1..
2  -------------------------------------------------------------------
3  NOTE: This file is automatically generated by running clang-tblgen
4  -gen-opt-docs. Do not edit this file by hand!!
5  -------------------------------------------------------------------
6
7=====================================
8Clang command line argument reference
9=====================================
10.. contents::
11   :local:
12
13Introduction
14============
15
16This page lists the command line arguments currently supported by the
17GCC-compatible ``clang`` and ``clang++`` drivers.
18
19
20.. program:: clang
21.. option:: -B<prefix>, --prefix <arg>, --prefix=<arg>
22
23Search $prefix$file for executables, libraries, and data files. If $prefix is a directory, search $prefix/$file
24
25.. option:: -F<arg>
26
27Add directory to framework include search path
28
29.. option:: -ObjC
30
31Treat source input files as Objective-C inputs
32
33.. program:: clang1
34.. option:: -ObjC++
35.. program:: clang
36
37Treat source input files as Objective-C++ inputs
38
39.. option:: -Qn, -fno-ident
40
41Do not emit metadata containing compiler name and version
42
43.. option:: -Qunused-arguments
44
45Don't emit warning for unused driver arguments
46
47.. option:: -Qy, -fident
48
49Emit metadata containing compiler name and version
50
51.. option:: -Wa,<arg>,<arg2>...
52
53Pass the comma separated arguments in <arg> to the assembler
54
55.. option:: -Wlarge-by-value-copy=<arg>
56
57.. option:: -Xarch\_<arg1> <arg2>
58
59.. program:: clang1
60.. option:: -Xarch\_device <arg>
61.. program:: clang
62
63Pass <arg> to the CUDA/HIP device compilation
64
65.. program:: clang2
66.. option:: -Xarch\_host <arg>
67.. program:: clang
68
69Pass <arg> to the CUDA/HIP host compilation
70
71.. option:: -Xcuda-fatbinary <arg>
72
73Pass <arg> to fatbinary invocation
74
75.. option:: -Xcuda-ptxas <arg>
76
77Pass <arg> to the ptxas assembler
78
79.. option:: -Xopenmp-target <arg>
80
81Pass <arg> to the target offloading toolchain.
82
83.. program:: clang1
84.. option:: -Xopenmp-target=<triple> <arg>
85.. program:: clang
86
87Pass <arg> to the target offloading toolchain identified by <triple>.
88
89.. option:: -Z<arg>
90
91.. option:: -a<arg>, --profile-blocks
92
93.. option:: -all\_load
94
95.. option:: -allowable\_client <arg>
96
97.. option:: --analyze
98
99Run the static analyzer
100
101.. option:: --analyzer-no-default-checks
102
103.. option:: --analyzer-output<arg>
104
105Static analyzer report output format (html\|plist\|plist-multi-file\|plist-html\|sarif\|sarif-html\|text).
106
107.. option:: -arch <arg>
108
109.. program:: clang1
110.. option:: -arch\_errors\_fatal
111.. program:: clang
112
113.. program:: clang2
114.. option:: -arch\_only <arg>
115.. program:: clang
116
117.. option:: -arcmt-migrate-emit-errors
118
119Emit ARC errors even if the migrator can fix them
120
121.. option:: -arcmt-migrate-report-output <arg>
122
123Output path for the plist report
124
125.. option:: --autocomplete=<arg>
126
127.. option:: -bind\_at\_load
128
129.. option:: -bundle
130
131.. program:: clang1
132.. option:: -bundle\_loader <arg>
133.. program:: clang
134
135.. option:: -client\_name<arg>
136
137.. option:: -compatibility\_version<arg>
138
139.. option:: --config <arg>
140
141Specifies configuration file
142
143.. option:: --constant-cfstrings
144
145.. option:: --cuda-compile-host-device
146
147Compile CUDA code for both host and device (default).  Has no effect on non-CUDA compilations.
148
149.. option:: --cuda-device-only
150
151Compile CUDA code for device only
152
153.. option:: --cuda-host-only
154
155Compile CUDA code for host only.  Has no effect on non-CUDA compilations.
156
157.. option:: --cuda-include-ptx=<arg>, --no-cuda-include-ptx=<arg>
158
159Include PTX for the following GPU architecture (e.g. sm\_35) or 'all'. May be specified more than once.
160
161.. option:: --cuda-noopt-device-debug, --no-cuda-noopt-device-debug
162
163Enable device-side debug info generation. Disables ptxas optimizations.
164
165.. option:: -cuid=<arg>
166
167An ID for compilation unit, which should be the same for the same compilation unit but different for different compilation units. It is used to externalize device-side static variables for single source offloading languages CUDA and HIP so that they can be accessed by the host code of the same compilation unit.
168
169.. option:: -current\_version<arg>
170
171.. option:: -dead\_strip
172
173.. option:: -dependency-dot <arg>
174
175Filename to write DOT-formatted header dependencies to
176
177.. option:: -dependency-file <arg>
178
179Filename (or -) to write dependency output to
180
181.. option:: -dsym-dir<dir>
182
183Directory to output dSYM's (if any) to
184
185.. option:: -dumpmachine
186
187.. option:: -dumpversion
188
189.. option:: --dyld-prefix=<arg>, --dyld-prefix <arg>
190
191.. option:: -dylib\_file <arg>
192
193.. option:: -dylinker
194
195.. program:: clang1
196.. option:: -dylinker\_install\_name<arg>
197.. program:: clang
198
199.. option:: -dynamic
200
201.. option:: -dynamiclib
202
203.. option:: -emit-ast
204
205Emit Clang AST files for source inputs
206
207.. option:: --emit-static-lib
208
209Enable linker job to emit a static library.
210
211.. option:: -enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
212
213Trivial automatic variable initialization to zero is only here for benchmarks, it'll eventually be removed, and I'm OK with that because I'm only using it to benchmark
214
215.. option:: --end-no-unused-arguments
216
217Start emitting warnings for unused driver arguments
218
219.. option:: -exported\_symbols\_list <arg>
220
221.. option:: -faligned-new=<arg>
222
223.. option:: -fautomatic
224
225.. option:: -ffixed-r19
226
227Reserve register r19 (Hexagon only)
228
229.. option:: -fgpu-flush-denormals-to-zero, -fcuda-flush-denormals-to-zero, -fno-gpu-flush-denormals-to-zero
230
231Flush denormal floating point values to zero in CUDA/HIP device mode.
232
233.. option:: -fheinous-gnu-extensions
234
235.. option:: -flat\_namespace
236
237.. option:: -fopenmp-targets=<arg1>,<arg2>...
238
239Specify comma-separated list of triples OpenMP offloading targets to be supported
240
241.. option:: -force\_cpusubtype\_ALL
242
243.. program:: clang1
244.. option:: -force\_flat\_namespace
245.. program:: clang
246
247.. program:: clang2
248.. option:: -force\_load <arg>
249.. program:: clang
250
251.. option:: -fplugin-arg-<name>-<arg>
252
253Pass <arg> to plugin <name>
254
255.. option:: -framework <arg>
256
257.. option:: -frtlib-add-rpath, -fno-rtlib-add-rpath
258
259Add -rpath with architecture-specific resource directory to the linker flags
260
261.. option:: -fsanitize-system-ignorelist=<arg>, -fsanitize-system-blacklist=<arg>
262
263Path to system ignorelist file for sanitizers
264
265.. option:: -fshow-skipped-includes
266
267#include files may be "skipped" due to include guard optimization
268             or #pragma once. This flag makes -H show also such includes.
269
270.. option:: -fsystem-module
271
272Build this module as a system module. Only used with -emit-module
273
274.. option:: -fuse-cuid=<arg>
275
276Method to generate ID's for compilation units for single source offloading languages CUDA and HIP: 'hash' (ID's generated by hashing file path and command line options) \| 'random' (ID's generated as random numbers) \| 'none' (disabled). Default is 'hash'. This option will be overridden by option '-cuid=\[ID\]' if it is specified.
277
278.. option:: --gcc-toolchain=<arg>
279
280Search for GCC installation in the specified directory on targets which commonly use GCC. The directory usually contains 'lib{,32,64}/gcc{,-cross}/$triple' and 'include'. If specified, sysroot is skipped for GCC detection. Note: executables (e.g. ld) used by the compiler are not overridden by the selected GCC installation
281
282.. option:: -gcodeview
283
284Generate CodeView debug information
285
286.. option:: -gcodeview-ghash, -gno-codeview-ghash
287
288Emit type record hashes in a .debug$H section
289
290.. option:: --gpu-instrument-lib=<arg>
291
292Instrument device library for HIP, which is a LLVM bitcode containing \_\_cyg\_profile\_func\_enter and \_\_cyg\_profile\_func\_exit
293
294.. option:: --gpu-max-threads-per-block=<arg>
295
296Default max threads per block for kernel launch bounds for HIP
297
298.. option:: -headerpad\_max\_install\_names<arg>
299
300.. option:: -help, --help
301
302Display available options
303
304.. option:: --help-hidden
305
306Display help for hidden options
307
308.. option:: --hip-link
309
310Link clang-offload-bundler bundles for HIP
311
312.. option:: --hip-version=<arg>
313
314HIP version in the format of major.minor.patch
315
316.. option:: -ibuiltininc
317
318Enable builtin #include directories even when -nostdinc is used before or after -ibuiltininc. Using -nobuiltininc after the option disables it
319
320.. option:: -image\_base <arg>
321
322.. option:: -index-header-map
323
324Make the next included directory (-I or -F) an indexer header map
325
326.. option:: -init <arg>
327
328.. option:: -install\_name <arg>
329
330.. option:: -interface-stub-version=<arg>
331
332.. option:: -keep\_private\_externs
333
334.. option:: -lazy\_framework <arg>
335
336.. program:: clang1
337.. option:: -lazy\_library <arg>
338.. program:: clang
339
340.. option:: -mbig-endian, -EB
341
342.. option:: -mbranch-protection=<arg>
343
344Enforce targets of indirect branches and function returns
345
346.. option:: -menable-unsafe-fp-math
347
348Allow unsafe floating-point math optimizations which may decrease precision
349
350.. option:: -mharden-sls=<arg>
351
352Select straight-line speculation hardening scope
353
354.. option:: --migrate
355
356Run the migrator
357
358.. option:: -mios-simulator-version-min=<arg>, -miphonesimulator-version-min=<arg>
359
360.. option:: -mlinker-version=<arg>
361
362.. option:: -mlittle-endian, -EL
363
364.. option:: -mllvm <arg>
365
366Additional arguments to forward to LLVM's option processing
367
368.. option:: -module-dependency-dir <arg>
369
370Directory to dump module dependencies to
371
372.. option:: -mtvos-simulator-version-min=<arg>, -mappletvsimulator-version-min=<arg>
373
374.. option:: -multi\_module
375
376.. option:: -multiply\_defined <arg>
377
378.. program:: clang1
379.. option:: -multiply\_defined\_unused <arg>
380.. program:: clang
381
382.. option:: -mwatchos-simulator-version-min=<arg>, -mwatchsimulator-version-min=<arg>
383
384.. option:: --no-cuda-version-check
385
386Don't error out if the detected version of the CUDA install is too low for the requested CUDA gpu architecture.
387
388.. option:: -no-integrated-cpp, --no-integrated-cpp
389
390.. option:: -no\_dead\_strip\_inits\_and\_terms
391
392.. option:: -nobuiltininc
393
394Disable builtin #include directories
395
396.. option:: -nodefaultlibs
397
398.. option:: -nofixprebinding
399
400.. option:: -nogpuinc, -nocudainc
401
402Do not add include paths for CUDA/HIP and do not include the default CUDA/HIP wrapper headers
403
404.. option:: -nogpulib, -nocudalib
405
406Do not link device library for CUDA/HIP device compilation
407
408.. option:: -nohipwrapperinc
409
410Do not include the default HIP wrapper headers and include paths
411
412.. option:: -nolibc
413
414.. option:: -nomultidefs
415
416.. option:: -nopie, -no-pie
417
418.. option:: -noprebind
419
420.. option:: -noprofilelib
421
422.. option:: -noseglinkedit
423
424.. option:: -nostdinc, --no-standard-includes
425
426.. program:: clang1
427.. option:: -nostdinc++
428.. program:: clang
429
430Disable standard #include directories for the C++ standard library
431
432.. option:: -nostdlib++
433
434.. option:: -nostdlibinc
435
436.. option:: -o<file>, --output <arg>, --output=<arg>
437
438Write output to <file>
439
440.. option:: -objcmt-allowlist-dir-path=<arg>, -objcmt-white-list-dir-path=<arg>, -objcmt-whitelist-dir-path=<arg>
441
442Only modify files with a filename contained in the provided directory path
443
444.. option:: -objcmt-atomic-property
445
446Make migration to 'atomic' properties
447
448.. option:: -objcmt-migrate-all
449
450Enable migration to modern ObjC
451
452.. option:: -objcmt-migrate-annotation
453
454Enable migration to property and method annotations
455
456.. option:: -objcmt-migrate-designated-init
457
458Enable migration to infer NS\_DESIGNATED\_INITIALIZER for initializer methods
459
460.. option:: -objcmt-migrate-instancetype
461
462Enable migration to infer instancetype for method result type
463
464.. option:: -objcmt-migrate-literals
465
466Enable migration to modern ObjC literals
467
468.. option:: -objcmt-migrate-ns-macros
469
470Enable migration to NS\_ENUM/NS\_OPTIONS macros
471
472.. option:: -objcmt-migrate-property
473
474Enable migration to modern ObjC property
475
476.. option:: -objcmt-migrate-property-dot-syntax
477
478Enable migration of setter/getter messages to property-dot syntax
479
480.. option:: -objcmt-migrate-protocol-conformance
481
482Enable migration to add protocol conformance on classes
483
484.. option:: -objcmt-migrate-readonly-property
485
486Enable migration to modern ObjC readonly property
487
488.. option:: -objcmt-migrate-readwrite-property
489
490Enable migration to modern ObjC readwrite property
491
492.. option:: -objcmt-migrate-subscripting
493
494Enable migration to modern ObjC subscripting
495
496.. option:: -objcmt-ns-nonatomic-iosonly
497
498Enable migration to use NS\_NONATOMIC\_IOSONLY macro for setting property's 'atomic' attribute
499
500.. option:: -objcmt-returns-innerpointer-property
501
502Enable migration to annotate property with NS\_RETURNS\_INNER\_POINTER
503
504.. option:: -object
505
506.. option:: -object-file-name=<file>, -object-file-name <arg>
507
508Set the output <file> for debug infos
509
510.. option:: --offload-arch=<arg>, --cuda-gpu-arch=<arg>, --no-offload-arch=<arg>
511
512CUDA offloading device architecture (e.g. sm\_35), or HIP offloading target ID in the form of a device architecture followed by target ID features delimited by a colon. Each target ID feature is a pre-defined string followed by a plus or minus sign (e.g. gfx908:xnack+:sramecc-).  May be specified more than once.
513
514.. option:: --offload=<arg1>,<arg2>...
515
516Specify comma-separated list of offloading target triples (CUDA and HIP only)
517
518.. option:: -p, --profile
519
520.. option:: -pagezero\_size<arg>
521
522.. option:: -pg
523
524Enable mcount instrumentation
525
526.. option:: -pipe, --pipe
527
528Use pipes between commands, when possible
529
530.. option:: -prebind
531
532.. program:: clang1
533.. option:: -prebind\_all\_twolevel\_modules
534.. program:: clang
535
536.. option:: -preload
537
538.. option:: --print-diagnostic-categories
539
540.. option:: -print-effective-triple, --print-effective-triple
541
542Print the effective target triple
543
544.. option:: -print-file-name=<file>, --print-file-name=<file>, --print-file-name <arg>
545
546Print the full library path of <file>
547
548.. option:: -print-ivar-layout
549
550Enable Objective-C Ivar layout bitmap print trace
551
552.. option:: -print-libgcc-file-name, --print-libgcc-file-name
553
554Print the library path for the currently used compiler runtime library ("libgcc.a" or "libclang\_rt.builtins.\*.a")
555
556.. option:: -print-multi-directory, --print-multi-directory
557
558.. option:: -print-multi-lib, --print-multi-lib
559
560.. option:: -print-multiarch, --print-multiarch
561
562Print the multiarch target triple
563
564.. option:: -print-prog-name=<name>, --print-prog-name=<name>, --print-prog-name <arg>
565
566Print the full program path of <name>
567
568.. option:: -print-resource-dir, --print-resource-dir
569
570Print the resource directory pathname
571
572.. option:: -print-rocm-search-dirs, --print-rocm-search-dirs
573
574Print the paths used for finding ROCm installation
575
576.. option:: -print-runtime-dir, --print-runtime-dir
577
578Print the directory pathname containing clangs runtime libraries
579
580.. option:: -print-search-dirs, --print-search-dirs
581
582Print the paths used for finding libraries and programs
583
584.. option:: -print-target-triple, --print-target-triple
585
586Print the normalized target triple
587
588.. option:: -print-targets, --print-targets
589
590Print the registered targets
591
592.. option:: -private\_bundle
593
594.. option:: -pthread, -no-pthread
595
596Support POSIX threads in generated code
597
598.. option:: -pthreads
599
600.. option:: -read\_only\_relocs <arg>
601
602.. option:: -relocatable-pch, --relocatable-pch
603
604Whether to build a relocatable precompiled header
605
606.. option:: -remap
607
608.. option:: -rewrite-legacy-objc
609
610Rewrite Legacy Objective-C source to C++
611
612.. option:: -rtlib=<arg>, --rtlib=<arg>, --rtlib <arg>
613
614Compiler runtime library to use
615
616.. option:: -save-stats=<arg>, --save-stats=<arg>, -save-stats (equivalent to -save-stats=cwd), --save-stats (equivalent to -save-stats=cwd)
617
618Save llvm statistics.
