| History log of /src/lib/libpthread/arch/i386/pthread_md.h |
| Revision | | Date | Author | Comments |
| 1.21 |
| 25-May-2023 |
riastradh | libpthread: New pthread__smt_wait to put CPU in low power for spin.
This is now distinct from pthread__smt_pause, which is for spin lock
backoff with no paired wakeup.
On Arm, there is a single-bit event register per CPU, and there are two
instructions to manage it:
- wfe, wait for event -- if event register is clear, enter low power
mode and wait until event register is set; then exit low power mode
and clear event register
- sev, signal event -- sets event register on all CPUs (other
circumstances like interrupts also set the event register and cause
wfe to wake)
These can be used to reduce the power consumption of spinning for a
lock, but only if they are actually paired -- if there's no sev, wfe
might hang indefinitely. Currently only pthread_spin(3) actually
pairs them; the other lock primitives (internal lock, mutex, rwlock)
do not -- they have spin lock backoff loops, but no corresponding
wakeup to cancel a wfe.
It may be worthwhile to teach the other lock primitives to pair
wfe/sev, but that requires some performance measurement to verify
it's actually worthwhile. So for now, we just make sure not to use
wfe when there's no sev, and keep everything else the same -- this
should fix severe performance degredation in libpthread on Arm
without hurting anything else.
No change in the generated code on amd64 and i386. No change in the
generated code for pthread_spin.c on arm and aarch64 -- changes only
the generated code for pthread_lock.c, pthread_mutex.c, and
pthread_rwlock.c, as intended.
PR port-arm/57437
XXX pullup-10
|
| 1.20 |
| 02-Mar-2012 |
joerg | branches: 1.20.24; 1.20.34; 1.20.42;
Avoid getcontext() as it triggers clobbering warnings. Use inline
assembler to get the fields directly. Saves a system call as side
effect.
|
| 1.19 |
| 24-Feb-2011 |
joerg | branches: 1.19.4;
Allow storing and receiving the LWP private pointer via ucontext_t
on all platforms except VAX and IA64. Add fast access via register for
AMD64, i386 and SH3 ports. Use this fast access in libpthread to replace
the stack based pthread_self(). Implement skeleton support for Alpha,
HPPA, PowerPC, SPARC and SPARC64, but leave it disabled.
Ports that support this feature provide __HAVE____LWP_GETPRIVATE_FAST in
machine/types.h and a corresponding __lwp_getprivate_fast in
machine/mcontext.h.
This material is based upon work partially supported by
The NetBSD Foundation under a contract with Joerg Sonnenberger.
|
| 1.18 |
| 25-Jan-2011 |
christos | make pthread__sp unsigned long.
|
| 1.17 |
| 16-May-2009 |
ad | branches: 1.17.2;
Remove unused code that's confusing when using cscope/opengrok.
|
| 1.16 |
| 29-Mar-2009 |
ad | - Make the threadreg code use _lwp_setprivate() instead of MD hooks.
XXX This must not be enabled by default because the LWP private mechanism
is reserved for TLS. It is provided only as a test/demo.
XXX Since ucontext_t does not contain the thread private variable, for a
short time after threads are created their thread specific data is unset.
If a signal arrives during that time we are screwed.
- No longer need pthread__osrev.
- Rearrange _lwp_ctl() calls slightly.
|
| 1.15 |
| 23-Jun-2008 |
ad | branches: 1.15.6; 1.15.8;
pthread__threadreg_get: mark it const.
|
| 1.14 |
| 28-Apr-2008 |
martin | branches: 1.14.2;
Remove clause 3 and 4 from TNF licenses
|
| 1.13 |
| 22-Mar-2008 |
ad | branches: 1.13.2;
Cheat and add inlines for _atomic_cas_ptr() to work around gcc emitting
unneeded PIC stuff in mutex_lock() and mutex_unlock(), when a thread
register is used.
|
| 1.12 |
| 10-Feb-2008 |
ad | - Remove libpthread's atomic ops.
- Remove the old spinlock-based mutex and rwlock implementations.
- Use the atomic ops from libc.
|
| 1.11 |
| 13-Nov-2007 |
ad | Mutexes:
- Play scrooge again and chop more cycles off acquire/release.
- Spin while the lock holder is running on another CPU (adaptive mutexes).
- Do non-atomic release.
Threadreg:
- Add the necessary hooks to use a thread register.
- Add the code for i386, using %gs.
