qemu-e2k/TODO
bellard 56abbcfff3 update
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@3942 c046a42c-6fe2-441c-8c8c-71466251a162
2008-02-01 10:03:18 +00:00

61 lines
2.0 KiB
Plaintext

short term:
----------
- x86_64: fxsave/fxrestore bugs
- x86_64: lcall/ljmp intel/amd differences
- x86_64: cmpxchgl
- x86_64: cmovl bug
- x86: monitor invalid
- cycle counter for all archs
- cpu_interrupt() win32/SMP fix
- support variable tsc freq
- USB host async
- IDE async
- debug option in 'configure' script + disable -fomit-frame-pointer
- Precise VGA timings for old games/demos (malc patch)
- merge PIC spurious interrupt patch
- warning for OS/2: must not use 128 MB memory (merge bochs cmos patch ?)
- config file (at least for windows/Mac OS X)
- update doc: PCI infos.
- basic VGA optimizations
- better code fetch (different exception handling + CS.limit support)
- do not resize vga if invalid size.
- avoid looping if only exceptions
- TLB code protection support for PPC
- see openMosix Doc
- disable SMC handling for ARM/SPARC/PPC (not finished)
- see undefined flags for BTx insn
- user/kernel PUSHL/POPL in helper.c
- keyboard output buffer filling timing emulation
- return UD exception if LOCK prefix incorrectly used
- test ldt limit < 7 ?
- tests for each target CPU
- fix CCOP optimisation
- fix all remaining thread lock issues (must put TBs in a specific invalid
state, find a solution for tb_flush()).
ppc specific:
------------
- TLB invalidate not needed if msr_pr changes
- enable shift optimizations ?
linux-user specific:
-------------------
- add IPC syscalls
- handle rare page fault cases (in particular if page fault in helpers or
in syscall emulation code).
- more syscalls (in particular all 64 bit ones, IPCs, fix 64 bit
issues, fix 16 bit uid issues)
- use page_unprotect_range in every suitable syscall to handle all
cases of self modifying code.
- fix thread stack freeing (use kernel 2.5.x CLONE_CHILD_CLEARTID)
- use kernel traps for unaligned accesses on ARM ?
lower priority:
--------------
- int15 ah=86: use better timing
- suppress shift_mem ops
- fix some 16 bit sp push/pop overflow (pusha/popa, lcall lret)
- optimize FPU operations (evaluate x87 stack pointer statically)
- use -msoft-float on ARM