619
620.. option:: -save-temps=<arg>, --save-temps=<arg>, -save-temps (equivalent to -save-temps=cwd), --save-temps (equivalent to -save-temps=cwd)
621
622Save intermediate compilation results.
623
624.. option:: -sectalign <arg1> <arg2> <arg3>
625
626.. option:: -sectcreate <arg1> <arg2> <arg3>
627
628.. option:: -sectobjectsymbols <arg1> <arg2>
629
630.. option:: -sectorder <arg1> <arg2> <arg3>
631
632.. option:: -seg1addr<arg>
633
634.. option:: -seg\_addr\_table <arg>
635
636.. program:: clang1
637.. option:: -seg\_addr\_table\_filename <arg>
638.. program:: clang
639
640.. option:: -segaddr <arg1> <arg2>
641
642.. option:: -segcreate <arg1> <arg2> <arg3>
643
644.. option:: -seglinkedit
645
646.. option:: -segprot <arg1> <arg2> <arg3>
647
648.. option:: -segs\_read\_<arg>
649
650.. program:: clang1
651.. option:: -segs\_read\_only\_addr <arg>
652.. program:: clang
653
654.. program:: clang2
655.. option:: -segs\_read\_write\_addr <arg>
656.. program:: clang
657
658.. option:: -serialize-diagnostics <arg>, --serialize-diagnostics <arg>
659
660Serialize compiler diagnostics to a file
661
662.. option:: -shared-libgcc
663
664.. option:: -shared-libsan, -shared-libasan
665
666Dynamically link the sanitizer runtime
667
668.. option:: -single\_module
669
670.. option:: --start-no-unused-arguments
671
672Don't emit warnings about unused arguments for the following arguments
673
674.. option:: -static-libgcc
675
676.. option:: -static-libsan
677
678Statically link the sanitizer runtime
679
680.. option:: -static-libstdc++
681
682.. option:: -static-openmp
683
684Use the static host OpenMP runtime while linking.
685
686.. option:: -std-default=<arg>
687
688.. option:: -stdlib=<arg>, --stdlib=<arg>, --stdlib <arg>
689
690C++ standard library to use
691
692.. option:: -sub\_library<arg>
693
694.. program:: clang1
695.. option:: -sub\_umbrella<arg>
696.. program:: clang
697
698.. option:: --sysroot=<arg>, --sysroot <arg>
699
700.. option:: --target-help
701
702.. option:: --target=<arg>, -target <arg>
703
704Generate code for the given target
705
706.. option:: -time
707
708Time individual commands
709
710.. option:: -traditional, --traditional
711
712.. option:: -traditional-cpp, --traditional-cpp
713
714Enable some traditional CPP emulation
715
716.. option:: -twolevel\_namespace
717
718.. program:: clang1
719.. option:: -twolevel\_namespace\_hints
720.. program:: clang
721
722.. option:: -umbrella <arg>
723
724.. option:: -unexported\_symbols\_list <arg>
725
726.. option:: -unwindlib=<arg>, --unwindlib=<arg>
727
728Unwind library to use
729
730.. option:: -v, --verbose
731
732Show commands to run and use verbose output
733
734.. option:: --verify-debug-info
735
736Verify the binary representation of debug output
737
738.. option:: --version
739
740Print version information
741
742.. option:: -w, --no-warnings
743
744Suppress all warnings
745
746.. option:: -weak-l<arg>
747
748.. option:: -weak\_framework <arg>
749
750.. program:: clang1
751.. option:: -weak\_library <arg>
752.. program:: clang
753
754.. program:: clang2
755.. option:: -weak\_reference\_mismatches <arg>
756.. program:: clang
757
758.. option:: -whatsloaded
759
760.. option:: -why\_load, -whyload
761
762.. option:: -working-directory<arg>, -working-directory=<arg>
763
764Resolve file paths relative to the specified directory
765
766.. option:: -x<language>, --language <arg>, --language=<arg>
767
768Treat subsequent input files as having type <language>
769
770.. option:: -y<arg>
771
772Actions
773=======
774The action to perform on the input.
775
776.. option:: -E, --preprocess
777
778Only run the preprocessor
779
780.. option:: -S, --assemble
781
782Only run preprocess and compilation steps
783
784.. option:: -c, --compile
785
786Only run preprocess, compile, and assemble steps
787
788.. option:: -emit-interface-stubs
789
790Generate Interface Stub Files.
791
792.. option:: -emit-llvm
793
794Use the LLVM representation for assembler and object files
795
796.. option:: -emit-merged-ifs
797
798Generate Interface Stub Files, emit merged text not binary.
799
800.. option:: -extract-api
801
802Extract API information
803
804.. option:: -fopenmp-new-driver
805
806Use the new driver for OpenMP offloading.
807
808.. option:: -fsyntax-only
809
810.. option:: -module-file-info
811
812Provide information about a particular module file
813
814.. option:: --precompile
815
816Only precompile the input
817
818.. option:: -rewrite-objc
819
820Rewrite Objective-C source to C++
821
822.. option:: -verify-pch
823
824Load and verify that a pre-compiled header file is not stale
825
826Compilation flags
827=================
828
829Flags controlling the behavior of Clang during compilation. These flags have
830no effect during actions that do not perform compilation.
831
832.. option:: -Xassembler <arg>
833
834Pass <arg> to the assembler
835
836.. option:: -Xclang <arg>
837
838Pass <arg> to the clang compiler
839
840.. option:: -ansi, --ansi
841
842.. option:: -fc++-abi=<arg>
843
844C++ ABI to use. This will override the target C++ ABI.
845
846.. option:: -fclang-abi-compat=<version>
847
848Attempt to match the ABI of Clang <version>
849
850.. option:: -fcomment-block-commands=<arg>,<arg2>...
851
852Treat each comma separated argument in <arg> as a documentation comment block command
853
854.. option:: -fcomplete-member-pointers, -fno-complete-member-pointers
855
856Require member pointer base types to be complete if they would be significant under the Microsoft ABI
857
858.. option:: -fcrash-diagnostics-dir=<dir>
859
860Put crash-report files in <dir>
861
862.. option:: -fdeclspec, -fno-declspec
863
864Allow \_\_declspec as a keyword
865
866.. option:: -fdepfile-entry=<arg>
867
868.. option:: -fdiagnostics-fixit-info, -fno-diagnostics-fixit-info
869
870.. option:: -fdiagnostics-format=<arg>
871
872.. option:: -fdiagnostics-parseable-fixits
873
874Print fix-its in machine parseable form
875
876.. option:: -fdiagnostics-print-source-range-info
877
878Print source range spans in numeric form
879
880.. option:: -fdiagnostics-show-category=<arg>
881
882.. option:: -fdiscard-value-names, -fno-discard-value-names
883
884Discard value names in LLVM IR
885
886.. option:: -fexperimental-relative-c++-abi-vtables, -fno-experimental-relative-c++-abi-vtables
887
888Use the experimental C++ class ABI for classes with virtual tables
889
890.. option:: -fexperimental-strict-floating-point
891
892Enables experimental strict floating point in LLVM.
893
894.. option:: -ffine-grained-bitfield-accesses, -fno-fine-grained-bitfield-accesses
895
896Use separate accesses for consecutive bitfield runs with legal widths and alignments.
897
898.. option:: -fglobal-isel, -fexperimental-isel, -fno-global-isel
899
900Enables the global instruction selector
901
902.. option:: -finline-functions, -fno-inline-functions
903
904Inline suitable functions
905
906.. option:: -finline-hint-functions
907
908Inline functions which are (explicitly or implicitly) marked inline
909
910.. option:: -flegacy-pass-manager, -fno-legacy-pass-manager
911
912Use the legacy pass manager in LLVM (deprecated, to be removed in a future release)
913
914.. option:: -fno-crash-diagnostics
915
916Disable auto-generation of preprocessed source files and a script for reproduction during a clang crash
917
918.. option:: -fno-sanitize-ignorelist, -fno-sanitize-blacklist
919
920Don't use ignorelist file for sanitizers
921
922.. option:: -fparse-all-comments
923
924.. option:: -frecord-command-line, -fno-record-command-line, -frecord-gcc-switches
925
926.. option:: -fsanitize-address-destructor=<arg>
927
928Set destructor type used in ASan instrumentation
929
930.. option:: -fsanitize-address-field-padding=<arg>
931
932Level of field padding for AddressSanitizer
933
934.. option:: -fsanitize-address-globals-dead-stripping
935
936Enable linker dead stripping of globals in AddressSanitizer
937
938.. option:: -fsanitize-address-outline-instrumentation, -fno-sanitize-address-outline-instrumentation
939
940Always generate function calls for address sanitizer instrumentation
941
942.. option:: -fsanitize-address-poison-custom-array-cookie, -fno-sanitize-address-poison-custom-array-cookie
943
944Enable poisoning array cookies when using custom operator new\[\] in AddressSanitizer
945
946.. option:: -fsanitize-address-use-after-return=<mode>
947
948Select the mode of detecting stack use-after-return in AddressSanitizer: never \| runtime (default) \| always
949
950.. option:: -fsanitize-address-use-after-scope, -fno-sanitize-address-use-after-scope
951
952Enable use-after-scope detection in AddressSanitizer
953
954.. option:: -fsanitize-address-use-odr-indicator, -fno-sanitize-address-use-odr-indicator
955
956Enable ODR indicator globals to avoid false ODR violation reports in partially sanitized programs at the cost of an increase in binary size
957
958.. option:: -fsanitize-cfi-canonical-jump-tables, -fno-sanitize-cfi-canonical-jump-tables
959
960Make the jump table addresses canonical in the symbol table
961
962.. option:: -fsanitize-cfi-cross-dso, -fno-sanitize-cfi-cross-dso
963
964Enable control flow integrity (CFI) checks for cross-DSO calls.
965
966.. option:: -fsanitize-cfi-icall-generalize-pointers
967
968Generalize pointers in CFI indirect call type signature checks
969
970.. option:: -fsanitize-coverage-allowlist=<arg>, -fsanitize-coverage-whitelist=<arg>
971
972Restrict sanitizer coverage instrumentation exclusively to modules and functions that match the provided special case list, except the blocked ones
973
974.. option:: -fsanitize-coverage-ignorelist=<arg>, -fsanitize-coverage-blacklist=<arg>
975
976Disable sanitizer coverage instrumentation for modules and functions that match the provided special case list, even the allowed ones
977
978.. option:: -fsanitize-coverage=<arg1>,<arg2>..., -fno-sanitize-coverage=<arg1>,<arg2>...
979
980Specify the type of coverage instrumentation for Sanitizers
981
982.. option:: -fsanitize-hwaddress-abi=<arg>
983
984Select the HWAddressSanitizer ABI to target (interceptor or platform, default interceptor). This option is currently unused.
985
986.. option:: -fsanitize-hwaddress-experimental-aliasing, -fno-sanitize-hwaddress-experimental-aliasing
987
988Enable aliasing mode in HWAddressSanitizer
989
990.. option:: -fsanitize-ignorelist=<arg>, -fsanitize-blacklist=<arg>
991
992Path to ignorelist file for sanitizers
993
994.. option:: -fsanitize-link-c++-runtime, -fno-sanitize-link-c++-runtime
995
996.. option:: -fsanitize-link-runtime, -fno-sanitize-link-runtime
997
998.. option:: -fsanitize-memory-track-origins, -fno-sanitize-memory-track-origins
999
1000Enable origins tracking in MemorySanitizer
1001
1002.. program:: clang1
1003.. option:: -fsanitize-memory-track-origins=<arg>
1004.. program:: clang
1005
1006Enable origins tracking in MemorySanitizer
1007
1008.. option:: -fsanitize-memory-use-after-dtor, -fno-sanitize-memory-use-after-dtor
1009
1010Enable use-after-destroy detection in MemorySanitizer
1011
1012.. option:: -fsanitize-minimal-runtime, -fno-sanitize-minimal-runtime
1013
1014.. option:: -fsanitize-recover=<arg1>,<arg2>..., -fno-sanitize-recover=<arg1>,<arg2>..., -fsanitize-recover (equivalent to -fsanitize-recover=all)
1015
1016Enable recovery for specified sanitizers
1017
1018.. option:: -fsanitize-stats, -fno-sanitize-stats
1019
1020Enable sanitizer statistics gathering.
1021
1022.. option:: -fsanitize-thread-atomics, -fno-sanitize-thread-atomics
1023
1024Enable atomic operations instrumentation in ThreadSanitizer (default)
1025
1026.. option:: -fsanitize-thread-func-entry-exit, -fno-sanitize-thread-func-entry-exit
1027
1028Enable function entry/exit instrumentation in ThreadSanitizer (default)
1029
1030.. option:: -fsanitize-thread-memory-access, -fno-sanitize-thread-memory-access
1031
1032Enable memory access instrumentation in ThreadSanitizer (default)
1033
1034.. option:: -fsanitize-trap=<arg1>,<arg2>..., -fno-sanitize-trap=<arg1>,<arg2>..., -fsanitize-trap (equivalent to -fsanitize-trap=all), -fsanitize-undefined-trap-on-error (equivalent to -fsanitize-trap=undefined)
1035
1036Enable trapping for specified sanitizers
1037
1038.. option:: -fsanitize-undefined-strip-path-components=<number>
1039
1040Strip (or keep only, if negative) a given number of path components when emitting check metadata.
1041
1042.. option:: -fsanitize=<check>,<arg2>..., -fno-sanitize=<arg1>,<arg2>...
1043
1044Turn on runtime checks for various forms of undefined or suspicious behavior. See user manual for available checks
1045
1046.. option:: -moutline, -mno-outline
1047
1048Enable function outlining (AArch64 only)
1049
1050.. option:: -moutline-atomics, -mno-outline-atomics
1051
1052Generate local calls to out-of-line atomic operations
1053
1054.. option:: --param <arg>, --param=<arg>
1055
1056.. option:: -print-supported-cpus, --print-supported-cpus, -mcpu=?, -mtune=?
1057
1058Print supported cpu models for the given target (if target is not specified, it will print the supported cpus for the default target)
1059
1060.. option:: -std=<arg>, --std=<arg>, --std <arg>
1061
1062Language standard to compile for
1063
1064Preprocessor flags
1065~~~~~~~~~~~~~~~~~~
1066
1067Flags controlling the behavior of the Clang preprocessor.
1068
1069.. option:: -C, --comments
1070
1071Include comments in preprocessed output
1072
1073.. option:: -CC, --comments-in-macros
1074
1075Include comments from within macros in preprocessed output
1076
1077.. option:: -D<macro>=<value>, --define-macro <arg>, --define-macro=<arg>
1078
1079Define <macro> to <value> (or 1 if <value> omitted)
1080
1081.. option:: -H, --trace-includes
1082
1083Show header includes and nesting depth
1084
1085.. option:: -P, --no-line-commands
1086
1087Disable linemarker output in -E mode
1088
1089.. option:: -U<macro>, --undefine-macro <arg>, --undefine-macro=<arg>
1090
1091Undefine macro <macro>
1092
1093.. option:: -Wp,<arg>,<arg2>...
1094
1095Pass the comma separated arguments in <arg> to the preprocessor
1096
1097.. option:: -Xpreprocessor <arg>
1098
1099Pass <arg> to the preprocessor
1100
1101Include path management
1102-----------------------
1103
1104Flags controlling how ``#include``\s are resolved to files.
1105
1106.. option:: -I<dir>, --include-directory <arg>, --include-directory=<arg>
1107
1108Add directory to include search path. For C++ inputs, if
1109there are multiple -I options, these directories are searched
1110in the order they are given before the standard system directories
1111are searched. If the same directory is in the SYSTEM include search
1112paths, for example if also specified with -isystem, the -I option
1113will be ignored
1114
1115.. option:: -I-, --include-barrier
1116
1117Restrict all prior -I flags to double-quoted inclusion and remove current directory from include path
1118
1119.. option:: --amdgpu-arch-tool=<arg>
1120
1121Tool used for detecting AMD GPU arch in the system.