- Leave i386 code disabled until xen and COMPAT_NETBSD32 have the changes.
|
| 1.10 |
| 24-Sep-2007 |
skrll | Resurrect the function pointers for lock operations and allow each
architecture to provide asm versions of the RAS operations.
We do this because relying on the compiler to get the RAS right is not
sensible. (It gets alpha wrong and hppa is suboptimal)
Provide asm RAS ops for hppa.
(A slightly different version) reviewed by Andrew Doran.
|
| 1.9 |
| 08-Sep-2007 |
ad | - Get rid of self->pt_mutexhint and use pthread__mutex_owned() instead.
- Update some comments and fix minor bugs. Minor cosmetic changes.
- Replace some spinlocks with mutexes and rwlocks.
- Change the process private semaphores to use mutexes and condition
variables instead of doing the synchronization directly. Spinlocks
are no longer used by the semaphore code.
|
| 1.8 |
| 07-Sep-2007 |
ad | Add: pthread__atomic_cas_ptr, pthread__atomic_swap_ptr, pthread__membar_full
This is a stopgap until the thorpej-atomic branch is complete.
|
| 1.7 |
| 29-Mar-2006 |
cube | branches: 1.7.8; 1.7.12;
Instead of using hard-coded values for various registers, get them from the
current context. Valid values can change depending on how the kernel is
setup. i386 and amd64 happen to be setup differently.
|
| 1.6 |
| 24-Dec-2005 |
perry | Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
|
| 1.5 |
| 11-Feb-2004 |
nathanw | branches: 1.5.4; 1.5.6;
Add ucontext conversion macros for an "extra" register set.
|
| 1.4 |
| 18-Jan-2003 |
christos | delint
|
| 1.3 |
| 18-Jan-2003 |
christos | de-lint
|
| 1.2 |
| 18-Jan-2003 |
thorpej | Merge the nathanw_sa branch.
|
| 1.1 |
| 05-Mar-2001 |
nathanw | branches: 1.1.2;
file pthread_md.h was initially added on branch nathanw_sa.
|
| 1.1.2.11 |
| 17-Jan-2003 |
thorpej | Include <sys/ucontext.h> to get ucontext_t for prototypes.
|
| 1.1.2.10 |
| 16-Jan-2003 |
thorpej | * Move the pthread_sigmask() prototype to <signal.h>.
* Don't include <signal.h> in <pthread.h>.
* Add code to the signal trampoline to convert from the ucontext
to a sigcontext, and back again (XXX though, only callee-save
regs for _UC_USER contexts).
This is necessary in order to support e.g. GCC's libjava, which depends
on the traditional Unix semantics of changes made to the sigcontext
being visible when the handler returns.
|
| 1.1.2.9 |
| 06-Dec-2002 |
nathanw | Clear _UC_USER bit when converting from sigcontext to ucontext.
|
| 1.1.2.8 |
| 22-Oct-2002 |
nathanw | Define _INITCONTEXT_U_MD() to set segment registers and flag register
to sensible values.
|
| 1.1.2.7 |
| 21-Oct-2002 |
nathanw | Add some macros to convert between ucontext_t and struct sigcontext.
|
| 1.1.2.6 |
| 14-Aug-2002 |
nathanw | Define, initialize, and use function pointers for _{set,get,swap}context_u
routines, with different versions for saving FP state with fnsave or fxsave,
depending on the machdep.osfxsr sysctl.
|
| 1.1.2.5 |
| 24-Apr-2002 |
nathanw | Add macros to convert between struct ucontext and struct reg/fpreg.
|
| 1.1.2.4 |
| 14-Nov-2001 |
briggs | Add pthread__uc_pc() to pthread_md.h instead of making assumptions about
the contents of the m.d. mcontext.
|
| 1.1.2.3 |
| 04-Sep-2001 |
nathanw | Define STACKSPACE in pthread_md.h, not directly in pthread_switch.S
|
| 1.1.2.2 |
| 13-Jul-2001 |
nathanw | Note copyright.
