[PATCH 10/34] HP-PA: use is_end_of_stmt()
John David Anglin
dave.anglin@bell.net
Thu Feb 20 19:46:44 GMT 2025
On 2025-02-20 2:27 a.m., Jan Beulich wrote:
> On 19.02.2025 19:08, John David Anglin wrote:
>> On 2025-02-19 10:26 a.m., Jan Beulich wrote:
>>> ... instead of open-coding it.
>>
>> This might be correct but on HP-PA multiple statements can appear on a line
>> separated by the '!' character. We have this define in tc-hppa.h:
>>
>> /* On the PA, an exclamation point can appear in an instruction. It is
>> used in FP comparison instructions and as an end of line marker.
>> When used in an instruction it will always follow a comma. */
>> #define TC_EOL_IN_INSN(PTR) (*(PTR) == '!' && (PTR)[-1] == ',')
>
> Okay, but: What do you mean to suggest? Are you asking for me to make
> any (further) changes? The patch right now is a pure replacement from
> (undesirable) open-coding to using the designated predicate macro. No
> functional change whatsoever. Hence I'm unclear whether you mean to
> ask for further changes (which may better be in a separate patch), or
> whether that's just a remark with no real implication for the patch.
After more study, I believe your change to tc-hppa.c is correct and the
old code was wrong. is_end_of_line() doesn't appear to check for the '!' line
separator. is_end_of_stmt() checks for both end of line and line separators.
> Tangential question: The macro you quote doesn't permit for white
> space between exclamation mark and comma - is this intentional, or is
> this yet another of the many assumptions towards the scrubber eating
> white space, breaking when scrubbing is skipped (and potentially
> breaking when scrubber behavior changes)?
I believe this was intentional (i.e., whitespace is not allowed in the opcode
field).
--
John David Anglin dave.anglin@bell.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 840 bytes
Desc: OpenPGP digital signature
URL: <https://sourceware.org/pipermail/binutils/attachments/20250220/bebb6381/attachment.sig>
More information about the Binutils
mailing list