d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
4bfad7cc80
freebsd-skq/sys/amd64/include/profile.h

170 lines
5.0 KiB
C
Raw Normal View History

BSD 4.4 Lite Kernel Sources
1994-05-25 01:34:38 +00:00
/*
* Copyright (c) 1992, 1993
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)profile.h 8.1 (Berkeley) 6/11/93
$Id$ -> $FreeBSD$
1999-08-28 01:08:13 +00:00
* $FreeBSD$
BSD 4.4 Lite Kernel Sources
1994-05-25 01:34:38 +00:00
*/
Implemented non-statistical kernel profiling. This is based on looking at a high resolution clock for each of the following events: function call, function return, interrupt entry, interrupt exit, and interesting branches. The differences between the times of these events are added at appropriate places in a ordinary histogram (as if very fast statistical profiling sampled the pc at those places) so that ordinary gprof can be used to analyze the times. gmon.h: Histogram counters need to be 4 bytes for microsecond resolutions. They will need to be larger for the 586 clock. The comments were vax-centric and wrong even on vaxes. Does anyone disagree? gprof4.c: The standard gprof should support counters of all integral sizes and the size of the counter should be in the gmon header. This hack will do until then. (Use gprof4 -u to examine the results of non-statistical profiling.) config/*: Non-statistical profiling is configured with `config -pp'. `config -p' still gives ordinary profiling. kgmon/*: Non-statistical profiling is enabled with `kgmon -B'. `kgmon -b' still enables ordinary profiling (and distables non-statistical profiling) if non-statistical profiling is configured.
1995-12-29 15:30:05 +00:00
#ifndef _MACHINE_PROFILE_H_
#define _MACHINE_PROFILE_H_
Made idempotent. Reviewed by: Submitted by:
1994-08-21 04:55:31 +00:00
Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL" is an application space macro and the applications are supposed to be free to use it as they please (but cannot). This is consistant with the other BSD's who made this change quite some time ago. More commits to come.
1999-12-29 04:46:21 +00:00
#ifdef _KERNEL
Oops, fptrint_t still needs to be declared in <machine/profile.h> in the !KERNEL case. The kludge to get it declared in libc/gmon/mcount.c wasn't sufficient because fptrint_t is used in <sys/gmon.h>.
1998-07-10 09:26:41 +00:00
Moved definition of FUNCTION_ALIGNMENT to a machine-dependent place. Changed it from 4 to 16 for i386's. It can be anything for i386's, but compiler options limit it to a power of 2, and assembler and linker deficiencies limit it to a small power of 2 (<= 16). We use 16 in the kernel to get smaller tables (see Makefile.i386 and <machine/asmacros.h>). We still use the default of 4 in user mode. Use HISTCOUNTER instead of (*kcount) in the definition of KCOUNT() for consistency with other macros.
1997-02-13 10:47:29 +00:00
/*
* Config generates something to tell the compiler to align functions on 16
* byte boundaries. A strict alignment is good for keeping the tables small.
*/
#define FUNCTION_ALIGNMENT 16
Fixed user-mode mcount which I broke in the previous revision. Do it the old way for now. Moved recent additions around a lot to minimise ifdefs. Added prototypes.
1996-01-01 17:11:21 +00:00
/*
* The kernel uses assembler stubs instead of unportable inlines.
* This is mainly to save a little time when profiling is not enabled,
* which is the usual case for the kernel.
*/
#define _MCOUNT_DECL void mcount
#define MCOUNT
Cleaned up interrupt masking by declaring the state variable in a machine-dependent macro and passing it to all machine-dependent macros. Eliminated the state variable for the GUPROF case.
1996-08-28 20:15:32 +00:00
#ifdef GUPROF
#define CALIB_SCALE 1000
#define KCOUNT(p,index) ((p)->kcount[(index) \
Moved definition of FUNCTION_ALIGNMENT to a machine-dependent place. Changed it from 4 to 16 for i386's. It can be anything for i386's, but compiler options limit it to a power of 2, and assembler and linker deficiencies limit it to a small power of 2 (<= 16). We use 16 in the kernel to get smaller tables (see Makefile.i386 and <machine/asmacros.h>). We still use the default of 4 in user mode. Use HISTCOUNTER instead of (*kcount) in the definition of KCOUNT() for consistency with other macros.
1997-02-13 10:47:29 +00:00
/ (HISTFRACTION * sizeof(HISTCOUNTER))])
Cleaned up interrupt masking by declaring the state variable in a machine-dependent macro and passing it to all machine-dependent macros. Eliminated the state variable for the GUPROF case.