1122
1123.. option:: --cuda-path-ignore-env
1124
1125Ignore environment variables to detect CUDA installation
1126
1127.. option:: --cuda-path=<arg>
1128
1129CUDA installation path
1130
1131.. option:: -cxx-isystem<directory>
1132
1133Add directory to the C++ SYSTEM include search path
1134
1135.. option:: -fbuild-session-file=<file>
1136
1137Use the last modification time of <file> as the build session timestamp
1138
1139.. option:: -fbuild-session-timestamp=<time since Epoch in seconds>
1140
1141Time when the current build session started
1142
1143.. option:: -fmodule-file=\[<name>=\]<file>
1144
1145Specify the mapping of module name to precompiled module file, or load a module file if name is omitted.
1146
1147.. option:: -fmodules-cache-path=<directory>
1148
1149Specify the module cache path
1150
1151.. option:: -fmodules-disable-diagnostic-validation
1152
1153Disable validation of the diagnostic options when loading the module
1154
1155.. option:: -fmodules-prune-after=<seconds>
1156
1157Specify the interval (in seconds) after which a module file will be considered unused
1158
1159.. option:: -fmodules-prune-interval=<seconds>
1160
1161Specify the interval (in seconds) between attempts to prune the module cache
1162
1163.. option:: -fmodules-user-build-path <directory>
1164
1165Specify the module user build path
1166
1167.. option:: -fmodules-validate-once-per-build-session
1168
1169Don't verify input files for the modules if the module has been successfully validated or loaded during this build session
1170
1171.. option:: -fmodules-validate-system-headers, -fno-modules-validate-system-headers
1172
1173Validate the system headers that a module depends on when loading the module
1174
1175.. option:: -fprebuilt-module-path=<directory>
1176
1177Specify the prebuilt module path
1178
1179.. option:: --hip-path=<arg>
1180
1181HIP runtime installation path, used for finding HIP version and adding HIP include path.
1182
1183.. option:: -idirafter<arg>, --include-directory-after <arg>, --include-directory-after=<arg>
1184
1185Add directory to AFTER include search path
1186
1187.. option:: -iframework<arg>
1188
1189Add directory to SYSTEM framework search path
1190
1191.. option:: -iframeworkwithsysroot<directory>
1192
1193Add directory to SYSTEM framework search path, absolute paths are relative to -isysroot
1194
1195.. option:: -imacros<file>, --imacros<file>, --imacros=<arg>
1196
1197Include macros from file before parsing
1198
1199.. option:: -include<file>, --include<file>, --include=<arg>
1200
1201Include file before parsing
1202
1203.. option:: -include-pch <file>
1204
1205Include precompiled header file
1206
1207.. option:: -iprefix<dir>, --include-prefix <arg>, --include-prefix=<arg>
1208
1209Set the -iwithprefix/-iwithprefixbefore prefix
1210
1211.. option:: -iquote<directory>
1212
1213Add directory to QUOTE include search path
1214
1215.. option:: -isysroot<dir>
1216
1217Set the system root directory (usually /)
1218
1219.. option:: -isystem<directory>
1220
1221Add directory to SYSTEM include search path
1222
1223.. option:: -isystem-after<directory>
1224
1225Add directory to end of the SYSTEM include search path
1226
1227.. option:: -ivfsoverlay<arg>
1228
1229Overlay the virtual filesystem described by file over the real file system
1230
1231.. option:: -iwithprefix<dir>, --include-with-prefix <arg>, --include-with-prefix-after <arg>, --include-with-prefix-after=<arg>, --include-with-prefix=<arg>
1232
1233Set directory to SYSTEM include search path with prefix
1234
1235.. option:: -iwithprefixbefore<dir>, --include-with-prefix-before <arg>, --include-with-prefix-before=<arg>
1236
1237Set directory to include search path with prefix
1238
1239.. option:: -iwithsysroot<directory>
1240
1241Add directory to SYSTEM include search path, absolute paths are relative to -isysroot
1242
1243.. option:: --libomptarget-amdgpu-bc-path=<arg>, --libomptarget-amdgcn-bc-path=<arg>
1244
1245Path to libomptarget-amdgcn bitcode library
1246
1247.. option:: --libomptarget-nvptx-bc-path=<arg>
1248
1249Path to libomptarget-nvptx bitcode library
1250
1251.. option:: --ptxas-path=<arg>
1252
1253Path to ptxas (used for compiling CUDA code)
1254
1255.. option:: --rocm-path=<arg>
1256
1257ROCm installation path, used for finding and automatically linking required bitcode libraries.
1258
1259.. program:: clang1
1260.. option:: -stdlib++-isystem<directory>
1261.. program:: clang
1262
1263Use directory as the C++ standard library include path
1264
1265.. option:: --system-header-prefix=<prefix>, --no-system-header-prefix=<prefix>, --system-header-prefix <arg>
1266
1267Treat all #include paths starting with <prefix> as including a system header.
1268
1269Dependency file generation
1270--------------------------
1271
1272Flags controlling generation of a dependency file for ``make``-like build
1273systems.
1274
1275.. option:: -M, --dependencies
1276
1277Like -MD, but also implies -E and writes to stdout by default
1278
1279.. option:: -MD, --write-dependencies
1280
1281Write a depfile containing user and system headers
1282
1283.. option:: -MF<file>
1284
1285Write depfile output from -MMD, -MD, -MM, or -M to <file>
1286
1287.. option:: -MG, --print-missing-file-dependencies
1288
1289Add missing headers to depfile
1290
1291.. option:: -MJ<arg>
1292
1293Write a compilation database entry per input
1294
1295.. option:: -MM, --user-dependencies
1296
1297Like -MMD, but also implies -E and writes to stdout by default
1298
1299.. option:: -MMD, --write-user-dependencies
1300
1301Write a depfile containing user headers
1302
1303.. option:: -MP
1304
1305Create phony target for each dependency (other than main file)
1306
1307.. option:: -MQ<arg>
1308
1309Specify name of main file output to quote in depfile
1310
1311.. option:: -MT<arg>
1312
1313Specify name of main file output in depfile
1314
1315.. option:: -MV
1316
1317Use NMake/Jom format for the depfile
1318
1319Dumping preprocessor state
1320--------------------------
1321
1322Flags allowing the state of the preprocessor to be dumped in various ways.
1323
1324.. option:: -d
1325
1326.. program:: clang1
1327.. option:: -d<arg>
1328.. program:: clang
1329
1330.. option:: -dD
1331
1332Print macro definitions in -E mode in addition to normal output
1333
1334.. option:: -dI
1335
1336Print include directives in -E mode in addition to normal output
1337
1338.. option:: -dM
1339
1340Print macro definitions in -E mode instead of normal output
1341
1342Diagnostic flags
1343~~~~~~~~~~~~~~~~
1344
1345Flags controlling which warnings, errors, and remarks Clang will generate.
1346See the :doc:`full list of warning and remark flags <DiagnosticsReference>`.
1347
1348.. option:: -R<remark>
1349
1350Enable the specified remark
1351
1352.. option:: -Rpass-analysis=<arg>
1353
1354Report transformation analysis from optimization passes whose name matches the given POSIX regular expression
1355
1356.. option:: -Rpass-missed=<arg>
1357
1358Report missed transformations by optimization passes whose name matches the given POSIX regular expression
1359
1360.. option:: -Rpass=<arg>
1361
1362Report transformations performed by optimization passes whose name matches the given POSIX regular expression
1363
1364.. option:: -W<warning>, --extra-warnings, --warn-<arg>, --warn-=<arg>
1365
1366Enable the specified warning
1367
1368.. option:: -Wdeprecated, -Wno-deprecated
1369
1370Enable warnings for deprecated constructs and define \_\_DEPRECATED
1371
1372.. option:: -Wframe-larger-than=<arg>, -Wframe-larger-than
1373
1374.. option:: -Wnonportable-cfstrings<arg>, -Wno-nonportable-cfstrings<arg>
1375
1376Target-independent compilation options
1377~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1378.. option:: -fPIC, -fno-PIC
1379
1380.. option:: -fPIE, -fno-PIE
1381
1382.. option:: -faccess-control, -fno-access-control
1383
1384.. option:: -faddrsig, -fno-addrsig
1385
1386Emit an address-significance table
1387
1388.. option:: -falign-functions, -fno-align-functions
1389
1390.. program:: clang1
1391.. option:: -falign-functions=<arg>
1392.. program:: clang
1393
1394.. option:: -falign-loops=<N>
1395
1396N must be a power of two. Align loops to the boundary
1397
1398.. program:: clang1
1399.. option:: -faligned-allocation, -faligned-new, -fno-aligned-allocation
1400.. program:: clang
1401
1402Enable C++17 aligned allocation functions
1403
1404.. option:: -fallow-editor-placeholders, -fno-allow-editor-placeholders
1405
1406Treat editor placeholders as valid source code
1407
1408.. option:: -fallow-unsupported
1409
1410.. option:: -faltivec, -fno-altivec
1411
1412.. option:: -faltivec-src-compat=<arg>
1413
1414Source-level compatibility for Altivec vectors (for PowerPC targets). This includes results of vector comparison (scalar for 'xl', vector for 'gcc') as well as behavior when initializing with a scalar (splatting for 'xl', element zero only for 'gcc'). For 'mixed', the compatibility is as 'gcc' for 'vector bool/vector pixel' and as 'xl' for other types. Current default is 'mixed'.
1415
1416.. option:: -fansi-escape-codes
1417
1418Use ANSI escape codes for diagnostics
1419
1420.. option:: -fapple-kext, -findirect-virtual-calls, -fterminated-vtables
1421
1422Use Apple's kernel extensions ABI
1423
1424.. option:: -fapple-link-rtlib
1425
1426Force linking the clang builtins runtime library
1427
1428.. option:: -fapple-pragma-pack, -fno-apple-pragma-pack
1429
1430Enable Apple gcc-compatible #pragma pack handling
1431
1432.. option:: -fapplication-extension, -fno-application-extension
1433
1434Restrict code to those available for App Extensions
1435
1436.. option:: -fapprox-func, -fno-approx-func
1437
1438Allow certain math function calls to be replaced with an approximately equivalent calculation
1439
1440.. option:: -fasm, -fno-asm
1441
1442.. option:: -fasm-blocks, -fno-asm-blocks
1443
1444.. option:: -fassociative-math, -fno-associative-math
1445
1446.. option:: -fassume-sane-operator-new, -fno-assume-sane-operator-new
1447
1448.. option:: -fast
1449
1450.. option:: -fastcp
1451
1452.. option:: -fastf
1453
1454.. option:: -fasync-exceptions, -fno-async-exceptions
1455
1456Enable EH Asynchronous exceptions
1457
1458.. option:: -fasynchronous-unwind-tables, -fno-asynchronous-unwind-tables
1459
1460.. option:: -fautolink, -fno-autolink
1461
1462.. option:: -fbasic-block-sections=<arg>
1463
1464Generate labels for each basic block or place each basic block or a subset of basic blocks in its own section.
1465
1466.. option:: -fbinutils-version=<major.minor>
1467
1468Produced object files can use all ELF features supported by this binutils version and newer. If -fno-integrated-as is specified, the generated assembly will consider GNU as support. 'none' means that all ELF features can be used, regardless of binutils support. Defaults to 2.26.
1469
1470.. option:: -fblocks, -fno-blocks
1471
1472Enable the 'blocks' language feature
1473
1474.. option:: -fbootclasspath=<arg>, --bootclasspath <arg>, --bootclasspath=<arg>
1475
1476.. option:: -fborland-extensions, -fno-borland-extensions
1477
1478Accept non-standard constructs supported by the Borland compiler
1479
1480.. option:: -fbracket-depth=<arg>
1481
1482.. option:: -fbuiltin, -fno-builtin
1483
1484.. option:: -fbuiltin-module-map
1485
1486Load the clang builtins module map file.
1487
1488.. program:: clang1
1489.. option:: -fc++-static-destructors, -fno-c++-static-destructors
1490.. program:: clang
1491
1492.. option:: -fcaret-diagnostics, -fno-caret-diagnostics
1493
1494.. option:: -fcf-protection=<arg>, -fcf-protection (equivalent to -fcf-protection=full)
1495
1496Instrument control-flow architecture protection. Options: return, branch, full, none.
1497
1498.. option:: -fcf-runtime-abi=<arg>
1499
1500.. option:: -fchar8\_t, -fno-char8\_t
1501
1502Enable C++ builtin type char8\_t
1503
1504.. option:: -fclasspath=<arg>, --CLASSPATH <arg>, --CLASSPATH=<arg>, --classpath <arg>, --classpath=<arg>
1505
1506.. option:: -fcolor-diagnostics, -fno-color-diagnostics
1507
1508Enable colors in diagnostics
1509
1510.. option:: -fcommon, -fno-common
1511
1512Place uninitialized global variables in a common block
1513
1514.. option:: -fcompile-resource=<arg>, --resource <arg>, --resource=<arg>
1515
1516.. option:: -fconstant-cfstrings, -fno-constant-cfstrings
1517
1518.. option:: -fconstant-string-class=<arg>
1519
1520.. option:: -fconstexpr-backtrace-limit=<arg>
1521
1522.. option:: -fconstexpr-depth=<arg>
1523
1524.. option:: -fconstexpr-steps=<arg>
1525
1526.. option:: -fconvergent-functions
1527
1528Assume functions may be convergent
1529
1530.. option:: -fcoroutines-ts, -fno-coroutines-ts
1531
1532Enable support for the C++ Coroutines TS
1533
1534.. option:: -fcoverage-compilation-dir=<arg>
1535
1536The compilation directory to embed in the coverage mapping.
1537
1538.. option:: -fcoverage-mapping, -fno-coverage-mapping
1539
1540Generate coverage mapping to enable code coverage analysis
1541
1542.. option:: -fcoverage-prefix-map=<arg>
1543
1544remap file source paths in coverage mapping
1545
1546.. option:: -fcreate-profile
1547
1548.. option:: -fcs-profile-generate
1549
1550Generate instrumented code to collect context sensitive execution counts into default.profraw (overridden by LLVM\_PROFILE\_FILE env var)
1551
1552.. program:: clang1
1553.. option:: -fcs-profile-generate=<directory>
1554.. program:: clang
1555
1556Generate instrumented code to collect context sensitive execution counts into <directory>/default.profraw (overridden by LLVM\_PROFILE\_FILE env var)
1557
1558.. option:: -fcuda-approx-transcendentals, -fno-cuda-approx-transcendentals
1559
1560Use approximate transcendental functions
1561
1562.. option:: -fcuda-short-ptr, -fno-cuda-short-ptr
1563
1564Use 32-bit pointers for accessing const/local/shared address spaces
1565
1566.. option:: -fcxx-exceptions, -fno-cxx-exceptions
1567
1568Enable C++ exceptions
1569
1570.. option:: -fcxx-modules, -fno-cxx-modules
1571
1572Enable modules for C++
1573
1574.. option:: -fdata-sections, -fno-data-sections
1575
1576Place each data in its own section
1577
1578.. option:: -fdebug-compilation-dir=<arg>, -fdebug-compilation-dir <arg>
1579
1580The compilation directory to embed in the debug info
1581
1582.. option:: -fdebug-default-version=<arg>
1583
1584Default DWARF version to use, if a -g option caused DWARF debug info to be produced
1585
1586.. option:: -fdebug-info-for-profiling, -fno-debug-info-for-profiling
1587
1588Emit extra debug info to make sample profile more accurate
1589
1590.. option:: -fdebug-macro, -fno-debug-macro
1591
1592Emit macro debug information
1593
1594.. option:: -fdebug-pass-arguments
1595
1596.. option:: -fdebug-pass-structure
1597
1598.. option:: -fdebug-prefix-map=<arg>
1599
1600remap file source paths in debug info
1601
1602.. option:: -fdebug-ranges-base-address, -fno-debug-ranges-base-address
1603
1604Use DWARF base address selection entries in .debug\_ranges
1605
1606.. option:: -fdebug-types-section, -fno-debug-types-section
1607
1608Place debug types in their own section (ELF Only)
1609
1610.. option:: -fdelayed-template-parsing, -fno-delayed-template-parsing
1611
1612Parse templated function definitions at the end of the translation unit
1613
1614.. option:: -fdelete-null-pointer-checks, -fno-delete-null-pointer-checks
1615
1616Treat usage of null pointers as undefined behavior (default)
1617
1618.. option:: -fdenormal-fp-math=<arg>
1619
1620.. option:: -fdiagnostics-absolute-paths
1621
1622Print absolute paths in diagnostics
1623
1624.. option:: -fdiagnostics-color, -fno-diagnostics-color
1625
1626.. program:: clang1
1627.. option:: -fdiagnostics-color=<arg>
1628.. program:: clang
1629
1630.. option:: -fdiagnostics-hotness-threshold=<value>
1631
1632Prevent optimization remarks from being output if they do not have at least this profile count. Use 'auto' to apply the threshold from profile summary
1633
1634.. option:: -fdiagnostics-show-hotness, -fno-diagnostics-show-hotness
1635
1636Enable profile hotness information in diagnostic line
1637
1638.. option:: -fdiagnostics-show-note-include-stack, -fno-diagnostics-show-note-include-stack
1639
1640Display include stacks for diagnostic notes
1641
1642.. option:: -fdiagnostics-show-option, -fno-diagnostics-show-option
1643
1644Print option name with mappable diagnostics
1645
1646.. option:: -fdiagnostics-show-template-tree
1647
1648Print a template comparison tree for differing templates
1649
1650.. option:: -fdigraphs, -fno-digraphs
1651
1652Enable alternative token representations '<:', ':>', '<%', '%>', '%:', '%:%:' (default)
1653
1654.. option:: -fdirect-access-external-data, -fno-direct-access-external-data
1655
1656Don't use GOT indirection to reference external data symbols
1657
1658.. option:: -fdollars-in-identifiers, -fno-dollars-in-identifiers
1659
1660Allow '$' in identifiers
1661
1662.. option:: -fdouble-square-bracket-attributes, -fno-double-square-bracket-attributes
1663
1664Enable '\[\[\]\]' attributes in all C and C++ language modes
1665
1666.. option:: -fdwarf-directory-asm, -fno-dwarf-directory-asm
1667
1668.. option:: -fdwarf-exceptions
1669
1670Use DWARF style exceptions
1671
1672.. option:: -felide-constructors, -fno-elide-constructors
1673
1674.. option:: -feliminate-unused-debug-symbols, -fno-eliminate-unused-debug-symbols
1675
1676.. option:: -feliminate-unused-debug-types, -fno-eliminate-unused-debug-types
1677
1678Do not emit  debug info for defined but unused types
1679
1680.. option:: -fembed-bitcode=<option>, -fembed-bitcode (equivalent to -fembed-bitcode=all), -fembed-bitcode-marker (equivalent to -fembed-bitcode=marker)
1681
1682Embed LLVM bitcode (option: off, all, bitcode, marker)
1683
1684.. option:: -fembed-offload-object=<arg>
1685
1686Embed Offloading device-side binary into host object file as a section.
