source: trunk/source/lisp-kernel

Revision Log Mode:


Legend:

Added
Modified
Copied or renamed
Diff Rev Age Author Log Message
(edit) @14347   9 years rme Define SVN_REVISION in Linux and Darwin makefiles (x86, x86-64).
(edit) @14346   9 years rme Look for SVN_REVISION macro, and by means of ridiculous C preprocessor …
(edit) @14330   9 years gb Mach exception entry/return maintain tcr->last_lisp_frame.
(edit) @14329   9 years gb Don't try to file-map a readonly area whose size is 0.
(edit) @14328   9 years gb In count_cpus(), use _SC_NPROCESSORS_CONF . (Power management may …
(edit) @14321   9 years gb Build using the Android NDK, not via an ARM Linux-hosted toolchain. Be …
(edit) @14320   9 years gb Limit MAXIMUM_MAPPABLE_MEMORY to 512M (even though some current …
(edit) @14319   9 years gb relocate_area_contents(): need to relocate ARM entrypoints.
(edit) @14295   9 years rme Eliminate some (but not all) warnings produced when building with …
(edit) @14292   9 years rme In describe_memfault(): Windows doesn't have siginfo_t. On Windows, …
(edit) @14291   9 years rme In describe_memfault(), use new function describe_siginfo() to show …
(edit) @14281   9 years gb Do sigaltstack() via syscall on Android, because the C library doesn't.
(edit) @14280   9 years gb More Android conditionalization.
(edit) @14279   9 years gb Someone got paid to leave xDisposeThread in the CCL kernel all these years.
(edit) @14278   9 years gb Someone got paid to develop Android's C library.
(edit) @14277   9 years gb Android doesn't have ... well, you get the idea.
(edit) @14276   9 years gb Android doesn't have dladdr(), presumably because it might be useful.
(edit) @14275   9 years gb Android doesn't have dladdr(), presumably because it might be useful.
(edit) @14274   9 years gb New file; likely needs work.
(edit) @14273   9 years gb In forward_memoized_references(), try to keep function's entrypoint in …
(edit) @14272   9 years gb In write-barrier code, return without store-conditional if the …
(edit) @14271   9 years gb Some Android conditionalization.
(edit) @14270   9 years gb Some Android conditionalization. (Use 'Rfn' for r11, not 'fn').
(edit) @14269   9 years gb Some Android conditionalization. Check for ARM function's …
(edit) @14268   9 years gb Work-in-progress: probably needs to be built with the NDK toolchain, …
(edit) @14267   9 years rme Sprinkle some parentheses around to avoid trouble with operator …
(edit) @14260   9 years rme In debug_show_fpu(), check to see if Linux's xp->uc_mcontext.fpregs is …
(edit) @14233   9 years gb Don't use the (Linux-specific ?) MAP_GROWSDOWN mmap option when …
(edit) @14232   9 years gb In initial_stack_bottom(), don't assume that stack bounds have …
(edit) @14207   9 years rme Don't do arithmetic on pointers to void. GCC accepts this (it …
(edit) @14206   9 years rme Rename lisp kernel functions allocate() and deallocate() to …
(edit) @14205   9 years rme Do #include <mach-o/dyld.h> on Darwin to get prototype for …
(edit) @14201   9 years gb Really stop disabling EGC (as I thought I had several weeks ago.)
(edit) @14199   9 years gb Free the (separate) signal stack on thread exit on ARM Linux.
(edit) @14197   9 years rme Rename Threads.h to threads.h (with no capital letter).
(edit) @14190   9 years gb is_write_fault() returns true if the 0x800 bit is set in the fsr.
(edit) @14189   9 years gb Define xpFaultStatus (to access the fault status register copy in a …
(edit) @14188   9 years gb Ensure that stack pointer is aligned on an 8-byte boundary when …
(edit) @14185   9 years gb in describe_memfault(), there's not much to say if we don't have a …
(edit) @14184   9 years gb Change the signature of the ARM flush_cache_lines() - from …
(edit) @14169   9 years gb Lots more changes, most of which have to do with Mach exception …
(edit) @14168   9 years gb Lots-o-changes, not all of them the final word.
(edit) @14167   9 years gb Don't really need to #include <fenv.h>; don't do so on Darwin/ARM, …
(edit) @14152   9 years rme Remove stray backquote.
(edit) @14149   9 years gb If write-protecting a memory region fails with ENOMEM, map the region …
(edit) @14146   9 years gb Common struct defs, #defines in constants.h.
