[lld] process undefined atoms from shared library only once

Hi Nick,

--start-group/--end-group functionality with the Gnu flavor currently works only if the --start-group/--end-group contains archive files.

If they contain dynamic libraries, the undefined atoms from the dynamic libraries are processed whenever the group is iterated again.

I am trying to find out a way to make the resolver add undefined atoms from the shared library only *once*, when the SharedLibrary is first visited as part of the Group.

The only approach that I can think of is having a std::map in the resolver and using that to process undefined atoms from shared libraries the first time the shared library is visited. I dont like this approach as its not clean.

Any alternative approaches to handle this ?

Thanks

Shankar Easwaran

I’d do that with the nextFile abstraction like this: On the first iteration, a Group would return its member every time getNextFile() is called (the same as the current behavior). On the second and further iterations, the Group should skip all the members whose type is not Archive. By doing this, the core linker sees dynamic libraries (or regular object files) only once even if they are in --{start,end}-group.

Hi Rui,

I thought about this, but there is a catch. The undefined symbols still need to be resolved with exports from dynamic libraries even the second time, or any number of times.

For example :-

ld 1.o --start-group libc.so myprintf.a --end-group

1.o has a undef for myprintf, which is in myprintf.a and myprintf.a calls puts.

The first time when libc.so, undef atoms would be added, no shared library export atoms would be created because there is nothing that libc.so would resolve.

The second time, undefined symbols from shared libraries need to be skipped, and the external symbols that are present in the shared library libc.so need to be looked at.

To complicate things there is also --as-needed/--no-as-needed, where in a shared library is used only if its resolving symbols(the linker may need to handle undefined symbols/external symbols from shared library in any iteration).

Thanks

Shankar Easwaran

Hi Rui,

I thought about this, but there is a catch. The undefined symbols still need
to be resolved with exports from dynamic libraries even the second time, or
any number of times.

For example :-

ld 1.o --start-group libc.so myprintf.a --end-group

1.o has a undef for myprintf, which is in myprintf.a and myprintf.a calls
puts.

The first time when libc.so, undef atoms would be added, no shared library
export atoms would be created because there is nothing that libc.so would
resolve.

The second time, undefined symbols from shared libraries need to be skipped,
and the external symbols that are present in the shared library libc.so need
to be looked at.

To complicate things there is also --as-needed/--no-as-needed, where in a
shared library is used only if its resolving symbols(the linker may need to
handle undefined symbols/external symbols from shared library in any
iteration).

Thanks

Shankar Easwaran

In the --no-as-needed case you just add all the atoms when you hit the
library, and you're done. For the --as-needed case you need to only
add the undefs if something is resolved from it, and then skip it in
the future.

- Michael Spencer