diff options
Diffstat (limited to 'ld/ld.texinfo')
| -rw-r--r-- | ld/ld.texinfo | 69 |
1 files changed, 44 insertions, 25 deletions
diff --git a/ld/ld.texinfo b/ld/ld.texinfo index 420065b..bfd3464 100644 --- a/ld/ld.texinfo +++ b/ld/ld.texinfo @@ -5552,34 +5552,53 @@ interleaving might place the @code{.literal} sections from an initial group of files followed by the @code{.text} sections of that group of files. Then, the @code{.literal} sections from the rest of the files and the @code{.text} sections from the rest of the files would follow. -The non-interleaved order can still be specified as: -@smallexample -SECTIONS -@{ - .text : @{ - *(.literal) *(.text) - @} -@} -@end smallexample - -@cindex @code{--relax} on Xtensa +@cindex @option{--relax} on Xtensa @cindex relaxing on Xtensa +Relaxation is enabled by default for the Xtensa version of @command{ld} and +provides two important link-time optimizations. The first optimization +is to combine identical literal values to reduce code size. A redundant +literal will be removed and all the @code{L32R} instructions that use it +will be changed to reference an identical literal, as long as the +location of the replacement literal is within the offset range of all +the @code{L32R} instructions. The second optimization is to remove +unnecessary overhead from assembler-generated ``longcall'' sequences of +@code{L32R}/@code{CALLX@var{n}} when the target functions are within +range of direct @code{CALL@var{n}} instructions. + +For each of these cases where an indirect call sequence can be optimized +to a direct call, the linker will change the @code{CALLX@var{n}} +instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R} +instruction, and remove the literal referenced by the @code{L32R} +instruction if it is not used for anything else. Removing the +@code{L32R} instruction always reduces code size but can potentially +hurt performance by changing the alignment of subsequent branch targets. +By default, the linker will always preserve alignments, either by +switching some instructions between 24-bit encodings and the equivalent +density instructions or by inserting a no-op in place of the @code{L32R} +instruction that was removed. If code size is more important than +performance, the @option{--size-opt} option can be used to prevent the +linker from widening density instructions or inserting no-ops, except in +a few cases where no-ops are required for correctness. + +The following Xtensa-specific command-line options can be used to +control the linker: + +@cindex Xtensa options +@table @option @kindex --no-relax -The Xtensa version of @command{ld} enables the @option{--relax} option by -default to attempt to reduce space in the output image by combining -literals with identical values. It also provides the -@option{--no-relax} option to disable this optimization. When enabled, -the relaxation algorithm ensures that a literal will only be merged with -another literal when the new merged literal location is within the -offset range of all of its uses. - -The relaxation mechanism will also attempt to optimize -assembler-generated ``longcall'' sequences of -@code{L32R}/@code{CALLX@var{n}} when the target is known to fit into a -@code{CALL@var{n}} instruction encoding. The current optimization -converts the sequence into @code{NOP}/@code{CALL@var{n}} and removes the -literal referenced by the @code{L32R} instruction. +@item --no-relax +Since the Xtensa version of @code{ld} enables the @option{--relax} option +by default, the @option{--no-relax} option is provided to disable +relaxation. + +@item --size-opt +When optimizing indirect calls to direct calls, optimize for code size +more than performance. With this option, the linker will not insert +no-ops or widen density instructions to preserve branch target +alignment. There may still be some cases where no-ops are required to +preserve the correctness of the code. +@end table @ifclear GENERIC @lowersections |