1996-08-28 20:15:32 +00:00
#define MCOUNT_DECL(s)
#define MCOUNT_ENTER(s)
#define MCOUNT_EXIT(s)
Changed to the C9x draft spelling of the (unsigned) integral type suitable for holding object pointers (ptrint_t -> uintptr_t). Added corresponding signed type (intptr_t). Changed/added corresponding non-C9x types for function pointers to match. Don't use nonstandard types to implement these types, and don't comment on them in <machine/types.h>.
1998-07-14 05:09:48 +00:00
#define PC_TO_I(p, pc) ((uintfptr_t)(pc) - (uintfptr_t)(p)->lowpc)
Cleaned up interrupt masking by declaring the state variable in a machine-dependent macro and passing it to all machine-dependent macros. Eliminated the state variable for the GUPROF case.
1996-08-28 20:15:32 +00:00
#else
#define MCOUNT_DECL(s) u_long s;
Another round of lock pushdown. Add a simplelock to deal with disable_intr()/enable_intr() as used in UP kernel. UP kernel expects that this is enough to guarantee exclusive access to regions of code bracketed by these 2 functions. Add a simplelock to bracket clock accesses in clock.c: clock_lock. Help from: Bruce Evans <bde@zeta.org.au>
1997-08-30 08:08:10 +00:00
#ifdef SMP
Finish revs.1.23 and 1.24 so that MCOUNT_ENTER really actually compiles for SMP in the plain profiling case. It seems to work too. This error was not detected by LINT because LINT only compiles the GUPROF profiling case, which is is a superset of the plain profiling case for !SMP but which is so broken for SMP that the buggy code is not compiled.
2002-01-31 13:49:55 +00:00
extern int mcount_lock;
Commit MD parts of a loosely functional AMD64 port. This is based on a heavily stripped down FreeBSD/i386 (brutally stripped down actually) to attempt to get a stable base to start from. There is a lot missing still. Worth noting: - The kernel runs at 1GB in order to cheat with the pmap code. pmap uses a variation of the PAE code in order to avoid having to worry about 4 levels of page tables yet. - It boots in 64 bit "long mode" with a tiny trampoline embedded in the i386 loader. This simplifies locore.s greatly. - There are still quite a few fragments of i386-specific code that have not been translated yet, and some that I cheated and wrote dumb C versions of (bcopy etc). - It has both int 0x80 for syscalls (but using registers for argument passing, as is native on the amd64 ABI), and the 'syscall' instruction for syscalls. int 0x80 preserves all registers, 'syscall' does not. - I have tried to minimize looking at the NetBSD code, except in a couple of places (eg: to find which register they use to replace the trashed %rcx register in the syscall instruction). As a result, there is not a lot of similarity. I did look at NetBSD a few times while debugging to get some ideas about what I might have done wrong in my first attempt.
2003-05-01 01:05:25 +00:00
#define MCOUNT_ENTER(s) { s = read_rflags(); disable_intr(); \
Fix MCOUNT_ENTER() so it actually compiles in the profiling case. Pointy hat to: me Submitted by: Danny J. Zerkel <dzerkel@columbus.rr.com>
2001-07-14 21:40:53 +00:00
while (!atomic_cmpset_acq_int(&mcount_lock, 0, 1)) \
Get kernel profiling on SMP systems closer to working by replacing the mcount spin mutex with a very simple non-recursive spinlock implemented using atomic operations.
2001-06-28 04:03:29 +00:00
/* nothing */ ; }
#define MCOUNT_EXIT(s) { atomic_store_rel_int(&mcount_lock, 0); \
Commit MD parts of a loosely functional AMD64 port. This is based on a heavily stripped down FreeBSD/i386 (brutally stripped down actually) to attempt to get a stable base to start from. There is a lot missing still. Worth noting: - The kernel runs at 1GB in order to cheat with the pmap code. pmap uses a variation of the PAE code in order to avoid having to worry about 4 levels of page tables yet. - It boots in 64 bit "long mode" with a tiny trampoline embedded in the i386 loader. This simplifies locore.s greatly. - There are still quite a few fragments of i386-specific code that have not been translated yet, and some that I cheated and wrote dumb C versions of (bcopy etc). - It has both int 0x80 for syscalls (but using registers for argument passing, as is native on the amd64 ABI), and the 'syscall' instruction for syscalls. int 0x80 preserves all registers, 'syscall' does not. - I have tried to minimize looking at the NetBSD code, except in a couple of places (eg: to find which register they use to replace the trashed %rcx register in the syscall instruction). As a result, there is not a lot of similarity. I did look at NetBSD a few times while debugging to get some ideas about what I might have done wrong in my first attempt.
