On Fri, Feb 20, 2015 at 5:35 AM, Marcus Shawcroft
<marcus.shawcr...@gmail.com> wrote:
>> case SYMBOL_SMALL_GOTTPREL:
>> {
>> - rtx tmp_reg = gen_reg_rtx (Pmode);
>> + /* In ILP32, the mode of dest can be either SImode or DImode,
>> + while the got entry is always of SImode size. The mode of
>> + dest depends on how dest is used: if dest is assigned to a
>> + pointer (e.g. in the memory), it has SImode; it may have
>> + DImode if dest is dereferenced to access the memeory.
>> + This is why we have to handle three different tlsie_small
>> + patterns here (two patterns for ILP32). */
>> + enum machine_mode mode = GET_MODE (dest);
>> + rtx tmp_reg = gen_reg_rtx (mode);
>> rtx tp = aarch64_load_tp (NULL);
>> - emit_insn (gen_tlsie_small (tmp_reg, imm));
>> - emit_insn (gen_rtx_SET (Pmode, dest, gen_rtx_PLUS (Pmode, tp,
>> tmp_reg)));
>> +
>> + if (mode == ptr_mode)
>> + {
>> + if (mode == DImode)
>> + emit_insn (gen_tlsie_small_di (tmp_reg, imm));
>> + else
>> + {
>> + emit_insn (gen_tlsie_small_si (tmp_reg, imm));
>> + tp = gen_lowpart (mode, tp);
>> + }
>> + }
>> + else
>> + {
>> + gcc_assert (mode == Pmode);
>> + emit_insn (gen_tlsie_small_sidi (tmp_reg, imm));
>> + }
>
>
> Andrew, do you recall the circumstance that trigger the mode == SImode
> code path through the above code? I've been trying to trigger this
> while working on the equivalent tiny address mode code... with no
> success.
IIRC this happened during the compiling of glibc. That is with -fPIC
and with initial-exec tls_model. errno variable in fact.
Thanks,
Andrew Pinski
> /Marcus
