/[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 44 - (hide annotations)
Mon Oct 8 16:22:56 2007 UTC (12 years, 3 months ago) by dpavlin
File size: 24172 byte(s)
++ trunk/HISTORY	(local)
$Id: HISTORY,v 1.1632 2007/09/11 21:46:35 debug Exp $
20070616	Implementing the MIPS32/64 revision 2 "ror" instruction.
20070617	Adding a struct for each physpage which keeps track of which
		ranges within that page (base offset, length) that are
		continuously translatable. When running with native code
		generation enabled (-b), a range is added after each read-
		ahead loop.
		Experimenting with using the physical program counter sample
		data (implemented 20070608) together with the "translatable
		range" information, to figure out which physical address ranges
		would be worth translating to native code (if the number of
		samples falling within a range is above a certain threshold).
20070618	Adding automagic building of .index comment files for
		src/file/, src/promemul/, src src/useremul/ as well.
		Adding a "has been translated" bit to the ranges, so that only
		not-yet-translated ranges will be sampled.
20070619	Moving src/cpu.c and src/memory_rw.c into src/cpus/,
		src/device.c into src/devices/, and src/machine.c into
		src/machines/.
		Creating a skeleton cc/ld native backend module; beginning on
		the function which will detect cc command line, etc.
20070620	Continuing on the native code generation infrastructure.
20070621	Moving src/x11.c and src/console.c into a new src/console/
		subdir (for everything that is console or framebuffer related).
		Moving src/symbol*.c into a new src/symbol/, which should
		contain anything that is symbol handling related.
20070624	Making the program counter sampling threshold a "settings
		variable" (sampling_threshold), i.e. it can now be changed
		during runtime.
		Switching the RELEASE notes format from plain text to HTML.
		If the TMPDIR environment variable is set, it is used instead
		of "/tmp" for temporary files.
		Continuing on the cc/ld backend: simple .c code is generated,
		the compiler and linker are called, etc.
		Adding detection of host architecture to the configure script
		(again), and adding icache invalidation support (only
		implemented for Alpha hosts so far).
20070625	Simplifying the program counter sampling mechanism.
20070626	Removing the cc/ld native code generation stuff, program
		counter sampling, etc; it would not have worked well in the
		general case.
20070627	Removing everything related to native code generation.
20070629	Removing the (practically unusable) support for multiple
		emulations. (The single emulation allowed now still supports
		multiple simultaneous machines, as before.)
		Beginning on PCCTWO and M88K interrupts.
20070723	Adding a dummy skeleton for emulation of M32R processors.
20070901	Fixing a warning found by "gcc version 4.3.0 20070817
		(experimental)" on amd64.
20070905	Removing some more traces of the old "multiple emulations"
		code.
		Also looking in /usr/local/include and /usr/local/lib for
		X11 libs, when running configure.
20070909	Minor updates to the guest OS install instructions, in
		preparation for the NetBSD 4.0 release.
20070918	More testing of NetBSD 4.0 RC1.

1 dpavlin 44 $Id: TODO,v 1.556 2007/09/11 21:46:35 debug Exp $
2 dpavlin 2
3 dpavlin 42 Some things, in no specific order, that I'd like to fix:
4     (Some items in this list are perhaps already fixed.)
5 dpavlin 40
6 dpavlin 44 -------------------------------------------------------------------------------
7    
8     A first pass of installation regression testing of NetBSD 4.0 RC1 in GXemul:
9    
10     X pmax (including X Windows out-of-the-box)
11     X arc (1.6.2 -> 4.0! Yay!)
12     hpcmips
13     cobalt
14     evbmips
15     algor
16     sgimips
17     cats
18     evbarm
19     netwinder
20     prep nej, 2.0 är senaste som funkar :(
21     X macppc
22     pmppc
23     X dreamcast
24     X landisk
25    
26     X = done and worked fine
27    
28     -------------------------------------------------------------------------------
29    
30     Perform a second regression test pass, when the actual NetBSD 4.0 release
31     has happened.
32    
33     o) Test all guest OSes.
34     o) Update:
35     URLs
36     Versions
37    
38     o) Make a new GXemul release: 0.4.6.1
39    
40     -------------------------------------------------------------------------------
41    
42 dpavlin 42 M88K:
43     o) FIP != NIP + 4, in rte! (Simulate delayed branch stuff.)
44     o) cpu_dyntrans.c: MEMORY_USER_ACCESS implementation for M88K!
45     o) xmem: Set transaction registers!
46     o) CMMUs:
47     o) Translation invalidations, could be optimized.
48     o) Move initialization from dev_mvme187 to somewhere
49     more reasonable?
50     o) Instruction trace by using bits of ??IP control regs.
51     o) Interrupts (these are machine dependent, though).
52     o) Implement devices etc. for one or more machine modes,
53     to get some guest OS running. OpenBSD/mvme88k on MVME187
54     seems to be the smartest path to follow for now.
55     o) VME bus device
56     o) PCC2
57     o) Cirrus Logic serial port controller
58     o) Instruction disassembly, and implementation:
59     o) See http://www.panggih.staff.ugm.ac.id/download/GCC/info/gcc.i5
60     for some strange cases of when "div" can fail (?)
61     o) Floating point stuff
62     o) "Graphics" instructions (M88110-specific)
63 dpavlin 40
64 dpavlin 24 MIPS:
65 dpavlin 32 o) Nicer MIPS status bits in register dumps.
66     o) Floating point exception correctness.
67     o) Fix this? Triggered by NetBSD/sgimips? Hm:
68     to_be_translated(): TODO: unimplemented instruction:
69     000000000065102c: 00200800 (d) rot_00 at,zr,0
70     o) Some more work on opcodes.
71 dpavlin 24 x) MIPS64 revision 2.