2003-05-01 01:05:25 +00:00
write_rflags(s); }
Another round of lock pushdown. Add a simplelock to deal with disable_intr()/enable_intr() as used in UP kernel. UP kernel expects that this is enough to guarantee exclusive access to regions of code bracketed by these 2 functions. Add a simplelock to bracket clock accesses in clock.c: clock_lock. Help from: Bruce Evans <bde@zeta.org.au>
1997-08-30 08:08:10 +00:00
#else
Commit MD parts of a loosely functional AMD64 port. This is based on a heavily stripped down FreeBSD/i386 (brutally stripped down actually) to attempt to get a stable base to start from. There is a lot missing still. Worth noting: - The kernel runs at 1GB in order to cheat with the pmap code. pmap uses a variation of the PAE code in order to avoid having to worry about 4 levels of page tables yet. - It boots in 64 bit "long mode" with a tiny trampoline embedded in the i386 loader. This simplifies locore.s greatly. - There are still quite a few fragments of i386-specific code that have not been translated yet, and some that I cheated and wrote dumb C versions of (bcopy etc). - It has both int 0x80 for syscalls (but using registers for argument passing, as is native on the amd64 ABI), and the 'syscall' instruction for syscalls. int 0x80 preserves all registers, 'syscall' does not. - I have tried to minimize looking at the NetBSD code, except in a couple of places (eg: to find which register they use to replace the trashed %rcx register in the syscall instruction). As a result, there is not a lot of similarity. I did look at NetBSD a few times while debugging to get some ideas about what I might have done wrong in my first attempt.
2003-05-01 01:05:25 +00:00
#define MCOUNT_ENTER(s) { s = read_rflags(); disable_intr(); }
#define MCOUNT_EXIT(s) (write_rflags(s))
Another round of lock pushdown. Add a simplelock to deal with disable_intr()/enable_intr() as used in UP kernel. UP kernel expects that this is enough to guarantee exclusive access to regions of code bracketed by these 2 functions. Add a simplelock to bracket clock accesses in clock.c: clock_lock. Help from: Bruce Evans <bde@zeta.org.au>
1997-08-30 08:08:10 +00:00
#endif
Cleaned up interrupt masking by declaring the state variable in a machine-dependent macro and passing it to all machine-dependent macros. Eliminated the state variable for the GUPROF case.
1996-08-28 20:15:32 +00:00
#endif /* GUPROF */
Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL" is an application space macro and the applications are supposed to be free to use it as they please (but cannot). This is consistant with the other BSD's who made this change quite some time ago. More commits to come.
1999-12-29 04:46:21 +00:00
#else /* !_KERNEL */
Fixed user-mode mcount which I broke in the previous revision. Do it the old way for now. Moved recent additions around a lot to minimise ifdefs. Added prototypes.
1996-01-01 17:11:21 +00:00
Moved definition of FUNCTION_ALIGNMENT to a machine-dependent place. Changed it from 4 to 16 for i386's. It can be anything for i386's, but compiler options limit it to a power of 2, and assembler and linker deficiencies limit it to a small power of 2 (<= 16). We use 16 in the kernel to get smaller tables (see Makefile.i386 and <machine/asmacros.h>). We still use the default of 4 in user mode. Use HISTCOUNTER instead of (*kcount) in the definition of KCOUNT() for consistency with other macros.
1997-02-13 10:47:29 +00:00
#define FUNCTION_ALIGNMENT 4
Fixed user-mode mcount which I broke in the previous revision. Do it the old way for now. Moved recent additions around a lot to minimise ifdefs. Added prototypes.
1996-01-01 17:11:21 +00:00
#define _MCOUNT_DECL static __inline void _mcount
BSD 4.4 Lite Kernel Sources
1994-05-25 01:34:38 +00:00
Stylify (mainly line up macro EOL-continuation \'s), and add a dummy alternative for lint.
2002-04-21 10:49:00 +00:00
#ifdef __GNUC__
#define MCOUNT \
void \
mcount() \
{ \
uintfptr_t selfpc, frompc; \
/* \
* Find the return address for mcount, \
* and the return address for mcount's caller. \
* \
* selfpc = pc pushed by call to mcount \
*/ \
Commit MD parts of a loosely functional AMD64 port. This is based on a heavily stripped down FreeBSD/i386 (brutally stripped down actually) to attempt to get a stable base to start from. There is a lot missing still. Worth noting: - The kernel runs at 1GB in order to cheat with the pmap code. pmap uses a variation of the PAE code in order to avoid having to worry about 4 levels of page tables yet. - It boots in 64 bit "long mode" with a tiny trampoline embedded in the i386 loader. This simplifies locore.s greatly. - There are still quite a few fragments of i386-specific code that have not been translated yet, and some that I cheated and wrote dumb C versions of (bcopy etc). - It has both int 0x80 for syscalls (but using registers for argument passing, as is native on the amd64 ABI), and the 'syscall' instruction for syscalls. int 0x80 preserves all registers, 'syscall' does not. - I have tried to minimize looking at the NetBSD code, except in a couple of places (eg: to find which register they use to replace the trashed %rcx register in the syscall instruction). As a result, there is not a lot of similarity. I did look at NetBSD a few times while debugging to get some ideas about what I might have done wrong in my first attempt.
