doc: md.texi (Insn Splitting): Tweak wording for readability.
Checks
Commit Message
Jan, did I get this right? This was from your
r0-36413-g6b24c25948265c / svn r44249, now on its 22nd year!
I spot-checked the pdf for readability. Also calling on a
doc maintainer to check grammos etc. Ok to commit?
-- >8 --
I needed to check what was allowed in a define_split, but
had problems understanding what was meant by "Splitting of
jump instruction into sequence that over by another jump
instruction".
* doc/md.texi (Insn Splitting): Tweak wording for readability.
---
gcc/doc/md.texi | 30 +++++++++++++++---------------
1 file changed, 15 insertions(+), 15 deletions(-)
Comments
On 3/13/23 19:25, Hans-Peter Nilsson via Gcc-patches wrote:
> Jan, did I get this right? This was from your
> r0-36413-g6b24c25948265c / svn r44249, now on its 22nd year!
>
> I spot-checked the pdf for readability. Also calling on a
> doc maintainer to check grammos etc. Ok to commit?
>
> -- >8 --
> I needed to check what was allowed in a define_split, but
> had problems understanding what was meant by "Splitting of
> jump instruction into sequence that over by another jump
> instruction".
>
> * doc/md.texi (Insn Splitting): Tweak wording for readability.
Thanks for noticing this! I can't comment on technical correctness, but
I do have some further suggestions on wording below.
> ---
> gcc/doc/md.texi | 30 +++++++++++++++---------------
> 1 file changed, 15 insertions(+), 15 deletions(-)
>
> diff --git a/gcc/doc/md.texi b/gcc/doc/md.texi
> index 8e3113599fdc..786365143179 100644
> --- a/gcc/doc/md.texi
> +++ b/gcc/doc/md.texi
> @@ -8756,21 +8756,21 @@ insns that don't. Instead, write two separate @code{define_split}
> definitions, one for the insns that are valid and one for the insns that
> are not valid.
>
> -The splitter is allowed to split jump instructions into sequence of
> -jumps or create new jumps in while splitting non-jump instructions. As
> -the control flow graph and branch prediction information needs to be updated,
> -several restriction apply.
> -
> -Splitting of jump instruction into sequence that over by another jump
> -instruction is always valid, as compiler expect identical behavior of new
> -jump. When new sequence contains multiple jump instructions or new labels,
> -more assistance is needed. Splitter is required to create only unconditional
> -jumps, or simple conditional jump instructions. Additionally it must attach a
> -@code{REG_BR_PROB} note to each conditional jump. A global variable
> -@code{split_branch_probability} holds the probability of the original branch in case
> -it was a simple conditional jump, @minus{}1 otherwise. To simplify
> -recomputing of edge frequencies, the new sequence is required to have only
> -forward jumps to the newly created labels.
> +The splitter is allowed to split jump instructions into a sequence of jumps or
> +create new jumps while splitting non-jump instructions. As the control flow
> +graph and branch prediction information needs to be updated, several
> +restrictions apply.
Maybe "needs to be updated after the splitter runs"?
> +
> +Splitting of a jump instruction into a sequence that has another jump
> +instruction is always valid, as the compiler expects identical behavior of the
> +new jump.
Maybe "another jump instruction to the same label"?
> When the new sequence contains multiple jump instructions or new
> +labels, more assistance is needed. The splitter is required to create only
s/required/permitted/
> +unconditional jumps, or simple conditional jump instructions. Additionally it
> +must attach a @code{REG_BR_PROB} note to each conditional jump. A global
> +variable @code{split_branch_probability} holds the probability of the original
> +branch in case it was a simple conditional jump, @minus{}1 otherwise. To
> +simplify recomputing of edge frequencies, the new sequence is required to have
> +only forward jumps to the newly created labels.
s/required/permitted/ again
s/newly created/newly-created/
>
> @findex define_insn_and_split
> For the common case where the pattern of a define_split exactly matches the
-Sandra
@@ -8756,21 +8756,21 @@ insns that don't. Instead, write two separate @code{define_split}
definitions, one for the insns that are valid and one for the insns that
are not valid.
-The splitter is allowed to split jump instructions into sequence of
-jumps or create new jumps in while splitting non-jump instructions. As
-the control flow graph and branch prediction information needs to be updated,
-several restriction apply.
-
-Splitting of jump instruction into sequence that over by another jump
-instruction is always valid, as compiler expect identical behavior of new
-jump. When new sequence contains multiple jump instructions or new labels,
-more assistance is needed. Splitter is required to create only unconditional
-jumps, or simple conditional jump instructions. Additionally it must attach a
-@code{REG_BR_PROB} note to each conditional jump. A global variable
-@code{split_branch_probability} holds the probability of the original branch in case
-it was a simple conditional jump, @minus{}1 otherwise. To simplify
-recomputing of edge frequencies, the new sequence is required to have only
-forward jumps to the newly created labels.
+The splitter is allowed to split jump instructions into a sequence of jumps or
+create new jumps while splitting non-jump instructions. As the control flow
+graph and branch prediction information needs to be updated, several
+restrictions apply.
+
+Splitting of a jump instruction into a sequence that has another jump
+instruction is always valid, as the compiler expects identical behavior of the
+new jump. When the new sequence contains multiple jump instructions or new
+labels, more assistance is needed. The splitter is required to create only
+unconditional jumps, or simple conditional jump instructions. Additionally it
+must attach a @code{REG_BR_PROB} note to each conditional jump. A global
+variable @code{split_branch_probability} holds the probability of the original
+branch in case it was a simple conditional jump, @minus{}1 otherwise. To
+simplify recomputing of edge frequencies, the new sequence is required to have
+only forward jumps to the newly created labels.
@findex define_insn_and_split
For the common case where the pattern of a define_split exactly matches the