(edit) @14145   9 years gb Don't #include ppc-constants.h directly.
(edit) @14144   9 years gb Reinstate foreign_fpe tcr flag bit.
(edit) @14143   9 years gb Remove redundant/shared constant definitions.
(edit) @14142   9 years gb Remove redundant/common struct and constant definitions.
(edit) @14141   9 years gb Common constants here; need to remove duplicates from most …
(edit) @14140   9 years gb ARM conditionalization botch.
(edit) @14139   9 years gb Right Makefile.
(edit) @14124   9 years gb extra arg to install_signal_handler; ignored on PPC.
(edit) @14123   9 years gb Fix xpMXCSR macro definition.
(edit) @14122   9 years gb De-pessimize the handling of the on_altstack arg to …
(edit) @14119   9 years gb Changes from ARM branch. Need testing …
(edit) @14012   9 years rme Use the versions of acosh[f], asinh[f], and atanh[f] that we use on …
(edit) @13994   9 years rme Clear tcr.ffi_exception in .SPffcall/.SPffcall_return_registers before …
(edit) @13976   9 years gb Spell "true" with an "e".
(edit) @13970   9 years gb Kernel side of WAIT-FOR-SIGNAL. Probably needs to be more: a better …
(edit) @13960   9 years rme Keep trying to get xpMXCSR right on Solaris.
(edit) @13956   9 years rme Use /usr/bin/gm4 on Solaris systems. I don't know if Sun fixed the …
(edit) @13954   9 years gb In commit_pages(), try to ensure that we return false if …
(edit) @13953   9 years gb Spell xpMXCSR with a 'p'. Split 64/32-bit trap stuff into the right …
(edit) @13952   9 years rme New scheme for dealing with floating point exceptions on x86-64 …
(edit) @13943   9 years rme When we get EXC_ARITHMETIC in catch_exception_raise(), translate a …
(edit) @13930   9 years rme describe_memfault(): FreeBSD 6.x doesn't define the si_code constants.
(edit) @13895   9 years gb Conditionalize for USE_GC_NOTIFICATATION (true on X86, not yet on PPC).
(edit) @13882   9 years gb When setting GC nofitication threshold, reset our notion of whether or …
(edit) @13881   9 years gb Do GC notification if dynamic heap shrinks in ensure_static_conses().
(edit) @13880   9 years gb Declare did_gc_notification_since_last_full_gc.
(edit) @13870   9 years gb x86-exceptions.c: Ensure that SIGTRAP is defined (it's not on Win64). …
(edit) @13869   9 years gb For Windows: as suggested by 'maetbag' on #ccl: suspend_tcr() and …
(edit) @13814   9 years rme Delete .cvsignore files since we no longer use CVS.
(edit) @13792   9 years gb New pre-GC notification mechanism.
(edit) @13770   9 years palter Not on Windows
(edit) @13767   9 years rme When we find that we've take an exception in some foreign context, …
(edit) @13766   9 years rme Make debug_identify_exception() do a little something on non-PPC …
(edit) @13765   9 years rme Kernel debugger command "M" to show lisp memory areas.
(edit) @13649   9 years gb Remove some identical macro definitions (introduced in a recent change.)
(edit) @13642   9 years gb Fix win32 typos.
(edit) @13641   9 years gb use win64 constants/macros in platform-win64
(edit) @13640   9 years gb fix fpu.h location
(edit) @13639   9 years gb fix PLATFORM_H
(edit) @13638   9 years gb Move more PPC exception stuff, x86 exception stuff to platform …
(edit) @13637   9 years gb Move some PPC exception stuff to platform header.
(edit) @13628   9 years gb Not critical (yet), but add constants.h to CHEADERS.
(edit) @13627   9 years gb Include lisptypes.h and word-size/arch-specific constants file in …
(edit) @13626   9 years gb New file. (Actually, an old one.) Will be included by arch-specific …
(edit) @13625   9 years rme Put $(PLATFORM_H) in CHEADERS.
(edit) @13624   9 years rme Define, -include PLATFORM_H.
(edit) @13617   9 years gb Move IMAGE_BASE_ADDRESS, PURESPACE_RESERVE definitions to platform …
(edit) @13612   9 years gb Define, -include PLATFORM_H.
(edit) @13611   9 years gb Need to include <windows.h>
(edit) @13610   9 years gb Need <windows.h>
(edit) @13609   9 years gb Define, -include PLATFORM_H.
(edit) @13608   9 years gb Define PLATFORM{OS|CPU|WORD_SIZE} in PLATFORM_H, not by exhaustive …
(edit) @13607   9 years gb Define, -include PLATFORM_H.
(edit) @13606   9 years gb Define, -include PLATFORM_H. Ensure that <sys/types.h> is #included. …
Note: See TracRevisionLog for help on using the revision log.