2003-05-01 01:05:25 +00:00
asm("movq 8(%%rbp),%0" : "=r" (selfpc)); \
Stylify (mainly line up macro EOL-continuation \'s), and add a dummy alternative for lint.
2002-04-21 10:49:00 +00:00
/* \
* frompc = pc pushed by call to mcount's caller. \
* The caller's stack frame has already been built, so %ebp is \
* the caller's frame pointer. The caller's raddr is in the \
* caller's frame following the caller's caller's frame pointer.\
*/ \
Commit MD parts of a loosely functional AMD64 port. This is based on a heavily stripped down FreeBSD/i386 (brutally stripped down actually) to attempt to get a stable base to start from. There is a lot missing still. Worth noting: - The kernel runs at 1GB in order to cheat with the pmap code. pmap uses a variation of the PAE code in order to avoid having to worry about 4 levels of page tables yet. - It boots in 64 bit "long mode" with a tiny trampoline embedded in the i386 loader. This simplifies locore.s greatly. - There are still quite a few fragments of i386-specific code that have not been translated yet, and some that I cheated and wrote dumb C versions of (bcopy etc). - It has both int 0x80 for syscalls (but using registers for argument passing, as is native on the amd64 ABI), and the 'syscall' instruction for syscalls. int 0x80 preserves all registers, 'syscall' does not. - I have tried to minimize looking at the NetBSD code, except in a couple of places (eg: to find which register they use to replace the trashed %rcx register in the syscall instruction). As a result, there is not a lot of similarity. I did look at NetBSD a few times while debugging to get some ideas about what I might have done wrong in my first attempt.
2003-05-01 01:05:25 +00:00
asm("movq (%%rbp),%0" : "=r" (frompc)); \
Stylify (mainly line up macro EOL-continuation \'s), and add a dummy alternative for lint.
2002-04-21 10:49:00 +00:00
frompc = ((uintfptr_t *)frompc)[1]; \
_mcount(frompc, selfpc); \
BSD 4.4 Lite Kernel Sources
1994-05-25 01:34:38 +00:00
}
Stylify (mainly line up macro EOL-continuation \'s), and add a dummy alternative for lint.
2002-04-21 10:49:00 +00:00
#else /* __GNUC__ */
Fix a declaration that is actually supposed to be a macro definition. Submitted by: marius@alchemy.franken.de
2002-09-25 13:46:23 +00:00
#define MCOUNT \
Stylify (mainly line up macro EOL-continuation \'s), and add a dummy alternative for lint.
2002-04-21 10:49:00 +00:00
void \
mcount() \
{ \
}
#endif /* __GNUC__ */
Oops, fptrint_t still needs to be declared in <machine/profile.h> in the !KERNEL case. The kludge to get it declared in libc/gmon/mcount.c wasn't sufficient because fptrint_t is used in <sys/gmon.h>.
1998-07-10 09:26:41 +00:00
Commit MD parts of a loosely functional AMD64 port. This is based on a heavily stripped down FreeBSD/i386 (brutally stripped down actually) to attempt to get a stable base to start from. There is a lot missing still. Worth noting: - The kernel runs at 1GB in order to cheat with the pmap code. pmap uses a variation of the PAE code in order to avoid having to worry about 4 levels of page tables yet. - It boots in 64 bit "long mode" with a tiny trampoline embedded in the i386 loader. This simplifies locore.s greatly. - There are still quite a few fragments of i386-specific code that have not been translated yet, and some that I cheated and wrote dumb C versions of (bcopy etc). - It has both int 0x80 for syscalls (but using registers for argument passing, as is native on the amd64 ABI), and the 'syscall' instruction for syscalls. int 0x80 preserves all registers, 'syscall' does not. - I have tried to minimize looking at the NetBSD code, except in a couple of places (eg: to find which register they use to replace the trashed %rcx register in the syscall instruction). As a result, there is not a lot of similarity. I did look at NetBSD a few times while debugging to get some ideas about what I might have done wrong in my first attempt.
2003-05-01 01:05:25 +00:00
typedef unsigned long uintfptr_t;
Oops, fptrint_t still needs to be declared in <machine/profile.h> in the !KERNEL case. The kludge to get it declared in libc/gmon/mcount.c wasn't sufficient because fptrint_t is used in <sys/gmon.h>.
1998-07-10 09:26:41 +00:00
Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL" is an application space macro and the applications are supposed to be free to use it as they please (but cannot). This is consistant with the other BSD's who made this change quite some time ago. More commits to come.