Standardize RCS IDs.
|
| 1.1.2.1 |
| 05-Mar-2001 |
nathanw | The beginnings of a scheduler activations-based pthread library.
|
| 1.5.6.1 |
| 19-Apr-2006 |
tron | Pull up following revision(s) (requested by cube in ticket #1265):
lib/libpthread/arch/i386/pthread_md.h: revision 1.7
Instead of using hard-coded values for various registers, get them from the
current context. Valid values can change depending on how the kernel is
setup. i386 and amd64 happen to be setup differently.
|
| 1.5.4.1 |
| 21-Apr-2006 |
tron | Pull up following revision(s) (requested by cube in ticket #10448):
lib/libpthread/arch/i386/pthread_md.h: revision 1.7
Instead of using hard-coded values for various registers, get them from the
current context. Valid values can change depending on how the kernel is
setup. i386 and amd64 happen to be setup differently.
|
| 1.7.12.3 |
| 23-Mar-2008 |
matt | sync with HEAD
|
| 1.7.12.2 |
| 09-Jan-2008 |
matt | sync with HEAD
|
| 1.7.12.1 |
| 06-Nov-2007 |
matt | sync with HEAD
|
| 1.7.8.1 |
| 10-Sep-2007 |
skrll | Sync with HEAD.
|
| 1.13.2.1 |
| 18-May-2008 |
yamt | sync with head.
|
| 1.14.2.1 |
| 18-Sep-2008 |
wrstuden | Sync with wrstuden-revivesa-base-2.
|
| 1.15.8.1 |
| 13-May-2009 |
jym | Sync with HEAD.
Third (and last) commit. See http://mail-index.netbsd.org/source-changes/2009/05/13/msg221222.html
|
| 1.15.6.2 |
| 23-Jun-2008 |
ad | pthread__threadreg_get: mark it const.
|
| 1.15.6.1 |
| 23-Jun-2008 |
ad | file pthread_md.h was added on branch christos-time_t on 2008-06-23 10:39:39 +0000
|
| 1.17.2.2 |
| 05-Mar-2011 |
bouyer | Sync with HEAD
|
| 1.17.2.1 |
| 08-Feb-2011 |
bouyer | Sync with HEAD
|
| 1.19.4.1 |
| 17-Apr-2012 |
yamt | sync with head
|
| 1.20.42.1 |
| 01-Aug-2023 |
martin | Pull up following revision(s) (requested by riastradh in ticket #296):
lib/libpthread/arch/x86_64/pthread_md.h: revision 1.13
lib/libpthread/pthread_int.h: revision 1.110
lib/libpthread/pthread_int.h: revision 1.111
lib/libpthread/arch/i386/pthread_md.h: revision 1.21
lib/libpthread/arch/arm/pthread_md.h: revision 1.12
lib/libpthread/arch/arm/pthread_md.h: revision 1.13
lib/libpthread/pthread_spin.c: revision 1.11
lib/libpthread/arch/aarch64/pthread_md.h: revision 1.2
libpthread: Use __nothing, not /* nothing */, for empty macros.
No functional change intended -- just safer to do it this way in case
the macros are used in if branches or comma expressions.
PR port-arm/57437 (pthread__smt_pause/wake issue)
libpthread: New pthread__smt_wait to put CPU in low power for spin.
This is now distinct from pthread__smt_pause, which is for spin lock
backoff with no paired wakeup.
On Arm, there is a single-bit event register per CPU, and there are two
instructions to manage it:
- wfe, wait for event -- if event register is clear, enter low power
mode and wait until event register is set; then exit low power mode
and clear event register
- sev, signal event -- sets event register on all CPUs (other
circumstances like interrupts also set the event register and cause
wfe to wake)
These can be used to reduce the power consumption of spinning for a
lock, but only if they are actually paired -- if there's no sev, wfe
might hang indefinitely. Currently only pthread_spin(3) actually
pairs them; the other lock primitives (internal lock, mutex, rwlock)
do not -- they have spin lock backoff loops, but no corresponding
wakeup to cancel a wfe.
It may be worthwhile to teach the other lock primitives to pair
wfe/sev, but that requires some performance measurement to verify
it's actually worthwhile. So for now, we just make sure not to use
wfe when there's no sev, and keep everything else the same -- this
should fix severe performance degredation in libpthread on Arm
without hurting anything else.
No change in the generated code on amd64 and i386. No change in the
generated code for pthread_spin.c on arm and aarch64 -- changes only
the generated code for pthread_lock.c, pthread_mutex.c, and
pthread_rwlock.c, as intended.
