| Message ID | 1394705630-12384-4-git-send-email-takahiro.akashi@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, Mar 13, 2014 at 10:13:46AM +0000, AKASHI Takahiro wrote: > Recordmcount utility under scripts is run, after compiling each object, > to find out all the locations of calling _mcount() and put them into > specific seciton named __mcount_loc. > Then linker collects all such information into a table in the kernel image > (between __start_mcount_loc and __stop_mcount_loc) for later use by ftrace. > > This patch adds arm64 specific definitions to identify such locations. > There are two types of implementation, C and Perl. On arm64, only C version > is used to build the kernel now that CONFIG_HAVE_C_RECORDMCOUNT is on. > But Perl version is also maintained. > > This patch also contains a workaround just in case where a header file, > elf.h, on host machine doesn't have definitions of EM_AARCH64 nor > R_AARCH64_ABS64. Without them, compiling C version of recordmcount will > fail. [...] > diff --git a/scripts/recordmcount.c b/scripts/recordmcount.c > index 9c22317..047c8cd 100644 > --- a/scripts/recordmcount.c > +++ b/scripts/recordmcount.c > @@ -40,6 +40,11 @@ > #define R_METAG_NONE 3 > #endif > > +#ifndef EM_AARCH64 > +#define EM_AARCH64 183 > +#define R_AARCH64_ABS64 257 > +#endif > + > static int fd_map; /* File descriptor for file being modified. */ > static int mmap_failed; /* Boolean flag. */ > static void *ehdr_curr; /* current ElfXX_Ehdr * for resource cleanup */ > @@ -347,6 +352,10 @@ do_file(char const *const fname) > case EM_ARM: reltype = R_ARM_ABS32; > altmcount = "__gnu_mcount_nc"; > break; > + case EM_AARCH64: > + reltype = R_AARCH64_ABS64; > + altmcount = "_mcount"; Shouldn't you be setting gpfx = '_' instead of overriding altmcount like this? > + break; > case EM_IA_64: reltype = R_IA64_IMM64; gpfx = '_'; break; > case EM_METAG: reltype = R_METAG_ADDR32; > altmcount = "_mcount_wrapper"; > diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl > index 91280b8..397b6b8 100755 > --- a/scripts/recordmcount.pl > +++ b/scripts/recordmcount.pl > @@ -279,6 +279,11 @@ if ($arch eq "x86_64") { > $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_ARM_(CALL|PC24|THM_CALL)" . > "\\s+(__gnu_mcount_nc|mcount)\$"; > > +} elsif ($arch eq "arm64") { > + $alignment = 3; > + $section_type = '%progbits'; > + $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_AARCH64_CALL26\\s+_mcount\$"; What about "mcount" (i.e. no underscore)? Will
On 03/13/2014 09:38 PM, Will Deacon wrote: > On Thu, Mar 13, 2014 at 10:13:46AM +0000, AKASHI Takahiro wrote: >> Recordmcount utility under scripts is run, after compiling each object, >> to find out all the locations of calling _mcount() and put them into >> specific seciton named __mcount_loc. >> Then linker collects all such information into a table in the kernel image >> (between __start_mcount_loc and __stop_mcount_loc) for later use by ftrace. >> >> This patch adds arm64 specific definitions to identify such locations. >> There are two types of implementation, C and Perl. On arm64, only C version >> is used to build the kernel now that CONFIG_HAVE_C_RECORDMCOUNT is on. >> But Perl version is also maintained. >> >> This patch also contains a workaround just in case where a header file, >> elf.h, on host machine doesn't have definitions of EM_AARCH64 nor >> R_AARCH64_ABS64. Without them, compiling C version of recordmcount will >> fail. > > [...] > >> diff --git a/scripts/recordmcount.c b/scripts/recordmcount.c >> index 9c22317..047c8cd 100644 >> --- a/scripts/recordmcount.c >> +++ b/scripts/recordmcount.c >> @@ -40,6 +40,11 @@ >> #define R_METAG_NONE 3 >> #endif >> >> +#ifndef EM_AARCH64 >> +#define EM_AARCH64 183 >> +#define R_AARCH64_ABS64 257 >> +#endif >> + >> static int fd_map; /* File descriptor for file being modified. */ >> static int mmap_failed; /* Boolean flag. */ >> static void *ehdr_curr; /* current ElfXX_Ehdr * for resource cleanup */ >> @@ -347,6 +352,10 @@ do_file(char const *const fname) >> case EM_ARM: reltype = R_ARM_ABS32; >> altmcount = "__gnu_mcount_nc"; >> break; >> + case EM_AARCH64: >> + reltype = R_AARCH64_ABS64; >> + altmcount = "_mcount"; > > Shouldn't you be setting gpfx = '_' instead of overriding altmcount like > this? Yep. Fix it. >> + break; >> case EM_IA_64: reltype = R_IA64_IMM64; gpfx = '_'; break; >> case EM_METAG: reltype = R_METAG_ADDR32; >> altmcount = "_mcount_wrapper"; >> diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl >> index 91280b8..397b6b8 100755 >> --- a/scripts/recordmcount.pl >> +++ b/scripts/recordmcount.pl >> @@ -279,6 +279,11 @@ if ($arch eq "x86_64") { >> $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_ARM_(CALL|PC24|THM_CALL)" . >> "\\s+(__gnu_mcount_nc|mcount)\$"; >> >> +} elsif ($arch eq "arm64") { >> + $alignment = 3; >> + $section_type = '%progbits'; >> + $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_AARCH64_CALL26\\s+_mcount\$"; > > What about "mcount" (i.e. no underscore)? I don't think it works since such regex doesn't match to "CALL26 _mcount". Thanks, -Takahiro AKASHI > Will >