1687
1688.. option:: -femit-all-decls
1689
1690Emit all declarations, even if unused
1691
1692.. option:: -femulated-tls, -fno-emulated-tls
1693
1694Use emutls functions to access thread\_local variables
1695
1696.. option:: -fenable-matrix
1697
1698Enable matrix data type and related builtin functions
1699
1700.. option:: -fencoding=<arg>, --encoding <arg>, --encoding=<arg>
1701
1702.. option:: -ferror-limit=<arg>
1703
1704.. option:: -fescaping-block-tail-calls, -fno-escaping-block-tail-calls
1705
1706.. option:: -fexceptions, -fno-exceptions
1707
1708Enable support for exception handling
1709
1710.. option:: -fexec-charset=<arg>
1711
1712.. option:: -fexperimental-new-constant-interpreter
1713
1714Enable the experimental new constant interpreter
1715
1716.. option:: -fextdirs=<arg>, --extdirs <arg>, --extdirs=<arg>
1717
1718.. option:: -fextend-arguments=<arg>
1719
1720Controls how scalar integer arguments are extended in calls to unprototyped and varargs functions
1721
1722.. option:: -ffast-math, -fno-fast-math
1723
1724Allow aggressive, lossy floating-point optimizations
1725
1726.. option:: -ffile-compilation-dir=<arg>
1727
1728The compilation directory to embed in the debug info and coverage mapping.
1729
1730.. option:: -ffile-prefix-map=<arg>
1731
1732remap file source paths in debug info, predefined preprocessor macros and \_\_builtin\_FILE()
1733
1734.. option:: -ffinite-loops, -fno-finite-loops
1735
1736Assume all loops are finite.
1737
1738.. option:: -ffinite-math-only, -fno-finite-math-only
1739
1740.. option:: -ffixed-point, -fno-fixed-point
1741
1742Enable fixed point types
1743
1744.. option:: -ffor-scope, -fno-for-scope
1745
1746.. option:: -fforce-dwarf-frame, -fno-force-dwarf-frame
1747
1748Always emit a debug frame section
1749
1750.. option:: -fforce-emit-vtables, -fno-force-emit-vtables
1751
1752Emits more virtual tables to improve devirtualization
1753
1754.. option:: -fforce-enable-int128, -fno-force-enable-int128
1755
1756Enable support for int128\_t type
1757
1758.. option:: -ffp-contract=<arg>
1759
1760Form fused FP ops (e.g. FMAs): fast (fuses across statements disregarding pragmas) \| on (only fuses in the same statement unless dictated by pragmas) \| off (never fuses) \| fast-honor-pragmas (fuses across statements unless diectated by pragmas). Default is 'fast' for CUDA, 'fast-honor-pragmas' for HIP, and 'on' otherwise.
1761
1762.. option:: -ffp-exception-behavior=<arg>
1763
1764Specifies the exception behavior of floating-point operations.
1765
1766.. option:: -ffp-model=<arg>
1767
1768Controls the semantics of floating-point calculations.
1769
1770.. option:: -ffreestanding
1771
1772Assert that the compilation takes place in a freestanding environment
1773
1774.. option:: -ffunction-sections, -fno-function-sections
1775
1776Place each function in its own section
1777
1778.. option:: -fgnu-inline-asm, -fno-gnu-inline-asm
1779
1780.. option:: -fgnu-keywords, -fno-gnu-keywords
1781
1782Allow GNU-extension keywords regardless of language standard
1783
1784.. option:: -fgnu-runtime
1785
1786Generate output compatible with the standard GNU Objective-C runtime
1787
1788.. option:: -fgnu89-inline, -fno-gnu89-inline
1789
1790Use the gnu89 inline semantics
1791
1792.. option:: -fgnuc-version=<arg>
1793
1794Sets various macros to claim compatibility with the given GCC version (default is 4.2.1)
1795
1796.. option:: -fgpu-allow-device-init, -fno-gpu-allow-device-init
1797
1798Allow device side init function in HIP (experimental)
1799
1800.. option:: -fgpu-defer-diag, -fno-gpu-defer-diag
1801
1802Defer host/device related diagnostic messages for CUDA/HIP
1803
1804.. option:: -fgpu-rdc, -fcuda-rdc, -fno-gpu-rdc
1805
1806Generate relocatable device code, also known as separate compilation mode
1807
1808.. option:: -fgpu-sanitize, -fno-gpu-sanitize
1809
1810Enable sanitizer for AMDGPU target
1811
1812.. option:: -fhip-fp32-correctly-rounded-divide-sqrt, -fno-hip-fp32-correctly-rounded-divide-sqrt
1813
1814Specify that single precision floating-point divide and sqrt used in the program source are correctly rounded (HIP device compilation only)
1815
1816.. option:: -fhip-new-launch-api, -fno-hip-new-launch-api
1817
1818Use new kernel launching API for HIP
1819
1820.. option:: -fhonor-infinities, -fhonor-infinites, -fno-honor-infinities
1821
1822.. option:: -fhonor-nans, -fno-honor-nans
1823
1824.. option:: -fhosted
1825
1826.. option:: -fignore-exceptions
1827
1828Enable support for ignoring exception handling constructs
1829
1830.. option:: -fimplicit-module-maps, -fmodule-maps, -fno-implicit-module-maps
1831
1832Implicitly search the file system for module map files.
1833
1834.. option:: -fimplicit-modules, -fno-implicit-modules
1835
1836.. option:: -finput-charset=<arg>
1837
1838Specify the default character set for source files
1839
1840.. option:: -finstrument-function-entry-bare
1841
1842Instrument function entry only, after inlining, without arguments to the instrumentation call
1843
1844.. option:: -finstrument-functions
1845
1846Generate calls to instrument function entry and exit
1847
1848.. option:: -finstrument-functions-after-inlining
1849
1850Like -finstrument-functions, but insert the calls after inlining
1851
1852.. option:: -fintegrated-as, -fno-integrated-as, -integrated-as
1853
1854Enable the integrated assembler
1855
1856.. option:: -fintegrated-cc1, -fno-integrated-cc1
1857
1858Run cc1 in-process
1859
1860.. option:: -fjump-tables, -fno-jump-tables
1861
1862Use jump tables for lowering switches
1863
1864.. option:: -fkeep-static-consts, -fno-keep-static-consts
1865
1866Keep static const variables if unused
1867
1868.. option:: -flax-vector-conversions=<arg>, -flax-vector-conversions (equivalent to -flax-vector-conversions=integer), -fno-lax-vector-conversions (equivalent to -flax-vector-conversions=none)
1869
1870Enable implicit vector bit-casts
1871
1872.. option:: -flimited-precision=<arg>
1873
1874.. option:: -flto-jobs=<arg>
1875
1876Controls the backend parallelism of -flto=thin (default of 0 means the number of threads will be derived from the number of CPUs detected)
1877
1878.. option:: -flto=<arg>, -flto (equivalent to -flto=full), -flto=auto (equivalent to -flto=full), -flto=jobserver (equivalent to -flto=full)
1879
1880Set LTO mode to either 'full' or 'thin'
1881
1882.. option:: -fmacro-backtrace-limit=<arg>
1883
1884.. option:: -fmacro-prefix-map=<arg>
1885
1886remap file source paths in predefined preprocessor macros and \_\_builtin\_FILE()
1887
1888.. option:: -fmath-errno, -fno-math-errno
1889
1890Require math functions to indicate errors by setting errno
1891
1892.. option:: -fmax-tokens=<arg>
1893
1894Max total number of preprocessed tokens for -Wmax-tokens.
1895
1896.. option:: -fmax-type-align=<arg>
1897
1898Specify the maximum alignment to enforce on pointers lacking an explicit alignment
1899
1900.. option:: -fmemory-profile, -fno-memory-profile
1901
1902Enable heap memory profiling
1903
1904.. program:: clang1
1905.. option:: -fmemory-profile=<directory>
1906.. program:: clang
1907
1908Enable heap memory profiling and dump results into <directory>
1909
1910.. option:: -fmerge-all-constants, -fno-merge-all-constants
1911
1912Allow merging of constants
1913
1914.. option:: -fmessage-length=<arg>
1915
1916Format message diagnostics so that they fit within N columns
1917
1918.. option:: -fminimize-whitespace, -fno-minimize-whitespace
1919
1920Minimize whitespace when emitting preprocessor output
1921
1922.. option:: -fmodule-file-deps, -fno-module-file-deps
1923
1924.. option:: -fmodule-map-file=<file>
1925
1926Load this module map file
1927
1928.. option:: -fmodule-name=<name>, -fmodule-implementation-of <arg>
1929
1930Specify the name of the module to build
1931
1932.. option:: -fmodules, -fno-modules
1933
1934Enable the 'modules' language feature
1935
1936.. option:: -fmodules-decluse, -fno-modules-decluse
1937
1938Require declaration of modules used within a module
1939
1940.. option:: -fmodules-ignore-macro=<arg>
1941
1942Ignore the definition of the given macro when building and loading modules
1943
1944.. option:: -fmodules-search-all, -fno-modules-search-all
1945
1946Search even non-imported modules to resolve references
1947
1948.. option:: -fmodules-strict-decluse
1949
1950Like -fmodules-decluse but requires all headers to be in modules
1951
1952.. option:: -fmodules-ts
1953
1954Enable support for the C++ Modules TS
1955
1956.. option:: -fmodules-validate-input-files-content
1957
1958Validate PCM input files based on content if mtime differs
1959
1960.. option:: -fms-compatibility, -fno-ms-compatibility
1961
1962Enable full Microsoft Visual C++ compatibility
1963
1964.. option:: -fms-compatibility-version=<arg>
1965
1966Dot-separated value representing the Microsoft compiler version number to report in \_MSC\_VER (0 = don't define it (default))
1967
1968.. option:: -fms-extensions, -fno-ms-extensions
1969
1970Accept some non-standard constructs supported by the Microsoft compiler
1971
1972.. option:: -fms-hotpatch
1973
1974Ensure that all functions can be hotpatched at runtime
1975
1976.. option:: -fms-memptr-rep=<arg>
1977
1978.. option:: -fms-volatile
1979
1980.. option:: -fmsc-version=<arg>
1981
1982Microsoft compiler version number to report in \_MSC\_VER (0 = don't define it (default))
1983
1984.. option:: -fmudflap
1985
1986.. option:: -fmudflapth
1987
1988.. option:: -fnested-functions
1989
1990.. option:: -fnew-alignment=<align>, -fnew-alignment <arg>
1991
1992Specifies the largest alignment guaranteed by '::operator new(size\_t)'
1993
1994.. option:: -fnew-infallible, -fno-new-infallible
1995
1996Enable treating throwing global C++ operator new as always returning valid memory (annotates with \_\_attribute\_\_((returns\_nonnull)) and throw()). This is detectable in source.
1997
1998.. option:: -fnext-runtime
1999
2000.. option:: -fno-builtin-<arg>
2001
2002Disable implicit builtin knowledge of a specific function
2003
2004.. option:: -fno-elide-type
2005
2006Do not elide types when printing diagnostics
2007
2008.. option:: -fno-max-type-align
2009
2010.. option:: -fno-strict-modules-decluse
2011
2012.. option:: -fno-temp-file
2013
2014Directly create compilation output files. This may lead to incorrect incremental builds if the compiler crashes
2015
2016.. option:: -fno-working-directory
2017
2018.. option:: -fno\_modules-validate-input-files-content
2019
2020.. program:: clang1
2021.. option:: -fno\_pch-validate-input-files-content
2022.. program:: clang
2023
2024.. option:: -fnoxray-link-deps
2025
2026.. option:: -fobjc-abi-version=<arg>
2027
2028.. option:: -fobjc-arc, -fno-objc-arc
2029
2030Synthesize retain and release calls for Objective-C pointers
2031
2032.. option:: -fobjc-arc-exceptions, -fno-objc-arc-exceptions
2033
2034Use EH-safe code when synthesizing retains and releases in -fobjc-arc
2035
2036.. option:: -fobjc-convert-messages-to-runtime-calls, -fno-objc-convert-messages-to-runtime-calls
2037
2038.. option:: -fobjc-disable-direct-methods-for-testing
2039
2040Ignore attribute objc\_direct so that direct methods can be tested
2041
2042.. option:: -fobjc-encode-cxx-class-template-spec, -fno-objc-encode-cxx-class-template-spec
2043
2044Fully encode c++ class template specialization
2045
2046.. option:: -fobjc-exceptions, -fno-objc-exceptions
2047
2048Enable Objective-C exceptions
2049
2050.. option:: -fobjc-infer-related-result-type, -fno-objc-infer-related-result-type
2051
2052.. option:: -fobjc-legacy-dispatch, -fno-objc-legacy-dispatch
2053
2054.. option:: -fobjc-link-runtime
2055
2056.. option:: -fobjc-nonfragile-abi, -fno-objc-nonfragile-abi
2057
2058.. option:: -fobjc-nonfragile-abi-version=<arg>
2059
2060.. option:: -fobjc-runtime=<arg>
2061
2062Specify the target Objective-C runtime kind and version
2063
2064.. option:: -fobjc-sender-dependent-dispatch
2065
2066.. option:: -fobjc-weak, -fno-objc-weak
2067
2068Enable ARC-style weak references in Objective-C
2069
2070.. option:: -foffload-lto=<arg>, -foffload-lto (equivalent to -foffload-lto=full)
2071
2072Set LTO mode to either 'full' or 'thin' for offload compilation
2073
2074.. option:: -fomit-frame-pointer, -fno-omit-frame-pointer
2075
2076.. option:: -fopenmp, -fno-openmp
2077
2078Parse OpenMP pragmas and generate parallel code.
2079
2080.. option:: -fopenmp-extensions, -fno-openmp-extensions
2081
2082Enable all Clang extensions for OpenMP directives and clauses
2083
2084.. option:: -fopenmp-implicit-rpath, -fno-openmp-implicit-rpath
2085
2086Set rpath on OpenMP executables
2087
2088.. option:: -fopenmp-simd, -fno-openmp-simd
2089
2090Emit OpenMP code only for SIMD-based constructs.
2091
2092.. option:: -fopenmp-target-debug, -fno-openmp-target-debug
2093
2094Enable debugging in the OpenMP offloading device RTL
2095
2096.. option:: -fopenmp-target-new-runtime, -fno-openmp-target-new-runtime
2097
2098Use the new bitcode library for OpenMP offloading
2099
2100.. option:: -fopenmp-offload-mandatory
2101
2102Indicate that offloading to the device is mandatory and do not generate host-fallback code.
2103
2104.. option:: -fopenmp-version=<arg>
2105
2106Set OpenMP version (e.g. 45 for OpenMP 4.5, 50 for OpenMP 5.0). Default value is 50.
