OpenE2K/gcc
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Update documentation of AArch64 options for GCC6 to be more accurate,

Browse Source 
fix a few minor mistakes and remove some duplication.

        * doc/invoke.texi (AArch64 Options): Various updates.

From-SVN: r236290
This commit is contained in:
Wilco Dijkstra 2016-05-16 14:55:15 +00:00 committed by Wilco Dijkstra
parent 0bc8d0b3c1
commit 704b85c7d8
2 changed files with 36 additions and 40 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
ChangeLog
View File

@ -1,3 +1,7 @@
2016-05-16 Wilco Dijkstra <wdijkstr@arm.com>
* doc/invoke.texi (AArch64 Options): Various updates.
2016-05-09 Aaron Sawdey <acsawdey@linux.vnet.ibm.com> 2016-05-09 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
* MAINTAINERS (Write After Approval): Add myself. * MAINTAINERS (Write After Approval): Add myself.

72
gcc/doc/invoke.texi
View File

@ -12872,9 +12872,9 @@ These options are defined for AArch64 implementations:
@item -mabi=@var{name} @item -mabi=@var{name}
@opindex mabi @opindex mabi
Generate code for the specified data model. Permissible values Generate code for the specified data model. Permissible values
are @samp{ilp32} for SysV-like data model where int, long int and pointer are @samp{ilp32} for SysV-like data model where int, long int and pointers
are 32-bit, and @samp{lp64} for SysV-like data model where int is 32-bit, are 32 bits, and @samp{lp64} for SysV-like data model where int is 32 bits,
but long int and pointer are 64-bit. but long int and pointers are 64 bits.
The default depends on the specific target configuration. Note that The default depends on the specific target configuration. Note that
the LP64 and ILP32 ABIs are not link-compatible; you must compile your the LP64 and ILP32 ABIs are not link-compatible; you must compile your
@ -12899,25 +12899,24 @@ Generate little-endian code. This is the default when GCC is configured for an
@item -mcmodel=tiny @item -mcmodel=tiny
@opindex mcmodel=tiny @opindex mcmodel=tiny
Generate code for the tiny code model. The program and its statically defined Generate code for the tiny code model. The program and its statically defined
symbols must be within 1GB of each other. Pointers are 64 bits. Programs can symbols must be within 1MB of each other. Programs can be statically or
be statically or dynamically linked. This model is not fully implemented and dynamically linked.
mostly treated as @samp{small}.
@item -mcmodel=small @item -mcmodel=small
@opindex mcmodel=small @opindex mcmodel=small
Generate code for the small code model. The program and its statically defined Generate code for the small code model. The program and its statically defined
symbols must be within 4GB of each other. Pointers are 64 bits. Programs can symbols must be within 4GB of each other. Programs can be statically or
be statically or dynamically linked. This is the default code model. dynamically linked. This is the default code model.
@item -mcmodel=large @item -mcmodel=large
@opindex mcmodel=large @opindex mcmodel=large
Generate code for the large code model. This makes no assumptions about Generate code for the large code model. This makes no assumptions about
addresses and sizes of sections. Pointers are 64 bits. Programs can be addresses and sizes of sections. Programs can be statically linked only.
statically linked only.
@item -mstrict-align @item -mstrict-align
@opindex mstrict-align @opindex mstrict-align
Do not assume that unaligned memory references are handled by the system. Avoid generating memory accesses that may not be aligned on a natural object
boundary as described in the architecture specification.
@item -momit-leaf-frame-pointer @item -momit-leaf-frame-pointer
@itemx -mno-omit-leaf-frame-pointer @itemx -mno-omit-leaf-frame-pointer
@ -12939,7 +12938,7 @@ of TLS variables.
@item -mtls-size=@var{size} @item -mtls-size=@var{size}
@opindex mtls-size @opindex mtls-size
Specify bit size of immediate TLS offsets. Valid values are 12, 24, 32, 48. Specify bit size of immediate TLS offsets. Valid values are 12, 24, 32, 48.
This option depends on binutils higher than 2.25. This option requires binutils 2.26 or newer.
@item -mfix-cortex-a53-835769 @item -mfix-cortex-a53-835769
@itemx -mno-fix-cortex-a53-835769 @itemx -mno-fix-cortex-a53-835769
@ -12959,12 +12958,13 @@ corresponding flag to the linker.
@item -mlow-precision-recip-sqrt @item -mlow-precision-recip-sqrt
@item -mno-low-precision-recip-sqrt @item -mno-low-precision-recip-sqrt
@opindex -mlow-precision-recip-sqrt @opindex mlow-precision-recip-sqrt
@opindex -mno-low-precision-recip-sqrt @opindex mno-low-precision-recip-sqrt
When calculating the reciprocal square root approximation, Enable or disable reciprocal square root approximation.