PR port-arm/57437
|
| 1.20.34.1 |
| 04-Aug-2023 |
martin | Pull up following revision(s) (requested by riastradh in ticket #1700):
lib/libpthread/arch/x86_64/pthread_md.h: revision 1.13
lib/libpthread/pthread_int.h: revision 1.110
lib/libpthread/pthread_int.h: revision 1.111
lib/libpthread/arch/i386/pthread_md.h: revision 1.21
lib/libpthread/arch/arm/pthread_md.h: revision 1.12
lib/libpthread/arch/arm/pthread_md.h: revision 1.13
lib/libpthread/pthread_spin.c: revision 1.11
lib/libpthread/arch/aarch64/pthread_md.h: revision 1.2
libpthread: Use __nothing, not /* nothing */, for empty macros.
No functional change intended -- just safer to do it this way in case
the macros are used in if branches or comma expressions.
PR port-arm/57437 (pthread__smt_pause/wake issue)
libpthread: New pthread__smt_wait to put CPU in low power for spin.
This is now distinct from pthread__smt_pause, which is for spin lock
backoff with no paired wakeup.
On Arm, there is a single-bit event register per CPU, and there are two
instructions to manage it:
- wfe, wait for event -- if event register is clear, enter low power
mode and wait until event register is set; then exit low power mode
and clear event register
- sev, signal event -- sets event register on all CPUs (other
circumstances like interrupts also set the event register and cause
wfe to wake)
These can be used to reduce the power consumption of spinning for a
lock, but only if they are actually paired -- if there's no sev, wfe
might hang indefinitely. Currently only pthread_spin(3) actually
pairs them; the other lock primitives (internal lock, mutex, rwlock)
do not -- they have spin lock backoff loops, but no corresponding
wakeup to cancel a wfe.
It may be worthwhile to teach the other lock primitives to pair
wfe/sev, but that requires some performance measurement to verify
it's actually worthwhile. So for now, we just make sure not to use
wfe when there's no sev, and keep everything else the same -- this
should fix severe performance degredation in libpthread on Arm
without hurting anything else.
No change in the generated code on amd64 and i386. No change in the
generated code for pthread_spin.c on arm and aarch64 -- changes only
the generated code for pthread_lock.c, pthread_mutex.c, and
pthread_rwlock.c, as intended.
PR port-arm/57437
|
| 1.20.24.1 |
| 04-Aug-2023 |
martin | Pull up following revision(s) (requested by riastradh in ticket #1878):
lib/libpthread/arch/x86_64/pthread_md.h: revision 1.13
lib/libpthread/pthread_int.h: revision 1.110
lib/libpthread/pthread_int.h: revision 1.111
lib/libpthread/arch/i386/pthread_md.h: revision 1.21
lib/libpthread/arch/arm/pthread_md.h: revision 1.12
lib/libpthread/arch/arm/pthread_md.h: revision 1.13
lib/libpthread/pthread_spin.c: revision 1.11
lib/libpthread/arch/aarch64/pthread_md.h: revision 1.2
libpthread: Use __nothing, not /* nothing */, for empty macros.
No functional change intended -- just safer to do it this way in case
the macros are used in if branches or comma expressions.
PR port-arm/57437 (pthread__smt_pause/wake issue)
libpthread: New pthread__smt_wait to put CPU in low power for spin.
This is now distinct from pthread__smt_pause, which is for spin lock
backoff with no paired wakeup.
On Arm, there is a single-bit event register per CPU, and there are two
instructions to manage it:
- wfe, wait for event -- if event register is clear, enter low power
mode and wait until event register is set; then exit low power mode
and clear event register
- sev, signal event -- sets event register on all CPUs (other
circumstances like interrupts also set the event register and cause
wfe to wake)
These can be used to reduce the power consumption of spinning for a
lock, but only if they are actually paired -- if there's no sev, wfe
might hang indefinitely. Currently only pthread_spin(3) actually
pairs them; the other lock primitives (internal lock, mutex, rwlock)
do not -- they have spin lock backoff loops, but no corresponding
wakeup to cancel a wfe.
It may be worthwhile to teach the other lock primitives to pair
wfe/sev, but that requires some performance measurement to verify
it's actually worthwhile. So for now, we just make sure not to use
wfe when there's no sev, and keep everything else the same -- this
should fix severe performance degredation in libpthread on Arm
without hurting anything else.
No change in the generated code on amd64 and i386. No change in the
generated code for pthread_spin.c on arm and aarch64 -- changes only
the generated code for pthread_lock.c, pthread_mutex.c, and
pthread_rwlock.c, as intended.
PR port-arm/57437
|