2107
2108.. program:: clang1
2109.. option:: -fopenmp=<arg>
2110.. program:: clang
2111
2112.. option:: -foperator-arrow-depth=<arg>
2113
2114.. option:: -foperator-names, -fno-operator-names
2115
2116.. option:: -foptimization-record-file=<file>
2117
2118Specify the output name of the file containing the optimization remarks. Implies -fsave-optimization-record. On Darwin platforms, this cannot be used with multiple -arch <arch> options.
2119
2120.. option:: -foptimization-record-passes=<regex>
2121
2122Only include passes which match a specified regular expression in the generated optimization record (by default, include all passes)
2123
2124.. option:: -foptimize-sibling-calls, -fno-optimize-sibling-calls
2125
2126.. option:: -forder-file-instrumentation
2127
2128Generate instrumented code to collect order file into default.profraw file (overridden by '=' form of option or LLVM\_PROFILE\_FILE env var)
2129
2130.. option:: -foutput-class-dir=<arg>, --output-class-directory <arg>, --output-class-directory=<arg>
2131
2132.. option:: -fpack-struct, -fno-pack-struct
2133
2134.. program:: clang1
2135.. option:: -fpack-struct=<arg>
2136.. program:: clang
2137
2138Specify the default maximum struct packing alignment
2139
2140.. option:: -fpascal-strings, -fno-pascal-strings, -mpascal-strings
2141
2142Recognize and construct Pascal-style string literals
2143
2144.. option:: -fpass-plugin=<dsopath>
2145
2146Load pass plugin from a dynamic shared object file (only with new pass manager).
2147
2148.. option:: -fpatchable-function-entry=<N,M>
2149
2150Generate M NOPs before function entry and N-M NOPs after function entry
2151
2152.. option:: -fpcc-struct-return
2153
2154Override the default ABI to return all structs on the stack
2155
2156.. option:: -fpch-codegen, -fno-pch-codegen
2157
2158Generate code for uses of this PCH that assumes an explicit object file will be built for the PCH
2159
2160.. option:: -fpch-debuginfo, -fno-pch-debuginfo
2161
2162Generate debug info for types in an object file built from this PCH and do not generate them elsewhere
2163
2164.. option:: -fpch-instantiate-templates, -fno-pch-instantiate-templates
2165
2166Instantiate templates already while building a PCH
2167
2168.. option:: -fpch-preprocess
2169
2170.. option:: -fpch-validate-input-files-content
2171
2172Validate PCH input files based on content if mtime differs
2173
2174.. option:: -fpic, -fno-pic
2175
2176.. option:: -fpie, -fno-pie
2177
2178.. option:: -fplt, -fno-plt
2179
2180.. option:: -fplugin=<dsopath>
2181
2182Load the named plugin (dynamic shared object)
2183
2184.. option:: -fprebuilt-implicit-modules, -fno-prebuilt-implicit-modules
2185
2186Look up implicit modules in the prebuilt module path
2187
2188.. option:: -fpreserve-as-comments, -fno-preserve-as-comments
2189
2190.. option:: -fproc-stat-report<arg>
2191
2192Print subprocess statistics
2193
2194.. program:: clang1
2195.. option:: -fproc-stat-report=<arg>
2196.. program:: clang
2197
2198Save subprocess statistics to the given file
2199
2200.. option:: -fprofile-arcs, -fno-profile-arcs
2201
2202.. option:: -fprofile-dir=<arg>
2203
2204.. option:: -fprofile-exclude-files=<arg>
2205
2206Instrument only functions from files where names don't match all the regexes separated by a semi-colon
2207
2208.. option:: -fprofile-filter-files=<arg>
2209
2210Instrument only functions from files where names match any regex separated by a semi-colon
2211
2212.. option:: -fprofile-generate, -fno-profile-generate
2213
2214Generate instrumented code to collect execution counts into default.profraw (overridden by LLVM\_PROFILE\_FILE env var)
2215
2216.. program:: clang1
2217.. option:: -fprofile-generate=<directory>
2218.. program:: clang
2219
2220Generate instrumented code to collect execution counts into <directory>/default.profraw (overridden by LLVM\_PROFILE\_FILE env var)
2221
2222.. option:: -fprofile-instr-generate, -fno-profile-instr-generate
2223
2224Generate instrumented code to collect execution counts into default.profraw file (overridden by '=' form of option or LLVM\_PROFILE\_FILE env var)
2225
2226.. program:: clang1
2227.. option:: -fprofile-instr-generate=<file>
2228.. program:: clang
2229
2230Generate instrumented code to collect execution counts into <file> (overridden by LLVM\_PROFILE\_FILE env var)
2231
2232.. option:: -fprofile-instr-use, -fno-profile-instr-use, -fprofile-use
2233
2234.. program:: clang1
2235.. option:: -fprofile-instr-use=<arg>
2236.. program:: clang
2237
2238Use instrumentation data for profile-guided optimization
2239
2240.. option:: -fprofile-list=<arg>
2241
2242Filename defining the list of functions/files to instrument
2243
2244.. option:: -fprofile-remapping-file=<file>
2245
2246Use the remappings described in <file> to match the profile data against names in the program
2247
2248.. option:: -fprofile-sample-accurate, -fauto-profile-accurate, -fno-profile-sample-accurate
2249
2250Specifies that the sample profile is accurate. If the sample
2251               profile is accurate, callsites without profile samples are marked
2252               as cold. Otherwise, treat callsites without profile samples as if
2253               we have no profile
2254
2255.. option:: -fprofile-sample-use, -fauto-profile, -fno-profile-sample-use
2256
2257.. program:: clang1
2258.. option:: -fprofile-sample-use=<arg>, -fauto-profile=<arg>
2259.. program:: clang
2260
2261Enable sample-based profile guided optimizations
2262
2263.. option:: -fprofile-update=<method>
2264
2265Set update method of profile counters (atomic,prefer-atomic,single)
2266
2267.. program:: clang1
2268.. option:: -fprofile-use=<pathname>
2269.. program:: clang
2270
2271Use instrumentation data for profile-guided optimization. If pathname is a directory, it reads from <pathname>/default.profdata. Otherwise, it reads from file <pathname>.
2272
2273.. option:: -fprotect-parens, -fno-protect-parens
2274
2275Determines whether the optimizer honors parentheses when floating-point expressions are evaluated
2276
2277.. option:: -fpseudo-probe-for-profiling, -fno-pseudo-probe-for-profiling
2278
2279Emit pseudo probes for sample profiling
2280
2281.. option:: -freciprocal-math, -fno-reciprocal-math
2282
2283Allow division operations to be reassociated
2284
2285.. option:: -freg-struct-return
2286
2287Override the default ABI to return small structs in registers
2288
2289.. option:: -fregister-global-dtors-with-atexit, -fno-register-global-dtors-with-atexit
2290
2291Use atexit or \_\_cxa\_atexit to register global destructors
2292
2293.. option:: -frelaxed-template-template-args, -fno-relaxed-template-template-args
2294
2295Enable C++17 relaxed template template argument matching
2296
2297.. option:: -freroll-loops, -fno-reroll-loops
2298
2299Turn on loop reroller
2300
2301.. option:: -fretain-comments-from-system-headers
2302
2303.. option:: -frewrite-imports, -fno-rewrite-imports
2304
2305.. option:: -frewrite-includes, -fno-rewrite-includes
2306
2307.. option:: -frewrite-map-file=<arg>
2308
2309.. option:: -fropi, -fno-ropi
2310
2311Generate read-only position independent code (ARM only)
2312
2313.. option:: -frounding-math, -fno-rounding-math
2314
2315.. option:: -frtti, -fno-rtti
2316
2317.. option:: -frtti-data, -fno-rtti-data
2318
2319.. option:: -frwpi, -fno-rwpi
2320
2321Generate read-write position independent code (ARM only)
2322
2323.. option:: -fsanitize-memory-param-retval, -fno-sanitize-memory-param-retval
2324
2325Enable detection of uninitialized parameters and return values
2326
2327.. option:: -fsave-optimization-record, -fno-save-optimization-record
2328
2329Generate a YAML optimization record file
2330
2331.. program:: clang1
2332.. option:: -fsave-optimization-record=<format>
2333.. program:: clang
2334
2335Generate an optimization record file in a specific format
2336
2337.. option:: -fseh-exceptions
2338
2339Use SEH style exceptions
2340
2341.. option:: -fsemantic-interposition, -fno-semantic-interposition
2342
2343.. option:: -fshort-enums, -fno-short-enums
2344
2345Allocate to an enum type only as many bytes as it needs for the declared range of possible values
2346
2347.. option:: -fshort-wchar, -fno-short-wchar
2348
2349Force wchar\_t to be a short unsigned int
2350
2351.. option:: -fshow-column, -fno-show-column
2352
2353.. option:: -fshow-overloads=<arg>
2354
2355Which overload candidates to show when overload resolution fails: best\|all; defaults to all
2356
2357.. option:: -fshow-source-location, -fno-show-source-location
2358
2359.. option:: -fsignaling-math, -fno-signaling-math
2360
2361.. option:: -fsigned-bitfields
2362
2363.. option:: -fsigned-char, -fno-signed-char, --signed-char
2364
2365char is signed
2366
2367.. option:: -fsigned-zeros, -fno-signed-zeros
2368
2369.. option:: -fsized-deallocation, -fno-sized-deallocation
2370
2371Enable C++14 sized global deallocation functions
2372
2373.. option:: -fsjlj-exceptions
2374
2375Use SjLj style exceptions
2376
2377.. option:: -fslp-vectorize, -fno-slp-vectorize, -ftree-slp-vectorize
2378
2379Enable the superword-level parallelism vectorization passes
2380
2381.. option:: -fspell-checking, -fno-spell-checking
2382
2383.. option:: -fspell-checking-limit=<arg>
2384
2385.. option:: -fsplit-dwarf-inlining, -fno-split-dwarf-inlining
2386
2387Provide minimal debug info in the object/executable to facilitate online symbolication/stack traces in the absence of .dwo/.dwp files when using Split DWARF
2388
2389.. option:: -fsplit-lto-unit, -fno-split-lto-unit
2390
2391Enables splitting of the LTO unit
2392
2393.. option:: -fsplit-machine-functions, -fno-split-machine-functions
2394
2395Enable late function splitting using profile information (x86 ELF)
2396
2397.. option:: -fsplit-stack, -fno-split-stack
2398
2399Use segmented stack
2400
2401.. option:: -fstack-clash-protection, -fno-stack-clash-protection
2402
2403Enable stack clash protection
2404
2405.. option:: -fstack-protector, -fno-stack-protector
2406
2407Enable stack protectors for some functions vulnerable to stack smashing. This uses a loose heuristic which considers functions vulnerable if they contain a char (or 8bit integer) array or constant sized calls to alloca , which are of greater size than ssp-buffer-size (default: 8 bytes). All variable sized calls to alloca are considered vulnerable. A function with a stack protector has a guard value added to the stack frame that is checked on function exit. The guard value must be positioned in the stack frame such that a buffer overflow from a vulnerable variable will overwrite the guard value before overwriting the function's return address. The reference stack guard value is stored in a global variable.
2408
2409.. option:: -fstack-protector-all
2410
2411Enable stack protectors for all functions
2412
2413.. option:: -fstack-protector-strong
2414
2415Enable stack protectors for some functions vulnerable to stack smashing. Compared to -fstack-protector, this uses a stronger heuristic that includes functions containing arrays of any size (and any type), as well as any calls to alloca or the taking of an address from a local variable
2416
2417.. option:: -fstack-size-section, -fno-stack-size-section
2418
2419Emit section containing metadata on function stack sizes
2420
2421.. option:: -fstack-usage
2422
2423Emit .su file containing information on function stack sizes
2424
2425.. option:: -fstandalone-debug, -fno-limit-debug-info, -fno-standalone-debug
2426
2427Emit full debug info for all types used by the program
2428
2429.. option:: -fstrict-aliasing, -fno-strict-aliasing
2430
2431.. option:: -fstrict-enums, -fno-strict-enums
2432
2433Enable optimizations based on the strict definition of an enum's value range
2434
2435.. option:: -fstrict-float-cast-overflow, -fno-strict-float-cast-overflow
2436
2437Assume that overflowing float-to-int casts are undefined (default)
2438
2439.. option:: -fstrict-overflow, -fno-strict-overflow
2440
2441.. option:: -fstrict-return, -fno-strict-return
2442
2443.. option:: -fstrict-vtable-pointers, -fno-strict-vtable-pointers
2444
2445Enable optimizations based on the strict rules for overwriting polymorphic C++ objects
2446
2447.. option:: -fstruct-path-tbaa, -fno-struct-path-tbaa
2448
2449.. option:: -fswift-async-fp=<option>
2450
2451Control emission of Swift async extended frame info (option: auto, always, never)
2452
2453.. option:: -fsymbol-partition=<arg>
2454
2455.. option:: -ftabstop=<arg>
2456
2457.. option:: -ftemplate-backtrace-limit=<arg>
2458
2459.. option:: -ftemplate-depth-<arg>
2460
2461.. option:: -ftemplate-depth=<arg>
2462
2463.. option:: -ftest-coverage, -fno-test-coverage
2464
2465.. option:: -fthin-link-bitcode=<arg>
2466
2467Write minimized bitcode to <file> for the ThinLTO thin link only
2468
2469.. option:: -fthinlto-index=<arg>
2470
2471Perform ThinLTO importing using provided function summary index
2472
2473.. option:: -fthreadsafe-statics, -fno-threadsafe-statics
2474
2475.. option:: -ftime-report
2476
2477.. program:: clang1
2478.. option:: -ftime-report=<arg>
2479.. program:: clang
2480
2481(For new pass manager) "per-pass": one report for each pass; "per-pass-run": one report for each pass invocation
2482
2483.. option:: -ftime-trace
2484
2485
2486Turn on time profiler. Generates JSON file based on output filename. Results
2487can be analyzed with chrome://tracing or `Speedscope App
2488<https://www.speedscope.app>`_ for flamegraph visualization.
2489
2490.. option:: -ftime-trace-granularity=<arg>
2491
2492Minimum time granularity (in microseconds) traced by time profiler
2493
2494.. option:: -ftls-model=<arg>
2495
2496.. option:: -ftrap-function=<arg>
2497
2498Issue call to specified function rather than a trap instruction
2499
2500.. option:: -ftrapping-math, -fno-trapping-math
2501
2502.. option:: -ftrapv
2503
2504Trap on integer overflow
2505
2506.. option:: -ftrapv-handler <arg>
2507
2508.. program:: clang1
2509.. option:: -ftrapv-handler=<function name>
2510.. program:: clang
2511
2512Specify the function to be called on overflow
2513
2514.. option:: -ftrigraphs, -fno-trigraphs, -trigraphs, --trigraphs
2515
2516Process trigraph sequences
2517
2518.. option:: -ftrivial-auto-var-init-stop-after=<arg>
2519
2520Stop initializing trivial automatic stack variables after the specified number of instances
2521
2522.. option:: -ftrivial-auto-var-init=<arg>
2523
2524Initialize trivial automatic stack variables: uninitialized (default) \| pattern
2525
2526.. option:: -funique-basic-block-section-names, -fno-unique-basic-block-section-names
2527
2528Use unique names for basic block sections (ELF Only)
2529
2530.. option:: -funique-internal-linkage-names, -fno-unique-internal-linkage-names
2531
2532Uniqueify Internal Linkage Symbol Names by appending the MD5 hash of the module path
2533
2534.. option:: -funique-section-names, -fno-unique-section-names
2535
2536.. option:: -funit-at-a-time, -fno-unit-at-a-time
2537
2538.. option:: -funroll-loops, -fno-unroll-loops
2539
2540Turn on loop unroller
2541
2542.. option:: -funsafe-math-optimizations, -fno-unsafe-math-optimizations
2543
2544.. option:: -funsigned-bitfields
2545
2546.. option:: -funsigned-char, -fno-unsigned-char, --unsigned-char
2547
2548.. option:: -funwind-tables, -fno-unwind-tables
2549
2550.. option:: -fuse-cxa-atexit, -fno-use-cxa-atexit
2551
2552.. option:: -fuse-init-array, -fno-use-init-array
2553
2554.. option:: -fuse-ld=<arg>
2555
2556.. option:: -fuse-line-directives, -fno-use-line-directives
2557
2558Use #line in preprocessed output
2559
2560.. option:: -fvalidate-ast-input-files-content
2561
2562Compute and store the hash of input files used to build an AST. Files with mismatching mtime's are considered valid if both contents is identical
2563
2564.. option:: -fveclib=<arg>
2565
2566Use the given vector functions library
2567
2568.. option:: -fvectorize, -fno-vectorize, -ftree-vectorize
2569
2570Enable the loop vectorization passes
2571
2572.. option:: -fverbose-asm, -dA, -fno-verbose-asm
2573
2574Generate verbose assembly output
2575
2576.. option:: -fvirtual-function-elimination, -fno-virtual-function-elimination
2577
2578Enables dead virtual function elimination optimization. Requires -flto=full
2579
2580.. option:: -fvisibility-dllexport=<arg>
2581
2582The visibility for dllexport definitions \[-fvisibility-from-dllstorageclass\]
2583
2584.. option:: -fvisibility-externs-dllimport=<arg>
2585
2586The visibility for dllimport external declarations \[-fvisibility-from-dllstorageclass\]
2587
2588.. option:: -fvisibility-externs-nodllstorageclass=<arg>
2589
2590The visibility for external declarations without an explicit DLL dllstorageclass \[-fvisibility-from-dllstorageclass\]
2591
2592.. option:: -fvisibility-from-dllstorageclass, -fno-visibility-from-dllstorageclass
2593
2594Set the visibility of symbols in the generated code from their DLL storage class
2595
2596.. option:: -fvisibility-global-new-delete-hidden
2597
2598Give global C++ operator new and delete declarations hidden visibility
2599
2600.. option:: -fvisibility-inlines-hidden, -fno-visibility-inlines-hidden
2601
2602Give inline C++ member functions hidden visibility by default
2603
2604.. option:: -fvisibility-inlines-hidden-static-local-var, -fno-visibility-inlines-hidden-static-local-var
2605
2606When -fvisibility-inlines-hidden is enabled, static variables in inline C++ member functions will also be given hidden visibility by default
2607
2608.. option:: -fvisibility-ms-compat
2609
2610Give global types 'default' visibility and global functions and variables 'hidden' visibility by default
2611
2612.. option:: -fvisibility-nodllstorageclass=<arg>
2613
2614The visibility for defintiions without an explicit DLL export class \[-fvisibility-from-dllstorageclass\]
2615
2616.. option:: -fvisibility=<arg>
2617
2618Set the default symbol visibility for all global declarations
2619
2620.. option:: -fwasm-exceptions
2621
2622Use WebAssembly style exceptions
2623
2624.. option:: -fwhole-program-vtables, -fno-whole-program-vtables
2625
2626Enables whole-program vtable optimization. Requires -flto
2627
2628.. option:: -fwrapv, -fno-wrapv
2629
2630Treat signed integer overflow as two's complement
2631
2632.. option:: -fwritable-strings
2633
2634Store string literals as writable data
2635
2636.. option:: -fxl-pragma-pack, -fno-xl-pragma-pack
2637
2638Enable IBM XL #pragma pack handling
2639
2640.. option:: -fxray-always-emit-customevents, -fno-xray-always-emit-customevents
2641
2642Always emit \_\_xray\_customevent(...) calls even if the containing function is not always instrumented
2643
2644.. option:: -fxray-always-emit-typedevents, -fno-xray-always-emit-typedevents
2645
2646Always emit \_\_xray\_typedevent(...) calls even if the containing function is not always instrumented
2647
2648.. option:: -fxray-always-instrument=<arg>
2649
2650DEPRECATED: Filename defining the whitelist for imbuing the 'always instrument' XRay attribute.
