28 17195405 
/home/stas/src/glibc-dev/build/dlfcn/glreflib1.so

This is the output of my test-case in v9.
As you can see, I tried to code up the proof
already there.
"buffer not aligned" message comes from
the test that tries to submit an unaligned
buffer, so that's fine.
The next lines show you exactly the fact
that an elf was laid out per vaddr's. If it
wouldn't be the case then /proc/self/maps
would just have 1 reference to the .so file,
and in that example it does have 5.

Is this a proof?
If not - what kind of proof do you need?