/[gxemul]/trunk/TODO
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Annotation of /trunk/TODO

Parent Directory Parent Directory | Revision Log Revision Log


Revision 26 - (hide annotations)
Mon Oct 8 16:20:10 2007 UTC (13 years ago) by dpavlin
File size: 13416 byte(s)
++ trunk/HISTORY	(local)
$Id: HISTORY,v 1.1264 2006/06/25 11:08:04 debug Exp $
20060624	Replacing the error-prone machine type initialization stuff
		with something more reasonable.
		Finally removing the old "cpu_run" kludge; moving around stuff
		in machine.c and emul.c to better suit the dyntrans system.
		Various minor dyntrans cleanups (renaming translate_address to
		translate_v2p, and experimenting with template physpages).
20060625	Removing the speed hack which separated the vph entries into
		two halves (code vs data); things seem a lot more stable now.
		Minor performance hack: R2000/R3000 cache isolation now only
		clears address translations when going into isolation, not
		when going out of it.
		Fixing the MIPS interrupt problems by letting mtc0 immediately
		cause interrupts.

==============  RELEASE 0.4.0.1  ==============


1 dpavlin 26 $Id: TODO,v 1.298 2006/06/25 11:08:04 debug Exp $
2 dpavlin 2
3 dpavlin 22 Hm. This file is in random order, and not all parts of it are up-to-date.
4 dpavlin 20
5 dpavlin 24 --------------
6 dpavlin 2
7 dpavlin 24 x) FIX THE NON-R3000 TRANSLATION CACHE INVALIDATION BOTTLENECKS!
8 dpavlin 26 x) Find a way to get rid of the cpu_create_or_reset_tc in the
9     R2000/R3000 cache isolation code. (NetBSD works without it,
10     but not Ultrix and Linux yet.)
11     x) Formalize the statistics gathering stuff for dyntrans...
12     x) ... and use it to optimize MIPS dyntrans stuff.
13 dpavlin 24 x) Clock framework? Go through all clock devices, make sure they
14     return correct data, and run at correct speeds!
15     x) Optimizations, continuing on 64-bit issues etc with dyntrans
16     x) Dyntrans with SMP... lots of work to be done here.
17     x) Dyntrans with cache emulation... lots of work here as well.
18     x) Actually use the settings object, better debugger stuff, etc.
19     x) Wait for new releases of NetBSD, and test with those.
20    
21     --------------
22    
23     SMP:
24     o) dev_mp doesn't work well with dyntrans yet
25     o) In general, IPIs, CAS, LL/SC etc must be made to work with dyntrans
26    
27     MIPS:
28     o) Fix invalidate_asid so it works well for non-R3000 too!
29     x) [Re]add an interrupt-asserted bit for MIPS, to speed up
30     interrupt handling slightly?
31     +) Print a warning on the first reserved instruction.
32     +) Some more work on opcodes.
33     x) MIPS64 revision 2.
34     x) _MAYBE_ TX79 and R5900 actually differ in their
35     opcodes? Check this carefully!
36     o) Dyntrans: Count register updates are probably not 100% correct yet.
37     o) Dyntrans: SMP correctness
38     o) Refactor code for performance and readability/maintainability.
39     o) Instruction combinations? Possible candidates (but profile first!):
40     o) multiple loads/stores in a row
41     o) strlen, memset loops etc
42     o) compare + branch
43     o) DROTR32 and similar MIPS64 rev 2 instructions, which have
44     a rotation bit which differs from previous ISAs.
45     o) EI and DI instructions for MIPS64/32 rev 2. NOTE: These are
46     _NOT_ the same as for R5900!
47     o) R4000 and others:
48     x) watchhi/watchlo exceptions, and other exception
49     handling details
50     o) R10000 and others: (R12000, R14000 ?)
51     x) memory space, exceptions, ...
52     x) use cop0 framemask for tlb lookups
53     (http://techpubs.sgi.com/library/tpl/cgi-bin/getdoc.cgi/hdwr/bks/SGI_Developer/books/R10K_UM/sgi_html/t5.Ver.2.0.book_284.html)
54    
55 dpavlin 14 Dyntrans:
56 dpavlin 24 x) Move the mips_init_64bit_dummy_tables() etc calls into
57     src/cpu.c, for all 64-bit cpus?