On Thu, Mar 13, 2014 at 02:57:58PM +0000, AKASHI Takahiro wrote: > On 03/13/2014 09:38 PM, Will Deacon wrote: > > On Thu, Mar 13, 2014 at 10:13:46AM +0000, AKASHI Takahiro wrote: > >> diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl > >> index 91280b8..397b6b8 100755 > >> --- a/scripts/recordmcount.pl > >> +++ b/scripts/recordmcount.pl > >> @@ -279,6 +279,11 @@ if ($arch eq "x86_64") { > >> $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_ARM_(CALL|PC24|THM_CALL)" . > >> "\\s+(__gnu_mcount_nc|mcount)\$"; > >> > >> +} elsif ($arch eq "arm64") { > >> + $alignment = 3; > >> + $section_type = '%progbits'; > >> + $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_AARCH64_CALL26\\s+_mcount\$"; > > > > What about "mcount" (i.e. no underscore)? > > I don't think it works since such regex doesn't match to "CALL26 _mcount". What I meant was, do you need to match "mcount" as well? However, after speaking to our GCC guys, they say they only support "_mcount" (preferred) and "__mcount" (which GCC itself doesn't emit) so you can leave this regex as-is. Will
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 27bbcfc..340e344 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -27,12 +27,14 @@ config ARM64 select HARDIRQS_SW_RESEND select HAVE_ARCH_JUMP_LABEL select HAVE_ARCH_TRACEHOOK + select HAVE_C_RECORDMCOUNT select HAVE_DEBUG_BUGVERBOSE select HAVE_DEBUG_KMEMLEAK select HAVE_DMA_API_DEBUG select HAVE_DMA_ATTRS select HAVE_DMA_CONTIGUOUS select HAVE_EFFICIENT_UNALIGNED_ACCESS + select HAVE_FTRACE_MCOUNT_RECORD select HAVE_GENERIC_DMA_COHERENT select HAVE_HW_BREAKPOINT if PERF_EVENTS select HAVE_MEMBLOCK diff --git a/scripts/recordmcount.c b/scripts/recordmcount.c index 9c22317..047c8cd 100644 --- a/scripts/recordmcount.c +++ b/scripts/recordmcount.c @@ -40,6 +40,11 @@ #define R_METAG_NONE 3 #endif +#ifndef EM_AARCH64 +#define EM_AARCH64 183 +#define R_AARCH64_ABS64 257 +#endif + static int fd_map; /* File descriptor for file being modified. */ static int mmap_failed; /* Boolean flag. */ static void *ehdr_curr; /* current ElfXX_Ehdr * for resource cleanup */ @@ -347,6 +352,10 @@ do_file(char const *const fname) case EM_ARM: reltype = R_ARM_ABS32; altmcount = "__gnu_mcount_nc"; break; + case EM_AARCH64: + reltype = R_AARCH64_ABS64; + altmcount = "_mcount"; + break; case EM_IA_64: reltype = R_IA64_IMM64; gpfx = '_'; break; case EM_METAG: reltype = R_METAG_ADDR32; altmcount = "_mcount_wrapper"; diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl index 91280b8..397b6b8 100755 --- a/scripts/recordmcount.pl +++ b/scripts/recordmcount.pl @@ -279,6 +279,11 @@ if ($arch eq "x86_64") { $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_ARM_(CALL|PC24|THM_CALL)" . "\\s+(__gnu_mcount_nc|mcount)\$"; +} elsif ($arch eq "arm64") { + $alignment = 3; + $section_type = '%progbits'; + $mcount_regex = "^\\s*([0-9a-fA-F]+):\\s*R_AARCH64_CALL26\\s+_mcount\$"; + $type = ".quad"; } elsif ($arch eq "ia64") { $mcount_regex = "^\\s*([0-9a-fA-F]+):.*\\s_mcount\$"; $type = "data8";
Recordmcount utility under scripts is run, after compiling each object, to find out all the locations of calling _mcount() and put them into specific seciton named __mcount_loc. Then linker collects all such information into a table in the kernel image (between __start_mcount_loc and __stop_mcount_loc) for later use by ftrace. This patch adds arm64 specific definitions to identify such locations. There are two types of implementation, C and Perl. On arm64, only C version is used to build the kernel now that CONFIG_HAVE_C_RECORDMCOUNT is on. But Perl version is also maintained. This patch also contains a workaround just in case where a header file, elf.h, on host machine doesn't have definitions of EM_AARCH64 nor R_AARCH64_ABS64. Without them, compiling C version of recordmcount will fail. Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org> --- arch/arm64/Kconfig | 2 ++ scripts/recordmcount.c | 9 +++++++++ scripts/recordmcount.pl | 5 +++++ 3 files changed, 16 insertions(+)