This patch adds documentation for flto-toplevel-asm-heuristics
and mentions it in Remarks in Basic Asm.
gcc/ChangeLog:
* common.opt.urls: Regenerate.
* doc/extend.texi: Mention flto-toplevel-asm-heuristics.
* doc/invoke.texi: Add flto-toplevel-asm-heuristics.
---
gcc/common.opt.urls | 3 +++
gcc/doc/extend.texi | 5 +++++
gcc/doc/invoke.texi | 22 +++++++++++++++++++++-
3 files changed, 29 insertions(+), 1 deletion(-)
diff --git a/gcc/common.opt.urls b/gcc/common.opt.urls
index 6c094776187..93e730f2f5d 100644
--- a/gcc/common.opt.urls
+++ b/gcc/common.opt.urls
@@ -1044,6 +1044,9 @@
UrlSuffix(gcc/Optimize-Options.html#index-flto-incremental-cache-size)
flto-partition=
UrlSuffix(gcc/Optimize-Options.html#index-flto-partition)
+flto-toplevel-asm-heuristics
+UrlSuffix(gcc/Optimize-Options.html#index-flto-toplevel-asm-heuristics)
+
flto-compression-level=
UrlSuffix(gcc/Optimize-Options.html#index-flto-compression-level)
diff --git a/gcc/doc/extend.texi b/gcc/doc/extend.texi
index ac86a2e1484..5e481126069 100644
--- a/gcc/doc/extend.texi
+++ b/gcc/doc/extend.texi
@@ -11507,6 +11507,11 @@ with extended @code{asm}. See
from basic asm to extended asm} for information about how to perform this
conversion.
+LTO typically requires rewriting top-level basic @code{asm} statements to
+extended @code{asm}. Otherwise you will likely encounter missing symbol
+errors by linker. Alternatively you may use
+@option{-flto-toplevel-asm-heuristics}.
+
The compiler copies the assembler instructions in a basic @code{asm}
verbatim to the assembly language output file, without
processing dialects or any of the @samp{%} operators that are available with
diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi
index d690305a890..6ea59a2ea57 100644
--- a/gcc/doc/invoke.texi
+++ b/gcc/doc/invoke.texi
@@ -646,7 +646,8 @@ Objective-C and Objective-C++ Dialects}.
-floop-unroll-and-jam -floop-nest-optimize
-floop-parallelize-all -flra-remat -flto -flto-compression-level
-flto-partition=@var{alg} -flto-incremental=@var{path}
--flto-incremental-cache-size=@var{n} -fmalloc-dce -fmerge-all-constants
+-flto-incremental-cache-size=@var{n} -flto-toplevel-asm-heuristics
+-fmalloc-dce -fmerge-all-constants
-fmerge-constants -fmodulo-sched -fmodulo-sched-allow-regmoves
-fmove-loop-invariants -fmove-loop-stores -fno-branch-count-reg
-fno-defer-pop -fno-function-cse
@@ -15811,6 +15812,25 @@ Multiple GCC instances can use the same cache in
parallel.
Specifies number of cache entries in incremental LTO after which to prune
old entries. This is a soft limit, temporarily there may be more entries.
+@opindex flto-toplevel-asm-heuristics
+@item -flto-toplevel-asm-heuristics
+Enables heuristics to find symbols used in top-level basic @code{asm}.
+This will restrict link-time optimizations that could cause renaming
+or deletion of such symbols which would result in missing symbol errors by
+linker.
+
+This flag is intended for projects that have not converted to using top-level
+extended @code{asm} (@pxref{Extended Asm}), which specify the usage directly
+without any false positives.
+
+The heuristics are simple and do not parse the assembly.
+The heuristics scan through top-level assembly for all possible identifiers;
+if an identifier is found among declared symbols, the symbol will be marked to
+restrict link-time optimizations. Static symbols disable more optimizations.
+Identifiers followed by ':' disable more optimizations as well, because they
+might be a locally defined symbol in assembly, even when the declaration
+is marked 'extern'.
+
@opindex flto-compression-level
@item -flto-compression-level=@var{n}
This option specifies the level of compression used for intermediate
--
2.52.0