2651
2652.. option:: -fxray-attr-list=<arg>
2653
2654Filename defining the list of functions/types for imbuing XRay attributes.
2655
2656.. option:: -fxray-function-groups=<arg>
2657
2658Only instrument 1 of N groups
2659
2660.. option:: -fxray-function-index, -fno-xray-function-index
2661
2662.. option:: -fxray-ignore-loops, -fno-xray-ignore-loops
2663
2664Don't instrument functions with loops unless they also meet the minimum function size
2665
2666.. option:: -fxray-instruction-threshold<arg>
2667
2668.. program:: clang1
2669.. option:: -fxray-instruction-threshold=<arg>
2670.. program:: clang
2671
2672Sets the minimum function size to instrument with XRay
2673
2674.. option:: -fxray-instrument, -fno-xray-instrument
2675
2676Generate XRay instrumentation sleds on function entry and exit
2677
2678.. option:: -fxray-instrumentation-bundle=<arg>
2679
2680Select which XRay instrumentation points to emit. Options: all, none, function-entry, function-exit, function, custom. Default is 'all'.  'function' includes both 'function-entry' and 'function-exit'.
2681
2682.. option:: -fxray-link-deps
2683
2684Tells clang to add the link dependencies for XRay.
2685
2686.. option:: -fxray-modes=<arg>
2687
2688List of modes to link in by default into XRay instrumented binaries.
2689
2690.. option:: -fxray-never-instrument=<arg>
2691
2692DEPRECATED: Filename defining the whitelist for imbuing the 'never instrument' XRay attribute.
2693
2694.. option:: -fxray-selected-function-group=<arg>
2695
2696When using -fxray-function-groups, select which group of functions to instrument. Valid range is 0 to fxray-function-groups - 1
2697
2698.. option:: -fzero-call-used-regs=<arg>
2699
2700Clear call-used registers upon function return.
2701
2702.. option:: -fzero-initialized-in-bss, -fno-zero-initialized-in-bss
2703
2704.. option:: -fzvector, -fno-zvector, -mzvector
2705
2706Enable System z vector language extension
2707
2708.. option:: --gpu-bundle-output, --no-gpu-bundle-output
2709
2710Bundle output files of HIP device compilation
2711
2712.. option:: -pedantic, --pedantic, -no-pedantic, --no-pedantic
2713
2714Warn on language extensions
2715
2716.. option:: -pedantic-errors, --pedantic-errors
2717
2718OpenCL flags
2719------------
2720.. option:: -cl-denorms-are-zero
2721
2722OpenCL only. Allow denormals to be flushed to zero.
2723
2724.. option:: -cl-fast-relaxed-math
2725
2726OpenCL only. Sets -cl-finite-math-only and -cl-unsafe-math-optimizations, and defines \_\_FAST\_RELAXED\_MATH\_\_.
2727
2728.. option:: -cl-finite-math-only
2729
2730OpenCL only. Allow floating-point optimizations that assume arguments and results are not NaNs or +-Inf.
2731
2732.. option:: -cl-fp32-correctly-rounded-divide-sqrt
2733
2734OpenCL only. Specify that single precision floating-point divide and sqrt used in the program source are correctly rounded.
2735
2736.. option:: -cl-kernel-arg-info
2737
2738OpenCL only. Generate kernel argument metadata.
2739
2740.. option:: -cl-mad-enable
2741
2742OpenCL only. Allow use of less precise MAD computations in the generated binary.
2743
2744.. option:: -cl-no-signed-zeros
2745
2746OpenCL only. Allow use of less precise no signed zeros computations in the generated binary.
2747
2748.. option:: -cl-no-stdinc
2749
2750OpenCL only. Disables all standard includes containing non-native compiler types and functions.
2751
2752.. option:: -cl-opt-disable
2753
2754OpenCL only. This option disables all optimizations. By default optimizations are enabled.
2755
2756.. option:: -cl-single-precision-constant
2757
2758OpenCL only. Treat double precision floating-point constant as single precision constant.
2759
2760.. option:: -cl-std=<arg>
2761
2762OpenCL language standard to compile for.
2763
2764.. option:: -cl-strict-aliasing
2765
2766OpenCL only. This option is added for compatibility with OpenCL 1.0.
2767
2768.. option:: -cl-uniform-work-group-size
2769
2770OpenCL only. Defines that the global work-size be a multiple of the work-group size specified to clEnqueueNDRangeKernel
2771
2772.. option:: -cl-unsafe-math-optimizations
2773
2774OpenCL only. Allow unsafe floating-point optimizations.  Also implies -cl-no-signed-zeros and -cl-mad-enable.
2775
2776SYCL flags
2777----------
2778.. option:: -fsycl, -fno-sycl
2779
2780Enables SYCL kernels compilation for device
2781
2782.. option:: -sycl-std=<arg>
2783
2784SYCL language standard to compile for.
2785
2786Target-dependent compilation options
2787~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2788.. option:: -G<size>, -G=<arg>, -msmall-data-limit=<arg>, -msmall-data-threshold=<arg>
2789
2790Put objects of at most <size> bytes into small data section (MIPS / Hexagon)
2791
2792.. option:: -ffixed-x1
2793
2794Reserve the x1 register (AArch64/RISC-V only)
2795
2796.. option:: -ffixed-x10
2797
2798Reserve the x10 register (AArch64/RISC-V only)
2799
2800.. option:: -ffixed-x11
2801
2802Reserve the x11 register (AArch64/RISC-V only)
2803
2804.. option:: -ffixed-x12
2805
2806Reserve the x12 register (AArch64/RISC-V only)
2807
2808.. option:: -ffixed-x13
2809
2810Reserve the x13 register (AArch64/RISC-V only)
2811
2812.. option:: -ffixed-x14
2813
2814Reserve the x14 register (AArch64/RISC-V only)
2815
2816.. option:: -ffixed-x15
2817
2818Reserve the x15 register (AArch64/RISC-V only)
2819
2820.. option:: -ffixed-x16
2821
2822Reserve the x16 register (AArch64/RISC-V only)
2823
2824.. option:: -ffixed-x17
2825
2826Reserve the x17 register (AArch64/RISC-V only)
2827
2828.. option:: -ffixed-x18
2829
2830Reserve the x18 register (AArch64/RISC-V only)
2831
2832.. option:: -ffixed-x19
2833
2834Reserve the x19 register (AArch64/RISC-V only)
2835
2836.. option:: -ffixed-x2
2837
2838Reserve the x2 register (AArch64/RISC-V only)
2839
2840.. option:: -ffixed-x20
2841
2842Reserve the x20 register (AArch64/RISC-V only)
2843
2844.. option:: -ffixed-x21
2845
2846Reserve the x21 register (AArch64/RISC-V only)
2847
2848.. option:: -ffixed-x22
2849
2850Reserve the x22 register (AArch64/RISC-V only)
2851
2852.. option:: -ffixed-x23
2853
2854Reserve the x23 register (AArch64/RISC-V only)
2855
2856.. option:: -ffixed-x24
2857
2858Reserve the x24 register (AArch64/RISC-V only)
2859
2860.. option:: -ffixed-x25
2861
2862Reserve the x25 register (AArch64/RISC-V only)
2863
2864.. option:: -ffixed-x26
2865
2866Reserve the x26 register (AArch64/RISC-V only)
2867
2868.. option:: -ffixed-x27
2869
2870Reserve the x27 register (AArch64/RISC-V only)
2871
2872.. option:: -ffixed-x28
2873
2874Reserve the x28 register (AArch64/RISC-V only)
2875
2876.. option:: -ffixed-x29
2877
2878Reserve the x29 register (AArch64/RISC-V only)
2879
2880.. option:: -ffixed-x3
2881
2882Reserve the x3 register (AArch64/RISC-V only)
2883
2884.. option:: -ffixed-x30
2885
2886Reserve the x30 register (AArch64/RISC-V only)
2887
2888.. option:: -ffixed-x31
2889
2890Reserve the x31 register (AArch64/RISC-V only)
2891
2892.. option:: -ffixed-x4
2893
2894Reserve the x4 register (AArch64/RISC-V only)
2895
2896.. option:: -ffixed-x5
2897
2898Reserve the x5 register (AArch64/RISC-V only)
2899
2900.. option:: -ffixed-x6
2901
2902Reserve the x6 register (AArch64/RISC-V only)
2903
2904.. option:: -ffixed-x7
2905
2906Reserve the x7 register (AArch64/RISC-V only)
2907
2908.. option:: -ffixed-x8
2909
2910Reserve the x8 register (AArch64/RISC-V only)
2911
2912.. option:: -ffixed-x9
2913
2914Reserve the x9 register (AArch64/RISC-V only)
2915
2916.. option:: -ffuchsia-api-level=<arg>
2917
2918Set Fuchsia API level
2919
2920.. option:: -inline-asm=<arg>
2921
2922.. option:: -m16
2923
2924.. option:: -m32
2925
2926.. option:: -m64
2927
2928.. option:: -mabi=<arg>
2929
2930.. program:: clang1
2931.. option:: -mabi=vec-default
2932.. program:: clang
2933
2934Enable the default Altivec ABI on AIX (AIX only). Uses only volatile vector registers.
2935
2936.. program:: clang2
2937.. option:: -mabi=vec-extabi
2938.. program:: clang
2939
2940Enable the extended Altivec ABI on AIX (AIX only). Uses volatile and nonvolatile vector registers
2941
2942.. option:: -maix-struct-return
2943
2944Return all structs in memory (PPC32 only)
2945
2946.. option:: -malign-branch-boundary=<arg>
2947
2948Specify the boundary's size to align branches
2949
2950.. option:: -malign-branch=<arg1>,<arg2>...
2951
2952Specify types of branches to align
2953
2954.. option:: -malign-double
2955
2956Align doubles to two words in structs (x86 only)
2957
2958.. option:: -mamdgpu-ieee, -mno-amdgpu-ieee
2959
2960Sets the IEEE bit in the expected default floating point  mode register. Floating point opcodes that support exception flag gathering quiet and propagate signaling NaN inputs per IEEE 754-2008. This option changes the ABI. (AMDGPU only)
2961
2962.. option:: -march=<arg>
2963
2964.. option:: -masm=<arg>
2965
2966.. option:: -mbackchain, -mno-backchain
2967
2968Link stack frames through backchain on System Z
2969
2970.. option:: -mbranches-within-32B-boundaries
2971
2972Align selected branches (fused, jcc, jmp) within 32-byte boundary
2973
2974.. option:: -mcmodel=<arg>, -mcmodel=medany (equivalent to -mcmodel=medium), -mcmodel=medlow (equivalent to -mcmodel=small)
2975
2976.. option:: -mcode-object-v3, -mno-code-object-v3
2977
2978Legacy option to specify code object ABI V3 (AMDGPU only)
2979
2980.. option:: -mcode-object-version=<arg>
2981
2982Specify code object ABI version. Allowed values are 2, 3, 4, and 5. Defaults to 4. (AMDGPU only)
2983
2984.. option:: -mconsole<arg>
2985
2986.. program:: clang1
2987.. option:: -mcpu=<arg>, -mv5 (equivalent to -mcpu=hexagonv5), -mv55 (equivalent to -mcpu=hexagonv55), -mv60 (equivalent to -mcpu=hexagonv60), -mv62 (equivalent to -mcpu=hexagonv62), -mv65 (equivalent to -mcpu=hexagonv65), -mv66 (equivalent to -mcpu=hexagonv66), -mv67 (equivalent to -mcpu=hexagonv67), -mv67t (equivalent to -mcpu=hexagonv67t), -mv68 (equivalent to -mcpu=hexagonv68), -mv69 (equivalent to -mcpu=hexagonv69)
2988.. program:: clang
2989
2990.. option:: -mcrc, -mno-crc
2991
2992Allow use of CRC instructions (ARM/Mips only)
2993
2994.. option:: -mdefault-build-attributes<arg>, -mno-default-build-attributes<arg>
2995
2996.. option:: -mdll<arg>
2997
2998.. option:: -mdouble=<arg>
2999
3000Force double to be 32 bits or 64 bits
3001
3002.. option:: -mdynamic-no-pic<arg>
3003
3004.. option:: -meabi <arg>
3005
3006Set EABI type, e.g. 4, 5 or gnu (default depends on triple)
3007
3008.. option:: -menable-experimental-extensions
3009
3010Enable use of experimental RISC-V extensions.
3011
3012.. option:: -mfentry
3013
3014Insert calls to fentry at function entry (x86/SystemZ only)
3015
3016.. option:: -mfloat-abi=<arg>
3017
3018.. option:: -mfpmath=<arg>
3019
3020.. option:: -mfpu=<arg>
3021
3022.. option:: -mgeneral-regs-only
3023
3024Generate code which only uses the general purpose registers (AArch64/x86 only)
3025
3026.. option:: -mglobal-merge, -mno-global-merge
3027
3028Enable merging of globals
3029
3030.. option:: -mhard-float
3031
3032.. option:: -mhwdiv=<arg>, --mhwdiv <arg>, --mhwdiv=<arg>
3033
3034.. option:: -mhwmult=<arg>
3035
3036.. option:: -miamcu, -mno-iamcu
3037
3038Use Intel MCU ABI
3039
3040.. option:: -mibt-seal
3041
3042Optimize fcf-protection=branch/full (requires LTO).
3043
3044.. option:: -mignore-xcoff-visibility
3045
3046Not emit the visibility attribute for asm in AIX OS or give all symbols 'unspecified' visibility in XCOFF object file
3047
3048.. option:: -mimplicit-float, -mno-implicit-float
3049
3050.. option:: -mimplicit-it=<arg>
3051
3052.. option:: -mincremental-linker-compatible, -mno-incremental-linker-compatible
3053
3054(integrated-as) Emit an object file which can be used with an incremental linker
3055
3056.. option:: -miphoneos-version-min=<arg>, -mios-version-min=<arg>
3057
3058.. option:: -mkernel
3059
3060.. option:: -mlong-calls, -mno-long-calls
3061
3062Generate branches with extended addressability, usually via indirect jumps.