1999-12-29 04:46:21 +00:00
#endif /* _KERNEL */
Added MCOUNT_ENTER and MCOUNT_EXIT macros to profile.h Removed inb function since it's more correctly in pio.h Copied write_eflags and read_eflags over from npx.c (Some changes to the macros suggested by Bruce were not made at this time since his suggestions probably apply to all the macros and these inlined/macro definitions need a lot of cleaning up at some point in the future.) Reviewed by: Bruce
1994-09-15 16:27:14 +00:00
Implemented non-statistical kernel profiling. This is based on looking at a high resolution clock for each of the following events: function call, function return, interrupt entry, interrupt exit, and interesting branches. The differences between the times of these events are added at appropriate places in a ordinary histogram (as if very fast statistical profiling sampled the pc at those places) so that ordinary gprof can be used to analyze the times. gmon.h: Histogram counters need to be 4 bytes for microsecond resolutions. They will need to be larger for the 586 clock. The comments were vax-centric and wrong even on vaxes. Does anyone disagree? gprof4.c: The standard gprof should support counters of all integral sizes and the size of the counter should be in the gmon header. This hack will do until then. (Use gprof4 -u to examine the results of non-statistical profiling.) config/*: Non-statistical profiling is configured with `config -pp'. `config -p' still gives ordinary profiling. kgmon/*: Non-statistical profiling is enabled with `kgmon -B'. `kgmon -b' still enables ordinary profiling (and distables non-statistical profiling) if non-statistical profiling is configured.
1995-12-29 15:30:05 +00:00
/*
* An unsigned integral type that can hold non-negative difference between
* function pointers.
*/
Add support for low resolution SMP kernel profiling. - A nonprofiling version of s_lock (called s_lock_np) is used by mcount. - When profiling is active, more registers are clobbered in seemingly simple assembly routines. This means that some callers needed to save/restore extra registers. - The stack pointer must have space for a 'fake' return address in idle, to avoid stack underflow.
1997-12-15 02:18:35 +00:00
typedef u_int fptrdiff_t;
Implemented non-statistical kernel profiling. This is based on looking at a high resolution clock for each of the following events: function call, function return, interrupt entry, interrupt exit, and interesting branches. The differences between the times of these events are added at appropriate places in a ordinary histogram (as if very fast statistical profiling sampled the pc at those places) so that ordinary gprof can be used to analyze the times. gmon.h: Histogram counters need to be 4 bytes for microsecond resolutions. They will need to be larger for the 586 clock. The comments were vax-centric and wrong even on vaxes. Does anyone disagree? gprof4.c: The standard gprof should support counters of all integral sizes and the size of the counter should be in the gmon header. This hack will do until then. (Use gprof4 -u to examine the results of non-statistical profiling.) config/*: Non-statistical profiling is configured with `config -pp'. `config -p' still gives ordinary profiling. kgmon/*: Non-statistical profiling is enabled with `kgmon -B'. `kgmon -b' still enables ordinary profiling (and distables non-statistical profiling) if non-statistical profiling is configured.
1995-12-29 15:30:05 +00:00
Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL" is an application space macro and the applications are supposed to be free to use it as they please (but cannot). This is consistant with the other BSD's who made this change quite some time ago. More commits to come.
1999-12-29 04:46:21 +00:00
#ifdef _KERNEL
Improved non-statistical (GUPROF) profiling: - use a more accurate and more efficient method of compensating for overheads. The old method counted too much time against leaf functions. - normally use the Pentium timestamp counter if available. On Pentiums, the times are now accurate to within a couple of cpu clock cycles per function call in the (unlikely) event that there are no cache misses in or caused by the profiling code. - optionally use an arbitrary Pentium event counter if available. - optionally regress to using the i8254 counter. - scaled the i8254 counter by a factor of 128. Now the i8254 counters overflow slightly faster than the TSC counters for a 150MHz Pentium :-) (after about 16 seconds). This is to avoid fractional overheads. files.i386: permon.c temporarily has to be classified as a profiling-routine because a couple of functions in it may be called from profiling code. options.i386: - I586_CTR_GUPROF is currently unused (oops). - I586_PMC_GUPROF should be something like 0x70000 to enable (but not use unless prof_machdep.c is changed) support for Pentium event counters. 7 is a control mode and the counter number 0 is somewhere in the 0000 bits (see perfmon.h for the encoding). profile.h: - added declarations. - cleaned up separation of user mode declarations. prof_machdep.c: Mostly clock-select changes. The default clock can be changed by editing kmem. There should be a sysctl for this. subr_prof.c: - added copyright. - calibrate overheads for the new method. - documented new method. - fixed races and and machine dependencies in start/stop code. mcount.c: Use the new overhead compensation method. gmon.h: - changed GPROF4 counter type from unsigned to int. Oops, this should be machine-dependent and/or int32_t. - reorganized overhead counters. Submitted by: Pentium event counter changes mostly by wollman
1996-10-17 19:32:31 +00:00
Remove __P.