uses one less step than otherwise, thus reducing latency and precision. This option only has an effect if @option{-ffast-math} or
This is only relevant if @option{-ffast-math} enables the reciprocal square root @option{-funsafe-math-optimizations} is used as well. Enabling this reduces
approximation, which in turn depends on the target processor. precision of reciprocal square root results to about 16 bits for
single precision and to 32 bits for double precision.
@item -march=@var{name} @item -march=@var{name}
@opindex march @opindex march
@ -13001,17 +13001,15 @@ Specify the name of the target processor for which GCC should tune the
performance of the code. Permissible values for this option are: performance of the code. Permissible values for this option are:
@samp{generic}, @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a57}, @samp{generic}, @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a57},
@samp{cortex-a72}, @samp{exynos-m1}, @samp{qdf24xx}, @samp{thunderx}, @samp{cortex-a72}, @samp{exynos-m1}, @samp{qdf24xx}, @samp{thunderx},
@samp{xgene1}. @samp{xgene1}, @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
@samp{native}.
Additionally, this option can specify that GCC should tune the performance The values @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53}
of the code for a big.LITTLE system. Permissible values for this specify that GCC should tune for a big.LITTLE system.
option are: @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53}.
Additionally on native AArch64 GNU/Linux systems the value Additionally on native AArch64 GNU/Linux systems the value
@samp{native} is available. This option causes the compiler to pick @samp{native} tunes performance to the host system. This option has no effect
the architecture of and tune the performance of the code for the if the compiler is unable to recognize the processor of the host system.
processor of the host system. This option has no effect if the
compiler is unable to recognize the architecture of the host system.
Where none of @option{-mtune=}, @option{-mcpu=} or @option{-march=} Where none of @option{-mtune=}, @option{-mcpu=} or @option{-march=}
are specified, the code is tuned to perform well across a range are specified, the code is tuned to perform well across a range
@ -13031,12 +13029,6 @@ documented in the sub-section on
Feature Modifiers}. Where conflicting feature modifiers are Feature Modifiers}. Where conflicting feature modifiers are
specified, the right-most feature is used. specified, the right-most feature is used.
Additionally on native AArch64 GNU/Linux systems the value
@samp{native} is available. This option causes the compiler to tune
the performance of the code for the processor of the host system.
This option has no effect if the compiler is unable to recognize the
architecture of the host system.
GCC uses @var{name} to determine what kind of instructions it can emit when GCC uses @var{name} to determine what kind of instructions it can emit when
generating assembly code (as if by @option{-march}) and to determine generating assembly code (as if by @option{-march}) and to determine
the target processor for which to tune for performance (as if the target processor for which to tune for performance (as if
@ -13054,11 +13046,11 @@ across releases.
This option is only intended to be useful when developing GCC. This option is only intended to be useful when developing GCC.
@item -mpc-relative-literal-loads @item -mpc-relative-literal-loads
@opindex mpcrelativeliteralloads @opindex mpc-relative-literal-loads
Enable PC relative literal loads. If this option is used, literal Enable PC-relative literal loads. With this option literal pools are
pools are assumed to have a range of up to 1MiB and an appropriate accessed using a single instruction and emitted after each function. This
instruction sequence is used. This option has no impact when used limits the maximum size of functions to 1MB. This is enabled by default for
with @option{-mcmodel=tiny}. @option{-mcmodel=tiny}.
@end table @end table
@ -13089,9 +13081,9 @@ Enable Large System Extension instructions. This is on by default for
@end table @end table
That is, @option{crypto} implies @option{simd} implies @option{fp}. Feature @option{crypto} implies @option{simd}, which implies @option{fp}.
Conversely, @option{nofp} (or equivalently, @option{-mgeneral-regs-only}) Conversely, @option{nofp} implies @option{nosimd}, which implies
implies @option{nosimd} implies @option{nocrypto}. @option{nocrypto}.
@node Adapteva Epiphany Options @node Adapteva Epiphany Options
@subsection Adapteva Epiphany Options @subsection Adapteva Epiphany Options