3063
3064.. option:: -mlvi-cfi, -mno-lvi-cfi
3065
3066Enable only control-flow mitigations for Load Value Injection (LVI)
3067
3068.. option:: -mlvi-hardening, -mno-lvi-hardening
3069
3070Enable all mitigations for Load Value Injection (LVI)
3071
3072.. option:: -mmacosx-version-min=<arg>, -mmacos-version-min=<arg>
3073
3074Set Mac OS X deployment target
3075
3076.. option:: -mmcu=<arg>
3077
3078.. option:: -mms-bitfields, -mno-ms-bitfields
3079
3080Set the default structure layout to be compatible with the Microsoft compiler standard
3081
3082.. option:: -mnop-mcount
3083
3084Generate mcount/\_\_fentry\_\_ calls as nops. To activate they need to be patched in.
3085
3086.. option:: -momit-leaf-frame-pointer, -mno-omit-leaf-frame-pointer
3087
3088Omit frame pointer setup for leaf functions
3089
3090.. option:: -moslib=<arg>
3091
3092.. option:: -mpacked-stack, -mno-packed-stack
3093
3094Use packed stack layout (SystemZ only).
3095
3096.. option:: -mpad-max-prefix-size=<arg>
3097
3098Specify maximum number of prefixes to use for padding
3099
3100.. option:: -mprefer-vector-width=<arg>
3101
3102Specifies preferred vector width for auto-vectorization. Defaults to 'none' which allows target specific decisions.
3103
3104.. option:: -mqdsp6-compat
3105
3106Enable hexagon-qdsp6 backward compatibility
3107
3108.. option:: -mrecip
3109
3110.. program:: clang1
3111.. option:: -mrecip=<arg1>,<arg2>...
3112.. program:: clang
3113
3114.. option:: -mrecord-mcount
3115
3116Generate a \_\_mcount\_loc section entry for each \_\_fentry\_\_ call.
3117
3118.. option:: -mred-zone, -mno-red-zone
3119
3120.. option:: -mregparm=<arg>
3121
3122.. option:: -mrelax, -mno-relax
3123
3124Enable linker relaxation
3125
3126.. option:: -mrelax-all, -mno-relax-all
3127
3128(integrated-as) Relax all machine instructions
3129
3130.. option:: -mretpoline, -mno-retpoline
3131
3132.. option:: -mrtd, -mno-rtd
3133
3134Make StdCall calling convention the default
3135
3136.. option:: -mseses, -mno-seses
3137
3138Enable speculative execution side effect suppression (SESES). Includes LVI control flow integrity mitigations
3139
3140.. option:: -msign-return-address=<arg>
3141
3142Select return address signing scope
3143
3144.. option:: -msim
3145
3146.. option:: -mskip-rax-setup, -mno-skip-rax-setup
3147
3148Skip setting up RAX register when passing variable arguments (x86 only)
3149
3150.. option:: -msoft-float, -mno-soft-float
3151
3152Use software floating point
3153
3154.. option:: -mspeculative-load-hardening, -mno-speculative-load-hardening
3155
3156.. option:: -mstack-alignment=<arg>
3157
3158Set the stack alignment
3159
3160.. option:: -mstack-arg-probe, -mno-stack-arg-probe
3161
3162Enable stack probes
3163
3164.. option:: -mstack-probe-size=<arg>
3165
3166Set the stack probe size
3167
3168.. option:: -mstack-protector-guard-offset=<arg>
3169
3170Use the given offset for addressing the stack-protector guard
3171
3172.. option:: -mstack-protector-guard-reg=<arg>
3173
3174Use the given reg for addressing the stack-protector guard
3175
3176.. option:: -mstack-protector-guard=<arg>
3177
3178Use the given guard (global, tls) for addressing the stack-protector guard
3179
3180.. option:: -mstackrealign, -mno-stackrealign
3181
3182Force realign the stack at entry to every function
3183
3184.. option:: -msvr4-struct-return
3185
3186Return small structs in registers (PPC32 only)
3187
3188.. option:: -mtargetos=<arg>
3189
3190Set the deployment target to be the specified OS and OS version
3191
3192.. option:: -mthread-model <arg>
3193
3194The thread model to use, e.g. posix, single (posix by default)
3195
3196.. option:: -mthreads<arg>
3197
3198.. option:: -mthumb, -mno-thumb
3199
3200.. option:: -mtls-direct-seg-refs, -mno-tls-direct-seg-refs
3201
3202Enable direct TLS access through segment registers (default)
3203
3204.. option:: -mtls-size=<arg>
3205
3206Specify bit size of immediate TLS offsets (AArch64 ELF only): 12 (for 4KB) \| 24 (for 16MB, default) \| 32 (for 4GB) \| 48 (for 256TB, needs -mcmodel=large)
3207
3208.. program:: clang1
3209.. option:: -mtune=<arg>
3210.. program:: clang
3211
3212Only supported on X86 and RISC-V. Otherwise accepted for compatibility with GCC.
3213
3214.. option:: -mtvos-version-min=<arg>, -mappletvos-version-min=<arg>
3215
3216.. option:: -municode<arg>
3217
3218.. option:: -munsafe-fp-atomics, -mno-unsafe-fp-atomics
3219
3220Enable unsafe floating point atomic instructions (AMDGPU only)
3221
3222.. option:: -mvx, -mno-vx
3223
3224.. option:: -mwarn-nonportable-cfstrings, -mno-warn-nonportable-cfstrings
3225
3226.. option:: -mwatchos-version-min=<arg>
3227
3228.. option:: -mwavefrontsize64, -mno-wavefrontsize64
3229
3230Specify wavefront size 64 mode (AMDGPU only)
3231
3232.. option:: -mwindows<arg>
3233
3234.. option:: -mx32
3235
3236AARCH64
3237-------
3238.. option:: -fcall-saved-x10
3239
3240Make the x10 register call-saved (AArch64 only)
3241
3242.. option:: -fcall-saved-x11
3243
3244Make the x11 register call-saved (AArch64 only)
3245
3246.. option:: -fcall-saved-x12
3247
3248Make the x12 register call-saved (AArch64 only)
3249
3250.. option:: -fcall-saved-x13
3251
3252Make the x13 register call-saved (AArch64 only)
3253
3254.. option:: -fcall-saved-x14
3255
3256Make the x14 register call-saved (AArch64 only)
3257
3258.. option:: -fcall-saved-x15
3259
3260Make the x15 register call-saved (AArch64 only)
3261
3262.. option:: -fcall-saved-x18
3263
3264Make the x18 register call-saved (AArch64 only)
3265
3266.. option:: -fcall-saved-x8
3267
3268Make the x8 register call-saved (AArch64 only)
3269
3270.. option:: -fcall-saved-x9
3271
3272Make the x9 register call-saved (AArch64 only)
3273
3274.. option:: -mfix-cortex-a53-835769, -mno-fix-cortex-a53-835769
3275
3276Workaround Cortex-A53 erratum 835769 (AArch64 only)
3277
3278.. option:: -mmark-bti-property
3279
3280Add .note.gnu.property with BTI to assembly files (AArch64 only)
3281
3282.. option:: -msve-vector-bits=<arg>
3283
3284Specify the size in bits of an SVE vector register. Defaults to the vector length agnostic value of "scalable". (AArch64 only)
3285
3286.. option:: -mvscale-max=<arg>
3287
3288Specify the vscale maximum. Defaults to the vector length agnostic value of "0". (AArch64 only)
3289
3290.. option:: -mvscale-min=<arg>
3291
3292Specify the vscale minimum. Defaults to "1". (AArch64 only)
3293
3294AMDGPU
3295------
3296.. option:: -mcumode, -mno-cumode
3297
3298Specify CU wavefront execution mode (AMDGPU only)
3299
3300.. option:: -mtgsplit, -mno-tgsplit
3301
3302Enable threadgroup split execution mode (AMDGPU only)
3303
3304ARM
3305---
3306.. option:: -faapcs-bitfield-load
3307
3308Follows the AAPCS standard that all volatile bit-field write generates at least one load. (ARM only).
3309
3310.. option:: -faapcs-bitfield-width, -fno-aapcs-bitfield-width
3311
3312Follow the AAPCS standard requirement stating that volatile bit-field width is dictated by the field container type. (ARM only).
3313
3314.. option:: -ffixed-r9
3315
3316Reserve the r9 register (ARM only)
3317
3318.. option:: -mcmse
3319
3320Allow use of CMSE (Armv8-M Security Extensions)
3321
3322.. option:: -mexecute-only, -mno-execute-only, -mpure-code
3323
3324Disallow generation of data access to code sections (ARM only)
3325
3326.. option:: -mfix-cmse-cve-2021-35465, -mno-fix-cmse-cve-2021-35465
3327
3328Work around VLLDM erratum CVE-2021-35465 (ARM only)
3329
3330.. option:: -mno-bti-at-return-twice
3331
3332Do not add a BTI instruction after a setjmp or other return-twice construct (Arm only)
3333
3334.. option:: -mno-movt
3335
3336Disallow use of movt/movw pairs (ARM only)
3337
3338.. option:: -mno-neg-immediates
3339
3340Disallow converting instructions with negative immediates to their negation or inversion.
3341
3342.. option:: -mnocrc
3343
3344Disallow use of CRC instructions (ARM only)
3345
3346.. option:: -mrestrict-it, -mno-restrict-it
3347
3348Disallow generation of complex IT blocks.
3349
3350.. option:: -mtp=<arg>
3351
3352Thread pointer access method (AArch32/AArch64 only)
3353
3354.. option:: -munaligned-access, -mno-unaligned-access
3355
3356Allow memory accesses to be unaligned (AArch32/AArch64 only)
3357
3358Hexagon
3359-------
3360.. option:: -mhvx-ieee-fp, -mno-hvx-ieee-fp
3361
3362Enable Hexagon HVX IEEE floating-point
3363
3364.. option:: -mieee-rnd-near
3365
3366.. option:: -mmemops, -mno-memops
3367
3368Enable generation of memop instructions
3369
3370.. option:: -mnvj, -mno-nvj
3371
3372Enable generation of new-value jumps
3373
3374.. option:: -mnvs, -mno-nvs
3375
3376Enable generation of new-value stores
3377
3378.. option:: -mpackets, -mno-packets
3379
3380Enable generation of instruction packets
3381
3382Hexagon
3383-------
3384.. option:: -mhvx, -mno-hvx
3385
3386Enable Hexagon Vector eXtensions
3387
3388.. option:: -mhvx-length=<arg>
3389
3390Set Hexagon Vector Length
3391
3392.. option:: -mhvx-qfloat, -mno-hvx-qfloat
3393
3394Enable Hexagon HVX QFloat instructions
3395
3396.. program:: clang1
3397.. option:: -mhvx=<arg>
3398.. program:: clang
3399
3400Enable Hexagon Vector eXtensions
3401
3402M68k
3403----
3404.. option:: -ffixed-a0
3405
3406Reserve the a0 register (M68k only)
3407
3408.. option:: -ffixed-a1
3409
3410Reserve the a1 register (M68k only)
3411
3412.. option:: -ffixed-a2
3413
3414Reserve the a2 register (M68k only)
3415
3416.. option:: -ffixed-a3
3417
3418Reserve the a3 register (M68k only)
3419
3420.. option:: -ffixed-a4
3421
3422Reserve the a4 register (M68k only)
3423
3424.. option:: -ffixed-a5
3425
3426Reserve the a5 register (M68k only)
3427
3428.. option:: -ffixed-a6
3429
3430Reserve the a6 register (M68k only)
3431
3432.. option:: -ffixed-d0
3433
3434Reserve the d0 register (M68k only)
3435
3436.. option:: -ffixed-d1
3437
3438Reserve the d1 register (M68k only)
3439
3440.. option:: -ffixed-d2
3441
3442Reserve the d2 register (M68k only)
3443
3444.. option:: -ffixed-d3
3445
3446Reserve the d3 register (M68k only)
3447
3448.. option:: -ffixed-d4
3449
3450Reserve the d4 register (M68k only)
3451
3452.. option:: -ffixed-d5
3453
3454Reserve the d5 register (M68k only)
3455
3456.. option:: -ffixed-d6
3457
3458Reserve the d6 register (M68k only)
3459
3460.. option:: -ffixed-d7
3461
3462Reserve the d7 register (M68k only)
3463
3464.. option:: -m68000
3465
3466.. option:: -m68010
3467
3468.. option:: -m68020
3469
3470.. option:: -m68030
3471
3472.. option:: -m68040
3473
3474.. option:: -m68060
3475
3476MIPS
3477----
3478.. option:: -mabicalls, -mno-abicalls
3479
3480Enable SVR4-style position-independent code (Mips only)
3481
3482.. option:: -mabs=<arg>
3483
3484.. option:: -mcheck-zero-division, -mno-check-zero-division
3485
3486.. option:: -mcompact-branches=<arg>
3487
3488.. option:: -mdouble-float
3489
3490.. option:: -mdsp, -mno-dsp
3491
3492.. option:: -mdspr2, -mno-dspr2
3493
3494.. option:: -membedded-data, -mno-embedded-data
3495
3496Place constants in the .rodata section instead of the .sdata section even if they meet the -G <size> threshold (MIPS)
3497
3498.. option:: -mextern-sdata, -mno-extern-sdata
3499
3500Assume that externally defined data is in the small data if it meets the -G <size> threshold (MIPS)
3501
3502.. option:: -mfix4300
3503
3504.. option:: -mfp32
3505
3506Use 32-bit floating point registers (MIPS only)
3507
3508.. option:: -mfp64
3509
3510Use 64-bit floating point registers (MIPS only)
3511
3512.. option:: -mginv, -mno-ginv
3513
3514.. option:: -mgpopt, -mno-gpopt
3515
3516Use GP relative accesses for symbols known to be in a small data section (MIPS)
3517
3518.. option:: -mindirect-jump=<arg>
3519
3520Change indirect jump instructions to inhibit speculation
3521
3522.. option:: -mips16
3523
3524.. option:: -mldc1-sdc1, -mno-ldc1-sdc1
3525
3526.. option:: -mlocal-sdata, -mno-local-sdata
3527
3528Extend the -G behaviour to object local data (MIPS)
3529
3530.. option:: -mmadd4, -mno-madd4
3531
3532Enable the generation of 4-operand madd.s, madd.d and related instructions.