58     x) 64-bit "phystranslation" lookup as in 32-bit mode? Would probably
59     help performance a bit.
60     x) Common fatal_abort() function, which drops into the debugger
61     without continuing.
62 dpavlin 22 x) INVALIDATION should cause translations in _all_ cpus to be
63     invalidated, e.g. on a write to a write-protected page
64     (containing code)
65 dpavlin 24 x) better (formally defined) instr call statistics (-s command
66     line option?), multiple different types? (virtual pc, physical pc)
67 dpavlin 22 x) Call/return hints?
68     x) 16-bit encodings? (MIPS16, ARM Thumb, SH3, ...)
69 dpavlin 24 x) H8?
70 dpavlin 18 x) Lots of other stuff: see src/cpus/README_DYNTRANS
71 dpavlin 22 x) true recompilation backend? think carefully about this,
72     experiment in a separate project (not in GXemul)
73 dpavlin 24 x) Remove the dyntrans_alignment_check functionality; although
74     it gives slightly higher peformance sometimes, it increases
75     the complexity of the code too much!
76 dpavlin 18
77 dpavlin 24 Alpha:
78     o) Virtual memory (tlbs etc)
79     o) Get {NetBSD,OpenBSD,Linux}/alpha booting. :)
80    
81     SPARC:
82     o) Add all registers (floating point, control regs etc)
83     o) Save/restore register windows etc!
84     o) Disassemly of some more instructions.
85     o) Are sll etc 32-bit sign-extending or zero-extending?
86     o) Finish the cmp (subcc) flag computation code.
87     o) Finish the GDB register stuff.
88    
89     Debugger:
90     o) How does SMP debugging work? Does it simply use "threads"?
91     What if the guest OS (running on an emulated SMP machine)
92     has a usertask running, with userland threads?
93     o) Try to make the debugger more modular and, if possible, reentrant!
94     o) Remove the emul command? (But show network info if showing
95     machines?)
96     o) Generalize the expression evaluator. (debugger_expr.c?)
97     settable variables ("show nr of instructions on average")
98     emul[x] defaults to current emul
99     machine[x] defaults to current machine
100     cpu[x] defaults to currently focused cpu
101     registers cpu arch dependent (#-prefix)
102     symbols @-prefix
103     numeric constants decimal, hex, and octal ($-prefix)
104     boolean yes,no, true,false
105     operators (+ - * / % & | ^ !)
106     parentheses for grouping subexpressions
107     NOTE: the change from % to # for register prefix!
108     examples:
109     emul[0].machine[2].cpu[0].pc
110     machine[test2].cpu[1].ra = main
111     settings.show_trace_tree = yes
112    
113     Settings:
114     o) Remove a setting.
115     o) Read/write a setting given a name. (Read as
116     string and/or int64_t simultaneously?)
117    
118     Help command should have subsections! One for "expressions",
119     mirrored in the documentation, but the internal help should
120     be the one that should be considered correct.
121     o) see src/debugger.c for more
122    
123     POWER/PowerPC:
124     x) PPC optimizations; instr combs
125     x) 64-bit stuff
126     x) find and fix the bug which causes NetBSD/macppc to fail after
127     an install!
128     x) macppc: adb controller; keyboard (for framebuffer mode)
129     x) make OpenBSD/macppc work (PCI controller stuff)
130    
131     Algor:
132     PCI and ISA and LOCAL interrupts! --> wdc could start working
133     Add interrupt controller in dev_algor.c.
134    
135     ARM:
136     o) try to get netbsd/evbarm 3.x running (iq80321)
137     o) make the xscale counter registers (ccnt) work
138     o) make the ata controller usable for FreeBSD!
139     o) zaurus for openbsd...
140     o) debian/cats crashes because of unimplemented coproc stuff.
141     fix this?
142    
143     Cache simulation:
144     o) Separate from the CPU concept, so that multi-core CPUs sharing
145     e.g. a L2 cache can be simulated (?)
146     o) Instruction cache emulation is easiest (if separate from the
147     data cache); similar hack as the S;I; hack in cpu_dyntrans.c.
148     NOTE: if the architecture has a delay slot, then an instruction
149     slot can actually be executed as 2 instructions.