2002-03-20 05:48:58 +00:00
void mcount(uintfptr_t frompc, uintfptr_t selfpc);
void kmupetext(uintfptr_t nhighpc);
Improved non-statistical (GUPROF) profiling: - use a more accurate and more efficient method of compensating for overheads. The old method counted too much time against leaf functions. - normally use the Pentium timestamp counter if available. On Pentiums, the times are now accurate to within a couple of cpu clock cycles per function call in the (unlikely) event that there are no cache misses in or caused by the profiling code. - optionally use an arbitrary Pentium event counter if available. - optionally regress to using the i8254 counter. - scaled the i8254 counter by a factor of 128. Now the i8254 counters overflow slightly faster than the TSC counters for a 150MHz Pentium :-) (after about 16 seconds). This is to avoid fractional overheads. files.i386: permon.c temporarily has to be classified as a profiling-routine because a couple of functions in it may be called from profiling code. options.i386: - I586_CTR_GUPROF is currently unused (oops). - I586_PMC_GUPROF should be something like 0x70000 to enable (but not use unless prof_machdep.c is changed) support for Pentium event counters. 7 is a control mode and the counter number 0 is somewhere in the 0000 bits (see perfmon.h for the encoding). profile.h: - added declarations. - cleaned up separation of user mode declarations. prof_machdep.c: Mostly clock-select changes. The default clock can be changed by editing kmem. There should be a sysctl for this. subr_prof.c: - added copyright. - calibrate overheads for the new method. - documented new method. - fixed races and and machine dependencies in start/stop code. mcount.c: Use the new overhead compensation method. gmon.h: - changed GPROF4 counter type from unsigned to int. Oops, this should be machine-dependent and/or int32_t. - reorganized overhead counters. Submitted by: Pentium event counter changes mostly by wollman
1996-10-17 19:32:31 +00:00
#ifdef GUPROF
struct gmonparam;
Remove __P.
2002-03-20 05:48:58 +00:00
void nullfunc_loop_profiled(void);
void nullfunc_profiled(void);
void startguprof(struct gmonparam *p);
void stopguprof(struct gmonparam *p);
Fixed user-mode mcount which I broke in the previous revision. Do it the old way for now. Moved recent additions around a lot to minimise ifdefs. Added prototypes.
1996-01-01 17:11:21 +00:00
#else
Improved non-statistical (GUPROF) profiling: - use a more accurate and more efficient method of compensating for overheads. The old method counted too much time against leaf functions. - normally use the Pentium timestamp counter if available. On Pentiums, the times are now accurate to within a couple of cpu clock cycles per function call in the (unlikely) event that there are no cache misses in or caused by the profiling code. - optionally use an arbitrary Pentium event counter if available. - optionally regress to using the i8254 counter. - scaled the i8254 counter by a factor of 128. Now the i8254 counters overflow slightly faster than the TSC counters for a 150MHz Pentium :-) (after about 16 seconds). This is to avoid fractional overheads. files.i386: permon.c temporarily has to be classified as a profiling-routine because a couple of functions in it may be called from profiling code. options.i386: - I586_CTR_GUPROF is currently unused (oops). - I586_PMC_GUPROF should be something like 0x70000 to enable (but not use unless prof_machdep.c is changed) support for Pentium event counters. 7 is a control mode and the counter number 0 is somewhere in the 0000 bits (see perfmon.h for the encoding). profile.h: - added declarations. - cleaned up separation of user mode declarations. prof_machdep.c: Mostly clock-select changes. The default clock can be changed by editing kmem. There should be a sysctl for this. subr_prof.c: - added copyright. - calibrate overheads for the new method. - documented new method. - fixed races and and machine dependencies in start/stop code. mcount.c: Use the new overhead compensation method. gmon.h: - changed GPROF4 counter type from unsigned to int. Oops, this should be machine-dependent and/or int32_t. - reorganized overhead counters. Submitted by: Pentium event counter changes mostly by wollman
1996-10-17 19:32:31 +00:00
#define startguprof(p)
#define stopguprof(p)
#endif /* GUPROF */
Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL" is an application space macro and the applications are supposed to be free to use it as they please (but cannot). This is consistant with the other BSD's who made this change quite some time ago. More commits to come.