3533
3534.. option:: -mmicromips, -mno-micromips
3535
3536.. option:: -mmsa, -mno-msa
3537
3538Enable MSA ASE (MIPS only)
3539
3540.. option:: -mmt, -mno-mt
3541
3542Enable MT ASE (MIPS only)
3543
3544.. option:: -mnan=<arg>
3545
3546.. option:: -mno-mips16
3547
3548.. option:: -msingle-float
3549
3550.. option:: -mvirt, -mno-virt
3551
3552.. option:: -mxgot, -mno-xgot
3553
3554PowerPC
3555-------
3556.. option:: -maltivec, -mno-altivec
3557
3558.. option:: -mcmpb, -mno-cmpb
3559
3560.. option:: -mcrbits, -mno-crbits
3561
3562.. option:: -mcrypto, -mno-crypto
3563
3564.. option:: -mdirect-move, -mno-direct-move
3565
3566.. option:: -mefpu2
3567
3568.. option:: -mfloat128, -mno-float128
3569
3570.. option:: -mfprnd, -mno-fprnd
3571
3572.. option:: -mhtm, -mno-htm
3573
3574.. option:: -minvariant-function-descriptors, -mno-invariant-function-descriptors
3575
3576.. option:: -misel, -mno-isel
3577
3578.. option:: -mlongcall, -mno-longcall
3579
3580.. option:: -mmfocrf, -mmfcrf, -mno-mfocrf
3581
3582.. option:: -mmma, -mno-mma
3583
3584.. option:: -mpaired-vector-memops, -mno-paired-vector-memops
3585
3586.. option:: -mpcrel, -mno-pcrel
3587
3588.. option:: -mpopcntd, -mno-popcntd
3589
3590.. option:: -mpower10-vector, -mno-power10-vector
3591
3592.. option:: -mpower8-vector, -mno-power8-vector
3593
3594.. option:: -mpower9-vector, -mno-power9-vector
3595
3596.. option:: -mprefixed, -mno-prefixed
3597
3598.. option:: -mprivileged
3599
3600.. option:: -mrop-protect
3601
3602.. option:: -msecure-plt
3603
3604.. option:: -mspe, -mno-spe
3605
3606.. option:: -mvsx, -mno-vsx
3607
3608WebAssembly
3609-----------
3610.. option:: -matomics, -mno-atomics
3611
3612.. option:: -mbulk-memory, -mno-bulk-memory
3613
3614.. option:: -mexception-handling, -mno-exception-handling
3615
3616.. option:: -mmultivalue, -mno-multivalue
3617
3618.. option:: -mmutable-globals, -mno-mutable-globals
3619
3620.. option:: -mnontrapping-fptoint, -mno-nontrapping-fptoint
3621
3622.. option:: -mreference-types, -mno-reference-types
3623
3624.. option:: -mrelaxed-simd, -mno-relaxed-simd
3625
3626.. option:: -msign-ext, -mno-sign-ext
3627
3628.. option:: -msimd128, -mno-simd128
3629
3630.. option:: -mtail-call, -mno-tail-call
3631
3632.. option:: -mextended-const, -mno-extended-const
3633
3634WebAssembly Driver
3635------------------
3636.. option:: -mexec-model=<arg>
3637
3638Execution model (WebAssembly only)
3639
3640X86
3641---
3642.. option:: -m3dnow, -mno-3dnow
3643
3644.. option:: -m3dnowa, -mno-3dnowa
3645
3646.. option:: -madx, -mno-adx
3647
3648.. option:: -maes, -mno-aes
3649
3650.. option:: -mamx-bf16, -mno-amx-bf16
3651
3652.. option:: -mamx-int8, -mno-amx-int8
3653
3654.. option:: -mamx-tile, -mno-amx-tile
3655
3656.. option:: -mavx, -mno-avx
3657
3658.. option:: -mavx2, -mno-avx2
3659
3660.. option:: -mavx512bf16, -mno-avx512bf16
3661
3662.. option:: -mavx512bitalg, -mno-avx512bitalg
3663
3664.. option:: -mavx512bw, -mno-avx512bw
3665
3666.. option:: -mavx512cd, -mno-avx512cd
3667
3668.. option:: -mavx512dq, -mno-avx512dq
3669
3670.. option:: -mavx512er, -mno-avx512er
3671
3672.. option:: -mavx512f, -mno-avx512f
3673
3674.. option:: -mavx512fp16, -mno-avx512fp16
3675
3676.. option:: -mavx512ifma, -mno-avx512ifma
3677
3678.. option:: -mavx512pf, -mno-avx512pf
3679
3680.. option:: -mavx512vbmi, -mno-avx512vbmi
3681
3682.. option:: -mavx512vbmi2, -mno-avx512vbmi2
3683
3684.. option:: -mavx512vl, -mno-avx512vl
3685
3686.. option:: -mavx512vnni, -mno-avx512vnni
3687
3688.. option:: -mavx512vp2intersect, -mno-avx512vp2intersect
3689
3690.. option:: -mavx512vpopcntdq, -mno-avx512vpopcntdq
3691
3692.. option:: -mavxvnni, -mno-avxvnni
3693
3694.. option:: -mbmi, -mno-bmi
3695
3696.. option:: -mbmi2, -mno-bmi2
3697
3698.. option:: -mcldemote, -mno-cldemote
3699
3700.. option:: -mclflushopt, -mno-clflushopt
3701
3702.. option:: -mclwb, -mno-clwb
3703
3704.. option:: -mclzero, -mno-clzero
3705
3706.. option:: -mcrc32, -mno-crc32
3707
3708.. option:: -mcx16, -mno-cx16
3709
3710.. option:: -menqcmd, -mno-enqcmd
3711
3712.. option:: -mf16c, -mno-f16c
3713
3714.. option:: -mfma, -mno-fma
3715
3716.. option:: -mfma4, -mno-fma4
3717
3718.. option:: -mfsgsbase, -mno-fsgsbase
3719
3720.. option:: -mfxsr, -mno-fxsr
3721
3722.. option:: -mgfni, -mno-gfni
3723
3724.. option:: -mhreset, -mno-hreset
3725
3726.. option:: -minvpcid, -mno-invpcid
3727
3728.. option:: -mkl, -mno-kl
3729
3730.. option:: -mlwp, -mno-lwp
3731
3732.. option:: -mlzcnt, -mno-lzcnt
3733
3734.. option:: -mmmx, -mno-mmx
3735
3736.. option:: -mmovbe, -mno-movbe
3737
3738.. option:: -mmovdir64b, -mno-movdir64b
3739
3740.. option:: -mmovdiri, -mno-movdiri
3741
3742.. option:: -mmwaitx, -mno-mwaitx
3743
3744.. option:: -mpclmul, -mno-pclmul
3745
3746.. option:: -mpconfig, -mno-pconfig
3747
3748.. option:: -mpku, -mno-pku
3749
3750.. option:: -mpopcnt, -mno-popcnt
3751
3752.. option:: -mprefetchwt1, -mno-prefetchwt1
3753
3754.. option:: -mprfchw, -mno-prfchw
3755
3756.. option:: -mptwrite, -mno-ptwrite
3757
3758.. option:: -mrdpid, -mno-rdpid
3759
3760.. option:: -mrdrnd, -mno-rdrnd
3761
3762.. option:: -mrdseed, -mno-rdseed
3763
3764.. option:: -mretpoline-external-thunk, -mno-retpoline-external-thunk
3765
3766.. option:: -mrtm, -mno-rtm
3767
3768.. option:: -msahf, -mno-sahf
3769
3770.. option:: -mserialize, -mno-serialize
3771
3772.. option:: -msgx, -mno-sgx
3773
3774.. option:: -msha, -mno-sha
3775
3776.. option:: -mshstk, -mno-shstk
3777
3778.. option:: -msse, -mno-sse
3779
3780.. option:: -msse2, -mno-sse2
3781
3782.. option:: -msse3, -mno-sse3
3783
3784.. option:: -msse4.1, -mno-sse4.1
3785
3786.. program:: clang1
3787.. option:: -msse4.2, -mno-sse4.2, -msse4
3788.. program:: clang
3789
3790.. option:: -msse4a, -mno-sse4a
3791
3792.. option:: -mssse3, -mno-ssse3
3793
3794.. option:: -mtbm, -mno-tbm
3795
3796.. option:: -mtsxldtrk, -mno-tsxldtrk
3797
3798.. option:: -muintr, -mno-uintr
3799
3800.. option:: -mvaes, -mno-vaes
3801
3802.. option:: -mvpclmulqdq, -mno-vpclmulqdq
3803
3804.. option:: -mvzeroupper, -mno-vzeroupper
3805
3806.. option:: -mwaitpkg, -mno-waitpkg
3807
3808.. option:: -mwbnoinvd, -mno-wbnoinvd
3809
3810.. option:: -mwidekl, -mno-widekl
3811
3812.. option:: -mx87, -m80387, -mno-x87
3813
3814.. option:: -mxop, -mno-xop
3815
3816.. option:: -mxsave, -mno-xsave
3817
3818.. option:: -mxsavec, -mno-xsavec
3819
3820.. option:: -mxsaveopt, -mno-xsaveopt
3821
3822.. option:: -mxsaves, -mno-xsaves
3823
3824RISCV
3825-----
3826.. option:: -msave-restore, -mno-save-restore
3827
3828Enable using library calls for save and restore
3829
3830Long double flags
3831-----------------
3832Selects the long double implementation
3833
3834.. option:: -mlong-double-128
3835
3836Force long double to be 128 bits
3837
3838.. option:: -mlong-double-64
3839
3840Force long double to be 64 bits
3841
3842.. option:: -mlong-double-80
3843
3844Force long double to be 80 bits, padded to 128 bits for storage
3845
3846Optimization level
3847~~~~~~~~~~~~~~~~~~
3848
3849Flags controlling how much optimization should be performed.
3850
3851.. option:: -O<arg>, -O (equivalent to -O1), --optimize, --optimize=<arg>
3852
3853.. option:: -Ofast<arg>
3854
3855Debug information generation
3856~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3857
3858Flags controlling how much and what kind of debug information should be
3859generated.
3860
3861Kind and level of debug information
3862-----------------------------------
3863.. option:: -g, --debug, --debug=<arg>
3864
3865Generate source-level debug information
3866
3867.. option:: -gdwarf
3868
3869Generate source-level debug information with the default dwarf version
3870
3871.. option:: -gdwarf-2
3872
3873Generate source-level debug information with dwarf version 2
3874
3875.. option:: -gdwarf-3
3876
3877Generate source-level debug information with dwarf version 3
3878
3879.. option:: -gdwarf-4
3880
3881Generate source-level debug information with dwarf version 4
3882
3883.. option:: -gdwarf-5
3884
3885Generate source-level debug information with dwarf version 5
3886
3887.. option:: -gdwarf32
3888
3889Enables DWARF32 format for ELF binaries, if debug information emission is enabled.
3890
3891.. option:: -gdwarf64
3892
3893Enables DWARF64 format for ELF binaries, if debug information emission is enabled.
3894
3895.. option:: -gfull
3896
3897.. option:: -ginline-line-tables, -gno-inline-line-tables
3898
3899.. option:: -gused
3900
3901Debug level
3902___________
3903.. option:: -g0
3904
3905.. option:: -g2
3906
3907.. option:: -g3
3908
3909.. option:: -ggdb0
3910
3911.. option:: -ggdb1
3912
3913.. option:: -ggdb2
3914
3915.. option:: -ggdb3
3916
3917.. option:: -gline-directives-only
3918
3919Emit debug line info directives only
3920
3921.. option:: -gline-tables-only, -g1, -gmlt
3922
3923Emit debug line number tables only
3924
3925.. option:: -gmodules
3926
3927Generate debug info with external references to clang modules or precompiled headers
3928
3929Debugger to tune debug information for
3930______________________________________
3931.. option:: -gdbx
3932
3933.. option:: -ggdb
3934
3935.. option:: -glldb
3936
3937.. option:: -gsce
3938
3939Debug information flags
3940-----------------------
3941.. option:: -gcolumn-info, -gno-column-info
3942
3943.. option:: -gdwarf-aranges
3944
3945.. option:: -gembed-source, -gno-embed-source
3946
3947Embed source text in DWARF debug sections
3948
3949.. option:: -ggnu-pubnames, -gno-gnu-pubnames
3950
3951.. option:: -gpubnames, -gno-pubnames
3952
3953.. option:: -grecord-command-line, -gno-record-command-line, -grecord-gcc-switches
3954
3955.. option:: -gsimple-template-names, -gno-simple-template-names
3956
3957.. option:: -gsplit-dwarf, -gno-split-dwarf
3958
3959.. program:: clang1
3960.. option:: -gsplit-dwarf=<arg>
3961.. program:: clang
3962
3963Set DWARF fission mode to either 'split' or 'single'
3964
3965.. option:: -gstrict-dwarf, -gno-strict-dwarf
3966
3967.. option:: -gz=<arg>, -gz (equivalent to -gz=zlib)
3968
3969DWARF debug sections compression type
3970
3971Static analyzer flags
3972=====================
3973
3974Flags controlling the behavior of the Clang Static Analyzer.
3975
3976.. option:: -Xanalyzer <arg>
3977
3978Pass <arg> to the static analyzer
3979
3980Fortran compilation flags
3981=========================
3982
3983Flags that will be passed onto the ``gfortran`` compiler when Clang is given
3984a Fortran input.
3985
3986.. option:: -A<arg>, --assert <arg>, --assert=<arg>
3987
3988.. option:: -A-<arg>
3989
3990.. option:: -faggressive-function-elimination, -fno-aggressive-function-elimination
3991
3992.. option:: -falign-commons, -fno-align-commons
3993
3994.. option:: -fall-intrinsics, -fno-all-intrinsics
3995
3996.. option:: -fbacktrace, -fno-backtrace
3997
3998.. option:: -fblas-matmul-limit=<arg>
3999
4000.. option:: -fbounds-check, -fno-bounds-check
4001
4002.. option:: -fcheck-array-temporaries, -fno-check-array-temporaries
4003
4004.. option:: -fcheck=<arg>
4005
4006.. option:: -fcoarray=<arg>
4007
4008.. option:: -fconvert=<arg>
4009
4010.. option:: -fcray-pointer, -fno-cray-pointer
4011
4012.. option:: -fd-lines-as-code, -fno-d-lines-as-code
4013
4014.. option:: -fd-lines-as-comments, -fno-d-lines-as-comments
4015
4016.. option:: -fdollar-ok, -fno-dollar-ok
4017
4018.. option:: -fdump-fortran-optimized, -fno-dump-fortran-optimized
4019
4020.. option:: -fdump-fortran-original, -fno-dump-fortran-original
4021
4022.. option:: -fdump-parse-tree, -fno-dump-parse-tree
4023
4024.. option:: -fexternal-blas, -fno-external-blas
4025
4026.. option:: -ff2c, -fno-f2c
4027
4028.. option:: -ffpe-trap=<arg>
4029
4030.. option:: -ffree-line-length-<arg>
4031
4032.. option:: -ffrontend-optimize, -fno-frontend-optimize
4033
4034.. option:: -finit-character=<arg>
4035
4036.. option:: -finit-integer=<arg>
4037
4038.. option:: -finit-local-zero, -fno-init-local-zero
4039
4040.. option:: -finit-logical=<arg>
4041
4042.. option:: -finit-real=<arg>
4043
4044.. option:: -finteger-4-integer-8, -fno-integer-4-integer-8
4045
4046.. option:: -fmax-array-constructor=<arg>
4047
4048.. option:: -fmax-errors=<arg>
4049
4050.. option:: -fmax-identifier-length, -fno-max-identifier-length
4051
4052.. option:: -fmax-stack-var-size=<arg>
4053
4054.. option:: -fmax-subrecord-length=<arg>
4055
4056.. option:: -fmodule-private, -fno-module-private
4057
4058.. option:: -fpack-derived, -fno-pack-derived
4059
4060.. option:: -frange-check, -fno-range-check
4061
4062.. option:: -freal-4-real-10, -fno-real-4-real-10
4063
4064.. option:: -freal-4-real-16, -fno-real-4-real-16
4065
4066.. option:: -freal-4-real-8, -fno-real-4-real-8
4067
4068.. option:: -freal-8-real-10, -fno-real-8-real-10
4069
4070.. option:: -freal-8-real-16, -fno-real-8-real-16
4071
4072.. option:: -freal-8-real-4, -fno-real-8-real-4
4073
4074.. option:: -frealloc-lhs, -fno-realloc-lhs
4075
4076.. option:: -frecord-marker=<arg>
4077
4078.. option:: -frecursive, -fno-recursive
4079
4080.. option:: -frepack-arrays, -fno-repack-arrays
4081
4082.. option:: -fsecond-underscore, -fno-second-underscore
4083
4084.. option:: -fsign-zero, -fno-sign-zero
4085
4086.. option:: -fstack-arrays, -fno-stack-arrays
4087
4088.. option:: -funderscoring, -fno-underscoring
4089
4090.. option:: -fwhole-file, -fno-whole-file
4091
4092.. option:: -imultilib <arg>
4093
4094.. option:: -static-libgfortran
4095
4096Linker flags
4097============
4098Flags that are passed on to the linker
4099
4100.. option:: -L<dir>, --library-directory <arg>, --library-directory=<arg>
4101
4102Add directory to library search path
4103
4104.. option:: -Mach
4105
4106.. option:: -T<script>
4107
4108Specify <script> as linker script
4109
4110.. option:: -Tbss<addr>
4111
4112Set starting address of BSS to <addr>
4113
4114.. option:: -Tdata<addr>
4115
4116Set starting address of DATA to <addr>
4117
4118.. option:: -Ttext<addr>
4119
4120Set starting address of TEXT to <addr>
4121
4122.. option:: -Wl,<arg>,<arg2>...
4123
4124Pass the comma separated arguments in <arg> to the linker
4125
4126.. option:: -X
4127
4128.. option:: -Xlinker <arg>, --for-linker <arg>, --for-linker=<arg>
4129
4130Pass <arg> to the linker
4131
4132.. program:: clang1
4133.. option:: -Z
4134.. program:: clang
4135
4136.. option:: -b<arg>
4137
4138Pass -b <arg> to the linker on AIX (only).
4139
4140.. option:: -coverage, --coverage
4141
4142.. option:: -e<arg>, --entry
4143
4144.. option:: -filelist <arg>
4145
4146.. option:: --hip-device-lib=<arg>
4147
4148HIP device library
4149
4150.. option:: --hipspv-pass-plugin=<dsopath>
4151
4152path to a pass plugin for HIP to SPIR-V passes.
4153
4154.. option:: -l<arg>
4155
4156.. option:: --ld-path=<arg>
4157
4158.. option:: -nostartfiles
4159
4160.. program:: clang1
4161.. option:: -nostdlib, --no-standard-libraries
4162.. program:: clang
4163
4164.. option:: -pie
4165
4166.. option:: -r
4167
4168.. option:: -rdynamic
4169
4170.. option:: --rocm-device-lib-path=<arg>, --hip-device-lib-path=<arg>
4171
4172ROCm device library path. Alternative to rocm-path.
4173
4174.. option:: -rpath <arg>
4175
4176.. option:: -s
4177
4178.. option:: -shared, --shared
4179
4180.. option:: -specs=<arg>, --specs=<arg>
4181
4182.. option:: -static, --static
4183
4184.. option:: -static-pie
4185
4186.. option:: -t
4187
4188.. option:: -u<arg>, --force-link <arg>, --force-link=<arg>
4189
4190.. option:: -undef
4191
4192undef all system defines
4193
4194.. option:: -undefined<arg>, --no-undefined
4195
4196.. option:: -z <arg>
4197
4198Pass -z <arg> to the linker
4199
4200