[Bug dynamic-link/11767] RFE: dlopen of in-memory ET_DYN or ET_EXEC object

[stsp at users dot sourceforge.net]

https://sourceware.org/bugzilla/show_bug.cgi?id=11767

--- Comment #59 from Stas Sergeev <stsp at users dot sourceforge.net> ---
(In reply to Vinícius dos Santos Oliveira from comment #58)
> FreeBSD approach is just the logical conclusion to the same trend. It
> doesn't try to "innovate". IMO it's the approach that brings the least
> amount of surprises. IMO we should avoid unexpected surprises.

If fdlopen() existed, I'd be
using it for things like memfd
and tmpfs fds, I don't know how
safe is it to use st_ino/st_dev
for those. There is a risk that
some lib is loaded both from an
fs and from the in-memory copy
via memfd, and then you have
problems.
Also note that the proposal here
(and the patches I posted elsewhere)
are about the memory-based dlmem(),
which can't use st_ino stuff.
So while I don't think its a
real problem for fdlopen(), at
least the things like Build-ID
could be considered by those who
implement an extensions in this
area.

-- 
You are receiving this mail because:
You are on the CC list for the bug.