1999-12-29 04:46:21 +00:00
#else /* !_KERNEL */
Improved non-statistical (GUPROF) profiling: - use a more accurate and more efficient method of compensating for overheads. The old method counted too much time against leaf functions. - normally use the Pentium timestamp counter if available. On Pentiums, the times are now accurate to within a couple of cpu clock cycles per function call in the (unlikely) event that there are no cache misses in or caused by the profiling code. - optionally use an arbitrary Pentium event counter if available. - optionally regress to using the i8254 counter. - scaled the i8254 counter by a factor of 128. Now the i8254 counters overflow slightly faster than the TSC counters for a 150MHz Pentium :-) (after about 16 seconds). This is to avoid fractional overheads. files.i386: permon.c temporarily has to be classified as a profiling-routine because a couple of functions in it may be called from profiling code. options.i386: - I586_CTR_GUPROF is currently unused (oops). - I586_PMC_GUPROF should be something like 0x70000 to enable (but not use unless prof_machdep.c is changed) support for Pentium event counters. 7 is a control mode and the counter number 0 is somewhere in the 0000 bits (see perfmon.h for the encoding). profile.h: - added declarations. - cleaned up separation of user mode declarations. prof_machdep.c: Mostly clock-select changes. The default clock can be changed by editing kmem. There should be a sysctl for this. subr_prof.c: - added copyright. - calibrate overheads for the new method. - documented new method. - fixed races and and machine dependencies in start/stop code. mcount.c: Use the new overhead compensation method. gmon.h: - changed GPROF4 counter type from unsigned to int. Oops, this should be machine-dependent and/or int32_t. - reorganized overhead counters. Submitted by: Pentium event counter changes mostly by wollman
1996-10-17 19:32:31 +00:00
#include <sys/cdefs.h>
__BEGIN_DECLS
Ifdefed use of a GNU feature.
1998-02-03 20:32:38 +00:00
#ifdef __GNUC__
Remove __P.
2002-03-20 05:48:58 +00:00
void mcount(void) __asm(".mcount");
Make profiling work for ELF. gprof now autodetects the format of the executable file, so it will work for both a.out and ELF format files. I have split the object format specific code into separate source files. It's cleaner than it was before, but it's still pretty crufty. Don't cheat on your make world for this update. A lot of things have to be rebuilt for it to work, including the compiler and all of the profiled libraries.
1998-09-07 23:32:00 +00:00
#endif
Remove __P.
2002-03-20 05:48:58 +00:00
static void _mcount(uintfptr_t frompc, uintfptr_t selfpc);
Improved non-statistical (GUPROF) profiling: - use a more accurate and more efficient method of compensating for overheads. The old method counted too much time against leaf functions. - normally use the Pentium timestamp counter if available. On Pentiums, the times are now accurate to within a couple of cpu clock cycles per function call in the (unlikely) event that there are no cache misses in or caused by the profiling code. - optionally use an arbitrary Pentium event counter if available. - optionally regress to using the i8254 counter. - scaled the i8254 counter by a factor of 128. Now the i8254 counters overflow slightly faster than the TSC counters for a 150MHz Pentium :-) (after about 16 seconds). This is to avoid fractional overheads. files.i386: permon.c temporarily has to be classified as a profiling-routine because a couple of functions in it may be called from profiling code. options.i386: - I586_CTR_GUPROF is currently unused (oops). - I586_PMC_GUPROF should be something like 0x70000 to enable (but not use unless prof_machdep.c is changed) support for Pentium event counters. 7 is a control mode and the counter number 0 is somewhere in the 0000 bits (see perfmon.h for the encoding). profile.h: - added declarations. - cleaned up separation of user mode declarations. prof_machdep.c: Mostly clock-select changes. The default clock can be changed by editing kmem. There should be a sysctl for this. subr_prof.c: - added copyright. - calibrate overheads for the new method. - documented new method. - fixed races and and machine dependencies in start/stop code. mcount.c: Use the new overhead compensation method. gmon.h: - changed GPROF4 counter type from unsigned to int. Oops, this should be machine-dependent and/or int32_t. - reorganized overhead counters. Submitted by: Pentium event counter changes mostly by wollman
1996-10-17 19:32:31 +00:00
__END_DECLS
Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL" is an application space macro and the applications are supposed to be free to use it as they please (but cannot). This is consistant with the other BSD's who made this change quite some time ago. More commits to come.
1999-12-29 04:46:21 +00:00
#endif /* _KERNEL */
Fixed user-mode mcount which I broke in the previous revision. Do it the old way for now. Moved recent additions around a lot to minimise ifdefs. Added prototypes.
