Re: [v2] preloader: Allow a debugger to find ld.so's rendezvous-struct
Keno Fischer <keno(a)juliacomputing.com> writes:
@@ -123,6 +123,18 @@ static struct wine_preload_info preload_info[] = { 0, 0 } /* end of list */ };
+/* A fake .dynamic section through which we can inform the debugger of the + address of the chain-loaded dynamic linker's rendezvous struct */ +__asm__( +".section .dynamic,\"aw\"\n" +".global _DYNAMIC\n" +"_DYNAMIC: .quad 21\n" // DT_DEBUG +".quad 0\n" +".quad 0\n" +".quad 0\n" +".previous\n" +);
This is causing a warning here: gcc -m32 -c -o preloader.o preloader.c -I. -I../include -D__WINESRC__ -fno-builtin -Wall -pipe \ -fno-strict-aliasing -Wdeclaration-after-statement -Wempty-body -Wignored-qualifiers \ -Wstrict-prototypes -Wtype-limits -Wunused-but-set-parameter -Wvla -Wwrite-strings -Wpointer-arith \ -Wlogical-op -gdwarf-2 -gstrict-dwarf -fno-omit-frame-pointer -Werror -g -O2 -fno-diagnostics-show-caret {standard input}: Assembler messages: {standard input}:5: Warning: setting incorrect section attributes for .dynamic -- Alexandre Julliard julliard(a)winehq.org
Whether .dynamic is writable depends on the architecture (I believe the only one on which it isn't is MIPS, but I could be wrong about that). Looks like binutils doesn't take that into account when issuing this diagnostic in the assembler and expects .dynamic to have SHF_ALLOC only. Probably something to complain about to them. I don't believe there's an actual problem though, just an incorrect diagnostic.
Keno Fischer <keno(a)juliacomputing.com> writes:
Whether .dynamic is writable depends on the architecture (I believe the only one on which it isn't is MIPS, but I could be wrong about that). Looks like binutils doesn't take that into account when issuing this diagnostic in the assembler and expects .dynamic to have SHF_ALLOC only. Probably something to complain about to them. I don't believe there's an actual problem though, just an incorrect diagnostic.
We don't want to introduce build warnings though, so even if this is a binutils bug it would have to be avoided somehow. -- Alexandre Julliard julliard(a)winehq.org
Ok, that can probably be done with some linker script magic, since this is merely a warning in the assembler. I'll take a look at that. On Thu, Sep 8, 2016 at 10:36 PM, Alexandre Julliard <julliard(a)winehq.org> wrote:
Keno Fischer <keno(a)juliacomputing.com> writes:
Whether .dynamic is writable depends on the architecture (I believe the only one on which it isn't is MIPS, but I could be wrong about that). Looks like binutils doesn't take that into account when issuing this diagnostic in the assembler and expects .dynamic to have SHF_ALLOC only. Probably something to complain about to them. I don't believe there's an actual problem though, just an incorrect diagnostic.
We don't want to introduce build warnings though, so even if this is a binutils bug it would have to be avoided somehow.
-- Alexandre Julliard julliard(a)winehq.org
I had submitted v3 of this patch, but I don't see it on https://source.winehq.org/patches/ anymore. Was there some review comments I missed somewhere? On Thu, Sep 8, 2016 at 10:44 PM, Keno Fischer <keno(a)juliacomputing.com> wrote:
Ok, that can probably be done with some linker script magic, since this is merely a warning in the assembler. I'll take a look at that.
On Thu, Sep 8, 2016 at 10:36 PM, Alexandre Julliard <julliard(a)winehq.org> wrote:
Keno Fischer <keno(a)juliacomputing.com> writes:
Whether .dynamic is writable depends on the architecture (I believe the only one on which it isn't is MIPS, but I could be wrong about that). Looks like binutils doesn't take that into account when issuing this diagnostic in the assembler and expects .dynamic to have SHF_ALLOC only. Probably something to complain about to them. I don't believe there's an actual problem though, just an incorrect diagnostic.
We don't want to introduce build warnings though, so even if this is a binutils bug it would have to be avoided somehow.
-- Alexandre Julliard julliard(a)winehq.org
Bump on the question of what happened to this patch. On Mon, Oct 10, 2016 at 5:03 PM, Keno Fischer <keno(a)juliacomputing.com> wrote:
I had submitted v3 of this patch, but I don't see it on https://source.winehq.org/patches/ anymore. Was there some review comments I missed somewhere?
On Thu, Sep 8, 2016 at 10:44 PM, Keno Fischer <keno(a)juliacomputing.com> wrote:
Ok, that can probably be done with some linker script magic, since this is merely a warning in the assembler. I'll take a look at that.
On Thu, Sep 8, 2016 at 10:36 PM, Alexandre Julliard <julliard(a)winehq.org> wrote:
Keno Fischer <keno(a)juliacomputing.com> writes:
Whether .dynamic is writable depends on the architecture (I believe the only one on which it isn't is MIPS, but I could be wrong about that). Looks like binutils doesn't take that into account when issuing this diagnostic in the assembler and expects .dynamic to have SHF_ALLOC only. Probably something to complain about to them. I don't believe there's an actual problem though, just an incorrect diagnostic.
We don't want to introduce build warnings though, so even if this is a binutils bug it would have to be avoided somehow.
-- Alexandre Julliard julliard(a)winehq.org
Monthly bump on "Where'd my patch go?" On Sat, Nov 19, 2016 at 4:29 PM, Keno Fischer <keno(a)juliacomputing.com> wrote:
Bump on the question of what happened to this patch.
On Mon, Oct 10, 2016 at 5:03 PM, Keno Fischer <keno(a)juliacomputing.com> wrote:
I had submitted v3 of this patch, but I don't see it on https://source.winehq.org/patches/ anymore. Was there some review comments I missed somewhere?
On Thu, Sep 8, 2016 at 10:44 PM, Keno Fischer <keno(a)juliacomputing.com> wrote:
Ok, that can probably be done with some linker script magic, since this is merely a warning in the assembler. I'll take a look at that.
On Thu, Sep 8, 2016 at 10:36 PM, Alexandre Julliard <julliard(a)winehq.org> wrote:
Keno Fischer <keno(a)juliacomputing.com> writes:
Whether .dynamic is writable depends on the architecture (I believe the only one on which it isn't is MIPS, but I could be wrong about that). Looks like binutils doesn't take that into account when issuing this diagnostic in the assembler and expects .dynamic to have SHF_ALLOC only. Probably something to complain about to them. I don't believe there's an actual problem though, just an incorrect diagnostic.
We don't want to introduce build warnings though, so even if this is a binutils bug it would have to be avoided somehow.
-- Alexandre Julliard julliard(a)winehq.org
participants (2)
-
Alexandre Julliard -
Keno Fischer