summaryrefslogtreecommitdiff
path: root/Documentation/bpf/bpf_devel_QA.txt
diff options
context:
space:
mode:
authorJens Axboe <axboe@kernel.dk>2018-02-07 15:54:20 -0700
committerJens Axboe <axboe@kernel.dk>2018-02-07 15:54:20 -0700
commit61a695184fcc8f225327206b88320279dd8684af (patch)
tree6cbda2b0e7de94ebad95072d17cc0aa82c9d8ab3 /Documentation/bpf/bpf_devel_QA.txt
parent9e05c864993c5442227f83ae1694a737d7a102ed (diff)
parent581e400ff935d34d95811258586128bf11baef15 (diff)
Merge branch 'master' into test
* master: (1190 commits) ASoC: stm32: add of dependency for stm32 drivers ASoC: mt8173-rt5650: fix child-node lookup ASoC: dapm: fix debugfs read using path->connected platform/x86: samsung-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro platform/x86: ideapad-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro platform/x86: dell-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro seq_file: Introduce DEFINE_SHOW_ATTRIBUTE() helper macro Documentation/sysctl/user.txt: fix typo MAINTAINERS: update ARM/QUALCOMM SUPPORT patterns MAINTAINERS: update various PALM patterns MAINTAINERS: update "ARM/OXNAS platform support" patterns MAINTAINERS: update Cortina/Gemini patterns MAINTAINERS: remove ARM/CLKDEV SUPPORT file pattern MAINTAINERS: remove ANDROID ION pattern mm: docs: add blank lines to silence sphinx "Unexpected indentation" errors mm: docs: fix parameter names mismatch mm: docs: fixup punctuation pipe: read buffer limits atomically pipe: simplify round_pipe_size() pipe: reject F_SETPIPE_SZ with size over UINT_MAX ...
Diffstat (limited to 'Documentation/bpf/bpf_devel_QA.txt')
-rw-r--r--Documentation/bpf/bpf_devel_QA.txt31
1 files changed, 31 insertions, 0 deletions
diff --git a/Documentation/bpf/bpf_devel_QA.txt b/Documentation/bpf/bpf_devel_QA.txt
index cefef855dea4..84cbb302f2b5 100644
--- a/Documentation/bpf/bpf_devel_QA.txt
+++ b/Documentation/bpf/bpf_devel_QA.txt
@@ -516,4 +516,35 @@ A: LLVM has a -mcpu selector for the BPF back end in order to allow the
By the way, the BPF kernel selftests run with -mcpu=probe for better
test coverage.
+Q: In some cases clang flag "-target bpf" is used but in other cases the
+ default clang target, which matches the underlying architecture, is used.
+ What is the difference and when I should use which?
+
+A: Although LLVM IR generation and optimization try to stay architecture
+ independent, "-target <arch>" still has some impact on generated code:
+
+ - BPF program may recursively include header file(s) with file scope
+ inline assembly codes. The default target can handle this well,
+ while bpf target may fail if bpf backend assembler does not
+ understand these assembly codes, which is true in most cases.
+
+ - When compiled without -g, additional elf sections, e.g.,
+ .eh_frame and .rela.eh_frame, may be present in the object file
+ with default target, but not with bpf target.
+
+ - The default target may turn a C switch statement into a switch table
+ lookup and jump operation. Since the switch table is placed
+ in the global readonly section, the bpf program will fail to load.
+ The bpf target does not support switch table optimization.
+ The clang option "-fno-jump-tables" can be used to disable
+ switch table generation.
+
+ You should use default target when:
+
+ - Your program includes a header file, e.g., ptrace.h, which eventually
+ pulls in some header files containing file scope host assembly codes.
+ - You can add "-fno-jump-tables" to work around the switch table issue.
+
+ Otherwise, you can use bpf target.
+
Happy BPF hacking!