1996-01-01 17:11:21 +00:00
#ifdef GUPROF
Improved non-statistical (GUPROF) profiling: - use a more accurate and more efficient method of compensating for overheads. The old method counted too much time against leaf functions. - normally use the Pentium timestamp counter if available. On Pentiums, the times are now accurate to within a couple of cpu clock cycles per function call in the (unlikely) event that there are no cache misses in or caused by the profiling code. - optionally use an arbitrary Pentium event counter if available. - optionally regress to using the i8254 counter. - scaled the i8254 counter by a factor of 128. Now the i8254 counters overflow slightly faster than the TSC counters for a 150MHz Pentium :-) (after about 16 seconds). This is to avoid fractional overheads. files.i386: permon.c temporarily has to be classified as a profiling-routine because a couple of functions in it may be called from profiling code. options.i386: - I586_CTR_GUPROF is currently unused (oops). - I586_PMC_GUPROF should be something like 0x70000 to enable (but not use unless prof_machdep.c is changed) support for Pentium event counters. 7 is a control mode and the counter number 0 is somewhere in the 0000 bits (see perfmon.h for the encoding). profile.h: - added declarations. - cleaned up separation of user mode declarations. prof_machdep.c: Mostly clock-select changes. The default clock can be changed by editing kmem. There should be a sysctl for this. subr_prof.c: - added copyright. - calibrate overheads for the new method. - documented new method. - fixed races and and machine dependencies in start/stop code. mcount.c: Use the new overhead compensation method. gmon.h: - changed GPROF4 counter type from unsigned to int. Oops, this should be machine-dependent and/or int32_t. - reorganized overhead counters. Submitted by: Pentium event counter changes mostly by wollman
1996-10-17 19:32:31 +00:00
/* XXX doesn't quite work outside kernel yet. */
extern int cputime_bias;
__BEGIN_DECLS
Remove __P.
2002-03-20 05:48:58 +00:00
int cputime(void);
void empty_loop(void);
void mexitcount(uintfptr_t selfpc);
void nullfunc(void);
void nullfunc_loop(void);
Fixed user-mode mcount which I broke in the previous revision. Do it the old way for now. Moved recent additions around a lot to minimise ifdefs. Added prototypes.
1996-01-01 17:11:21 +00:00
__END_DECLS
Improved non-statistical (GUPROF) profiling: - use a more accurate and more efficient method of compensating for overheads. The old method counted too much time against leaf functions. - normally use the Pentium timestamp counter if available. On Pentiums, the times are now accurate to within a couple of cpu clock cycles per function call in the (unlikely) event that there are no cache misses in or caused by the profiling code. - optionally use an arbitrary Pentium event counter if available. - optionally regress to using the i8254 counter. - scaled the i8254 counter by a factor of 128. Now the i8254 counters overflow slightly faster than the TSC counters for a 150MHz Pentium :-) (after about 16 seconds). This is to avoid fractional overheads. files.i386: permon.c temporarily has to be classified as a profiling-routine because a couple of functions in it may be called from profiling code. options.i386: - I586_CTR_GUPROF is currently unused (oops). - I586_PMC_GUPROF should be something like 0x70000 to enable (but not use unless prof_machdep.c is changed) support for Pentium event counters. 7 is a control mode and the counter number 0 is somewhere in the 0000 bits (see perfmon.h for the encoding). profile.h: - added declarations. - cleaned up separation of user mode declarations. prof_machdep.c: Mostly clock-select changes. The default clock can be changed by editing kmem. There should be a sysctl for this. subr_prof.c: - added copyright. - calibrate overheads for the new method. - documented new method. - fixed races and and machine dependencies in start/stop code. mcount.c: Use the new overhead compensation method. gmon.h: - changed GPROF4 counter type from unsigned to int. Oops, this should be machine-dependent and/or int32_t. - reorganized overhead counters. Submitted by: Pentium event counter changes mostly by wollman
1996-10-17 19:32:31 +00:00
#endif
Implemented non-statistical kernel profiling. This is based on looking at a high resolution clock for each of the following events: function call, function return, interrupt entry, interrupt exit, and interesting branches. The differences between the times of these events are added at appropriate places in a ordinary histogram (as if very fast statistical profiling sampled the pc at those places) so that ordinary gprof can be used to analyze the times. gmon.h: Histogram counters need to be 4 bytes for microsecond resolutions. They will need to be larger for the 586 clock. The comments were vax-centric and wrong even on vaxes. Does anyone disagree? gprof4.c: The standard gprof should support counters of all integral sizes and the size of the counter should be in the gmon header. This hack will do until then. (Use gprof4 -u to examine the results of non-statistical profiling.) config/*: Non-statistical profiling is configured with `config -pp'. `config -p' still gives ordinary profiling. kgmon/*: Non-statistical profiling is enabled with `kgmon -B'. `kgmon -b' still enables ordinary profiling (and distables non-statistical profiling) if non-statistical profiling is configured.
1995-12-29 15:30:05 +00:00
Fixed user-mode mcount which I broke in the previous revision. Do it the old way for now. Moved recent additions around a lot to minimise ifdefs. Added prototypes.
1996-01-01 17:11:21 +00:00
#endif /* !_MACHINE_PROFILE_H_ */
Reference in New Issue Copy Permalink