This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Committed] Fix Darwin Bootstrap (and maybe arm build too) (PR target/39929)


Hi,
  The problem here is that during expanding of PHI nodes
current_ir_type () will return IR_GIMPLE but really we are expanding
from gimple to RTL so we don't mark the PIC offset table as being
needed (setting uses_pic_offset_table to 1).  This will cause problems
on Darwin and maybe arm.  This patch solves the issue by copying what
the MIPS back-end does into both the darwin and the arm back-ends.
The MIPS back-end checks currently_expanding_to_rtl also when it wants
to know we might be expanding to RTL.

Committed as obvious after a bootstrap/test on powerpc-darwin and a
quick build on arm-eabi.

Thanks,
Andrew Pinski

ChangeLog:
* config/darwin.c (machopic_gen_offset): Check
currently_expanding_to_rtl if current_ir_type returns IR_GIMPLE.
* config/arm/arm.c (require_pic_register): Likewise.

Attachment: fixdarwin.diff.txt
Description: Text document


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]