150     o) Data cache emulation = harder; each arch's load/store routines
151     must include support? running one instruction at a time and
152     having a cpu-dependant lookup function for each instruction
153     is another option (easier to implement, but very very slow).
154    
155     Documentation:
156     o) machines, cpus, devices.
157     o) Automagic documentation generation:
158     x) REMEMBER that several machines/devices can be in
159     the same source file!
160     o) Try to rewrite the install instructions for those machines
161     that use 3MAX into using CATS? (To remove the need to a raw
162     ffs partition using up all of the disk image.)
163    
164 dpavlin 22 More generic out_of_memory error reporting, and check everywhere!
165     Causes: OpenBSD has low default limits for normal users.
166     Host is 32-bit? (32-bit hosts are limited to 4 GB or less
167     of userspace memory.)
168     You are actually low on RAM. (As trivial as this might sound,
169     Unix systems usually allow processes to allocate virtual
170     memory beyond the amount of RAM in the machine.)
171    
172     Breakpoints: 32-bit vs 64-bit sign extension for MIPS, warnings, etc.
173     Use the debugger's symbolic name stuff. (which will have to be
174     extended soon to support stuff like "2*x + symbol + y" etc. cool
175     stuff)
176    
177     Sprite (guest OS for DECstation emulation)
178     x) Timing problems during bootup?
179    
180     The Device subsystem:
181     x) allow devices to be moved and/or changed in size (down to a
182     minimum size, etc, or up to a max size)
183     x) keep track of interrupts and busses? actually, allowing any device
184 dpavlin 24 to be a bus might be a nice idea.
185     x) turn interrupt controllers into devices? :-)
186 dpavlin 22 x) refactor various clocks/nvram/cmos into one device?
187    
188 dpavlin 24 Clocks:
189     x) General framework for automagic clock adjustment for _all_
190     kinds of clocks and timers. (Which should be possible to turn
191     off, of course, like the way DECstation emulation works now.)
192 dpavlin 22
193     PCI:
194 dpavlin 24 x) last write was ffffffff ==> fix this, it should be used
195     together with a mask to get the correct bits. also, not ALL
196     bits are size bits! (lowest 4 vs lowest 2?)
197 dpavlin 22 x) add support for address fixups
198     x) generalize the interrupt routing stuff (lines etc). this should
199     be per machine? or per bus, that's better
200 dpavlin 24 x) add a "pcn" NIC (AMD PCnet32 Lance 79c970 (PCI 1022:2000)),
201     could be useful for several machine modes (Malta, Algor, evbarm,
202     hp700?, macppc, etc.)
203 dpavlin 22
204     Network layer:
205     o) DHCP (for Debian and BSD installers :-)
206     o) increase performance
207     o) don't rely on NetBSD-ish usage
208     o) Multiple networks per emulation, and let different
209     NICs in machines connect to different networks.
210     o) many other issues: see src/net.c
211    
212     Busses:
213     o) Redesign the entire "mainbus" concept!
214     o) Busses should be placed in a hierarchical tree!
215     o) Easily configurable interrupt routing in SMP systems.
216     o) Specific clock/bus speeds, cpu speeds etc.
217     o) Synchronization over network? or at least in dyntrans within
218     one emulated machine
219     o) dev->bus: TurboChannel, PCMCIA, ADB?
220    
221     Config file parser:
222 dpavlin 24 o) Rewrite it from scratch!
223 dpavlin 22 o) Usage of any expression available through the debugger
224     o) Support for running debugger commands (like the -c
225     command line option)
226    
227     Floating point layer:
228     o) make it common enough to be used by _all_ emulation modes
229 dpavlin 24 o) implement more stuff
230 dpavlin 22 o) non-IEEE modes (i.e. x86)?
231    
232 dpavlin 14 Userland emulation:
233 dpavlin 22 x) Lots of stuff; freebsd and netbsd (and linux?) syscalls.
234 dpavlin 14 x) Dynamic linking? Hm.
235 dpavlin 12
236 dpavlin 22 Sound:
237     x) generic sound framework
238     x) add one or more sound cards as devices
239 dpavlin 12
240 dpavlin 24 ASC SCSI controller:
241     x) NetBSD/arc 2.0 uses the ASC controller in a way which GXemul
242     cannot yet handle. (NetBSD 1.6.2 works ok.) (Possibly a problem
243     in NetBSD itself, http://mail-index.netbsd.org/source-changes/
244     2005/11/06/0024.html suggests that.)
