[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20251212154253.GW3520958@port70.net>
Date: Fri, 12 Dec 2025 16:42:53 +0100
From: Szabolcs Nagy <nsz@...t70.net>
To: David Edelsohn <dje.gcc@...il.com>, musl@...ts.openwall.com,
Alex Rønne Petersen <alex@...xrp.com>
Subject: Re: [PATCH] powerpc64: add nop after b to __dlsym
* Szabolcs Nagy <nsz@...t70.net> [2025-12-12 16:06:25 +0100]:
> * David Edelsohn <dje.gcc@...il.com> [2025-12-11 20:35:48 -0500]:
> > On Thu, Dec 11, 2025 at 8:10 PM Alex Rønne Petersen <alex@...xrp.com> wrote:
> >
> > > To avoid errors like this:
> > >
> > > ld.lld: error:
> > > /home/ci/.cache/act/2c84ccdd36a924eb/hostexecutor/zig-global-cache/o/9e32a733e976f6f0194dc11bcf100f7a/libc.a(/home/ci/.cache/act/2c84ccdd36a924eb/hostexecutor/zig-global-cache/o/19fa681ef1630f481aadf66ec88d2634/dlsym.o):(function
> > > dlsym: .text+0xc): call to save___dlsym lacks nop, can't restore toc
> > > ---
> > > src/ldso/powerpc64/dlsym.s | 1 +
> > > 1 file changed, 1 insertion(+)
> > >
> > > diff --git a/src/ldso/powerpc64/dlsym.s b/src/ldso/powerpc64/dlsym.s
> > > index a14715fd..73e70db9 100644
> > > --- a/src/ldso/powerpc64/dlsym.s
> > > +++ b/src/ldso/powerpc64/dlsym.s
> > > @@ -8,4 +8,5 @@ dlsym:
> > > .localentry dlsym,.-dlsym
> > > mflr 5 # The return address is arg3.
> > > b __dlsym
> > > + nop
> > >
> >
> > The instruction is
> >
> > b __dlsym
> >
> > which is a branch -- a tail call. It cannot return. It will never need a
> > TOC restore instruction after the branch.
> >
> > As with the other patch, are you fixing actual bugs or bloating the code to
> > remove false-positive linker error messages?
> >
> > If the target of the call (bl) in the previous patch is known to be local,
> > it will never need a TOC restore.
>
> alan modra implies here that the linker
> logic is based on relocs (which is
> presumably the same for b and bl)
> https://sourceware.org/bugzilla/show_bug.cgi?id=26164
>
> i think the linker could check the instruction
> at the target location of a reloc in the inputs
> and decide how to generate the output based on
> that, but maybe that was deemed too complex for
on a second thought the assembler should be
able to generate @notoc for b, i.e. always
use the _NOTOC relocs automatically for b sym.
presumably there are reasons why the compiler
generates @notoc instead. i also dont know
why cpu=power10 is needed to get tail calls
and @notoc. but i wont dig further.
> the linker abi (iirc tocs and other stubs
> can already cause headaches as they change
> section layout late and thus need iterative
> adjustments that may not converge).
>
> >
> > David
> >
> >
> > > .size dlsym, .-dlsym
> > > --
> > > 2.51.0
> > >
> > >
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
