Re: [PATCH 2/3] rtld: Get architecture specific initializer in rtld_global

[Florian Weimer <fw@deneb.enyo.de>]

* H. J. Lu:

> On Sat, May 16, 2020 at 10:52 AM Florian Weimer <fw@deneb.enyo.de> wrote:
>>
>> * H. J. Lu via Libc-alpha:
>>
>> > Include <dl-procruntime.c> to get architecture specific initializer in
>> > rtld_global.
>> > ---
>> >  elf/rtld.c | 2 ++
>> >  1 file changed, 2 insertions(+)
>> >
>> > diff --git a/elf/rtld.c b/elf/rtld.c
>> > index 0016db86a7..44b9fb0b3c 100644
>> > --- a/elf/rtld.c
>> > +++ b/elf/rtld.c
>> > @@ -315,6 +315,8 @@ rtld_hidden_def (_dl_starting_up)
>> >     (except those which cannot be added for some reason).  */
>> >  struct rtld_global _rtld_global =
>> >    {
>> > +    /* Get architecture specific initializer.  */
>> > +#include <dl-procruntime.c>
>> >      /* Generally the default presumption without further information is an
>> >       * executable stack but this is not true for all platforms.  */
>> >      ._dl_stack_flags = DEFAULT_STACK_PERMS,
>>
>> This patch does not build on its own.  I'm surprised that it works in
>> combination with the third patch.
>
> It should build.

It doesn't if I do not apply the third patch.

>> Is _rtld_global really the right place for this data?  Is it even
>> needed from outside the dynamic loader?
>
> Yes,  it matches:
>
> struct rtld_global
> {
> ...
> #include <dl-procruntime.c>
> ...
> };
>
> in sysdeps/generic/ldsodefs.h.

I mean conceptually—it's not clear to me why these fields have to be
in _rtld_global.  It makes things harder to maintain.  Sorry, this is
a more general question, it's not about the patch itself.