245    
246 dpavlin 22 Caches / memory hierarchies: (this is mostly MIPS-specific)
247     o) MIPS coproc.c: bits in config registers should reflect
248     correct cache sizes for _all_ CPU types. (currently only
249     implemented for R4000, R1x000, and a few others)
250     o) src/memory*.c: Implement correct cache emulation for
251     all CPU types. (currently only R2000/R3000 is implemented)
252     (per CPU, multiple levels should be possible,
253     associativity etc!)
254     o) R2000/R3000 isn't _100%_ correct, just almost correct :)
255     o) Move the -S (fill mem with random) functionality into the
256     memory.c subsystem, not machine.c or wherever it is now
257     o) ECC stuff, simulation of memory errors? (Machine dependent)
258     o) More than 4GB of emulated RAM, when run on a 32-bit host?
259     (using manual swap-out of blocks to disk, ugly)
260     o) A global command line option should be used to turn
261     cache emulation on or off. When off, caches should be
262     faked like they are right now. When on, caches and
263     memory latencies should be emulated as correctly as
264     possible.
265 dpavlin 2
266 dpavlin 22 File/disk/symbol handling:
267     o) Better handling of tape files
268     o) Read function argument count and types from binaries? (ELF?)
269 dpavlin 24 o) Better demangling of C++ names. Note: GNU's C++ differs from e.g.
270     Microsoft's C++, so multiple schemes must be possible. See
271     URL at top of src/symbol_demangle.c for more info.
272 dpavlin 2
273 dpavlin 22 Userland ABI emulation:
274     o) see src/useremul.c
275 dpavlin 12
276 dpavlin 22 Terminal/console:
277     o) allow emulated serial ports to be connected to the outside
278     world in a more generic way, or even to other emulated
279     machines(?)
280 dpavlin 2
281 dpavlin 22 Save state of the whole emulated machine, to be able to load it back
282     in later? (Memory, all device's states, all registers and
283     so on. Like taking a snapshot. (SimOS seems to do this,
284     according to its website.))
285 dpavlin 6
286 dpavlin 22 Better framebuffer and X-windows functionality:
287     o) -Yx sometimes causes crashes.
288     o) Simple device access to framebuffer_blockcopyfill() etc,
289     and text output (using the built-in fonts), for dev_fb.
290     o) CLEAN UP the ugly event code
291     o) Mouse clicks can be "missed" in the current system; this is
292     not good. They should be put on a stack of some kind.
293     o) More 2D and 3D framebuffer acceleration.
294     o) Non-resizable windows? Or choose scaledown depending
295     on size (and center the image, with a black border).
296     o) Different scaledown on different windows?
297     o) Switch scaledown during runtime? (Ala CTRL-ALT-plus/minus)
298 dpavlin 24 o) Bug reported by Elijah Rutschman on MacOS with weird
299     keys (F5 = cursor down?).
300 dpavlin 22 o) Keyboard and mouse events:
301     x) Do this for more machines than just DECstation
302     x) more X11 cursor keycodes
303     x) Keys like CTRL, ALT, SHIFT do not get through
304     by themselves (these are necessary for example
305     to change the font of an xterm in X in the
306     emulator)
307     o) Generalize the framebuffer stuff by moving _ALL_ X11
308     specific code to src/x11.c!
309 dpavlin 2
310 dpavlin 22 Statistics: (this could be interesting)
311     o) Save to file and show graphics. It should be possible to
312     run gxemul after a simulation to just show the graphics,
313     or convert to a .ppm or .tga or similar.
314     o) memory accesses (to measure cache efficiency and
315     page coloring efficiency)
316     o) nr of simultaneous ASIDs in use in the TLB, for MIPS
317     o) percentage of time spent in different "states", such as
318     running userland code, kernel code, or idling (for CPUs
319     that have such an instruction, or whenever the PC is
320     inside a specific idle-function (address range)).
321     Possible additional state (for example on R3000): caches
322     disabled.
323     o) position of read/write on (SCSI) disks
324 dpavlin 2

  ViewVC Help
Powered by ViewVC 1.1.26