From: Matthew Malcomson
This commit just defines the new names -- as yet don't implement them.
Saving this commit because this is one decision, and recording
what the decision was and why:
Adding new floating point builtins for each floating point type that
is defined in the general code *except
From: Matthew Malcomson
Hello, this is an RFC for adding an atomic floating point fetch_add builtin
(and variants) to GCC. The atomic fetch_add operation is defined to work
on the base floating point types in the C++20 standard chapter 31.7.3, and
extended to work for all cv-unqualified floating
From: Matthew Malcomson
Signed-off-by: Matthew Malcomson
---
gcc/doc/extend.texi | 12
1 file changed, 12 insertions(+)
diff --git a/gcc/doc/extend.texi b/gcc/doc/extend.texi
index 66c99ef7a66..a3e3e7da5d6 100644
--- a/gcc/doc/extend.texi
+++ b/gcc/doc/extend.texi
@@ -13501,6 +135
From: Matthew Malcomson
Need to implement something in the
Things implemented in this patch:
1) Update the optabs definitions to include floating point versions of
atomic fetch_add variants.
2) When expanding into a CAS loop in RTL because the floating point
optab is not implemented, there
From: Matthew Malcomson
Do demo implementation in AArch64 since that's the backend I'm most
familiar with.
Nothing much else to say -- nice to see that the demo implementation
seems to work as expected (being used for fetch_add, add_fetch and
sub_fetch even though it's only defined for fetch_sub
From: Matthew Malcomson
Have a bit of an ugly mapping from floating point type to the builtin
using that type. Would like to find some code-sharing between this, the
function (in a later patch in this series) that finds the relevant mode
from a given builtin, and the general sync-builtins.def fi
From: Matthew Malcomson
Points to question here are:
1) Whether checking for this particular internal builtin is OK (this one
happens to be the one implementing the operation for a `double`, we
would have to rely on the approach that if anyone implements this
operation for a `double` the
From: Matthew Malcomson
As it stands there are still a few things to look at whether they could
be improved:
1) Need to find the exact version of automake to use. I'm using
automake 1.15.1 from https://ftp.gnu.org/gnu/automake/ but the
header is claiming I'm using automake 1.15.
2) The int
From: Matthew Malcomson
N.b. we match which function to use based on the MODE of the type for
two reasons:
1) Can't match directly on type as otherwise `typedef float x` would
mean that `x` could no longer be used with that intrinsic.
2) MODE (i.e. the types ABI) is the thing that we need to d
From: Matthew Malcomson
Hello,
This is the revision of the RFC posted here:
https://gcc.gnu.org/pipermail/gcc-patches/2024-September/663355.html
This patchset introduces floating point versions of atomic fetch_add,
fetch_sub, add_fetch and sub_fetch. Instructions for performing these
operation
From: Matthew Malcomson
N.b. including docs maintainers to ask about whether this flag should be
documented in invoke.texi -- I suspect that would be determined partly
based on the feedback I get on this patch about whether it should be
something mostly for the testsuite and otherwise internally
From: Matthew Malcomson
Not sure who to Cc for this. Honestly just guessing a bit here. Please
do redirect me if anyone knows of a better set of people to ask.
-- >8 --- 8< ---
Update the optabs definitions to include floating point versions of
atomic fetch_add variants
From: Matthew Malcomson
Cc'ing in build machinery maintainers who can hopefully check the
validity of those changes.
N.b. would appreciate information on who to Cc in for the libatomic
parts of this patch. I don't see anyone in the MAINTAINERS file against
libatomic.
-- >8 ---
From: Matthew Malcomson
N.b. including testsuite maintainers hoping for review of the new
effective targets I've added.
-- >8 --- 8< ---
Tests we include are:
- A set of tests checking the observed arithmetic behaviour of the
new builtins.
- A test ensuring the floating
From: Matthew Malcomson
Not sure who to Cc for this. Honestly just guessing a bit here. Please
do redirect me if anyone knows of a better set of people to ask.
-- >8 --- 8< ---
This commit just defines the new names -- as yet don't implement them.
Saving this commit bec
From: Matthew Malcomson
Use wrapper in builtins.cc to check if required. N.b. if there are any
floating point types that are available but not as arithmetic types or
simply not available, that should be handled by the sync_resolve_size
check against available types. We add an assertion in the C
From: Matthew Malcomson
Do not mention suffixed versions since integral suffixed versions are
not mentioned in existing docs.
gcc/ChangeLog:
* doc/extend.texi: Document ability to use floating point atomic
fetch_add/fetch_sub/add_fetch/sub_fetch builtins.
Signed-off-by: Matthew
From: Matthew Malcomson
Do demo implementation in AArch64 since that's the backend I'm most
familiar with.
Nothing much else to say -- nice to see that the demo implementation
seems to work as expected (being used for fetch_add, add_fetch and
sub_fetch even though it's only defined for fetch_sub
From: Matthew Malcomson
This is not to go into the upstream. This was used for testing.
This patch set is supposed to be upstreamed in combination with a
patchset automatically linking in libatomic when compiling.
The two patchsets are being worked on in parallel. Some of the tests in
this ser
From: Matthew Malcomson
This setting seems to better match the indentation that is used in GCC.
Adds an exra level of indentation after braces in a case statement.
Only manual testing done on the switch statements in
c-common.cc:resolve_overloaded_builtin and
alias.cc:record_component_aliases.
From: Matthew Malcomson
The previous value of TopLevel meant that the function name of
declarations would also be on a new line. This does not match the
current formatting of headers.
Manual testing done on c-common.h.
Also set BraceWrapping.BeforeWhile to true to match the formatting
specifie
From: Matthew Malcomson
I've posted the patch on the relevant Bugzilla, but also sending to
mailing list. If should have only done one please do mention.
- 8< --- >8
When making warnings trigger a failure in template substitution I
could not find any way
From: Matthew Malcomson
According to the editorconfig file format description, a match against
one of multiple different strings is described with those different
strings separated by commas and within curly braces. E.g.
[{x,y}.txt]
https://editorconfig.org/, under "Wildcard Patterns".
The
From: Matthew Malcomson
For the `gcc` and `g++` tools we often pass -B/path/to/object/dir in via
`TEST_ALWAYS_FLAGS` (see e.g. asan.exp where this is set).
In libitm.c++/c++.exp we pass that -B flag via the `tool_flags` argument
to `dg-runtest`.
Passing as the `tool_flags` argument means that th
24 matches
Mail list logo