72 dpavlin 28 o) Find out which actual CPUs implement the rev2 ISA!
73 dpavlin 42 o) DINS, DINSM, DINSU etc
74 dpavlin 32 o) DROTR32 and similar MIPS64 rev 2 instructions,
75     which have a rotation bit which differs from
76     previous ISAs.
77 dpavlin 24 x) _MAYBE_ TX79 and R5900 actually differ in their
78     opcodes? Check this carefully!
79     o) Dyntrans: Count register updates are probably not 100% correct yet.
80 dpavlin 28 o) (Re)implement 128-bit loads/stores for R5900.
81 dpavlin 42 o) Coprocessor 1x (i.e. 3) should cause cp1 exceptions, not 3?
82     (See http://lists.gnu.org/archive/html/qemu-devel/2007-05/msg00005.html)
83 dpavlin 24 o) R4000 and others:
84     x) watchhi/watchlo exceptions, and other exception
85     handling details
86 dpavlin 42 o) MIPS 5K* have 42 physical address bits, not 40/44?
87 dpavlin 24 o) R10000 and others: (R12000, R14000 ?)
88 dpavlin 34 x) The code before the line
89     /* reg[COP0_PAGEMASK] = cpu->cd.mips.coproc[0]->tlbs[0].mask & PAGEMASK_MASK; */
90     in cpu_mips.c is not correct for R10000 according to
91     Lemote's Godson patches for GXemul. TODO: Go through all
92     register definitions according to http://techpubs.sgi.com/library/tpl/cgi-bin/getdoc.cgi/hdwr/bks/SGI_Developer/books/R10K_UM/sgi_html/t5.Ver.2.0.book_263.html#HEADING334
93     and make sure everything works with R10000.
94     Then test with OpenBSD/sgi?
95 dpavlin 36 x) Entry LO mask (as above).
96 dpavlin 24 x) memory space, exceptions, ...
97     x) use cop0 framemask for tlb lookups
98     (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)
99    
100 dpavlin 32 SuperH:
101 dpavlin 42 x) Auto-generation of loads/stores! This should get rid of at least
102     the endianness check in each load/store.
103     x) Experiment with whether or not correct ITLB emulation is
104     actually needed. (20070522: I'm turning it off today.)
105 dpavlin 32 x) SH4 interrupt controller:
106 dpavlin 38 x) MASKING should be possible!
107 dpavlin 34 x) SH4 DMA (0xffa00000)
108     x) SH4 UBC (0xff200000)
109     x) Store queues can copy 32 bytes at a time, there's no need to
110     copy individual 32-bit words. (Performance improvement.)
111 dpavlin 32 x) SH4 BSC (Bus State Controller)
112 dpavlin 34 x) Instruction tracing should include symbols for branch targets,
113     and so on, to make the output more human readable.
114 dpavlin 38 x) SH3-specific devices: Pretty much everything!
115 dpavlin 42 x) NetBSD/evbsh3, hpcsh! Linux?
116 dpavlin 34 x) Floating point speed!
117 dpavlin 32 x) Floating point exception correctness.
118 dpavlin 40 x) Exceptions for unaligned load/stores. OpenBSD/landisk uses
119     this mechanism for its reboot code (machine_reset).
120 dpavlin 18
121 dpavlin 38 Landisk SH4:
122 dpavlin 40 x) When NetBSD/landisk 4.0 has been released, make sure it works
123     in the emulator. (Update documentation, etc.)
124 dpavlin 44 NetBSD HEAD (as of April 2007) hangs during bootup, because it
125     turns on/off interrupts in an unfortunately synchronized way
126     with dyntrans. This needs to be fixed.
127 dpavlin 38
128 dpavlin 32 Dreamcast:
129 dpavlin 34 x) G2 DMA
130     x) LAN adapter (dev_mb8696x.c). NetBSD root-on-nfs.
131 dpavlin 32 x) PVR: Lots of stuff. See dev_pvr.c.
132 dpavlin 38 x) Better GDROM support
133 dpavlin 34 x) Modem
134     x) PCI bridge/bus?
135 dpavlin 32 x) Maple bus:
136     x) Correct controller input
137     x) Mouse input
138 dpavlin 34 x) Software emulation of BIOS calls:
139     x) GD-ROM emulation: Use the GDROM device.
140     x) Use the VGA font as a fake ROM font. (Better than
141     nothing.)
142     x) Make as many as possible of the KOS examples run!
143 dpavlin 32 x) More homebrew demos/games.
144 dpavlin 34 x) SPU: Sound emulation (ARM cpu).
145     x) VME processor emulation? "(Sanyo LC8670 "Potato")" according to
146     Wikipedia, LC86K87 according to Comstedt's page. See
147     http://www.maushammer.com/vmu.html for a good description of
148     the differences between LC86104C and the one used in the VME.
149 dpavlin 32
150 dpavlin 24 Alpha:
151 dpavlin 32 x) OSF1 PALcode, Virtual memory support.
152     x) PALcode replacement! PAL1E etc opcodes...?
153     x) Interrupt/exception/trap handling.
154     x) Floating point exception correctness.
155     x) More work on bootup memory and register contents.
156     x) More Alpha machine types, so it could work with
157     OpenBSD, FreeBSD, and Linux too?
158 dpavlin 24
159 dpavlin 38 SPARC (both the ISA and the machines):
160 dpavlin 32 o) Implement Adress space identifiers; load/stores etc.
161 dpavlin 38 o) Exception/trap/interrupt handling.
162 dpavlin 34 o) Save/restore register windows etc! Both v9 and pre-v9!
163 dpavlin 30 o) Finish the subcc and addcc flag computation code.
164     o) Add more registers (floating point, control regs etc)
165 dpavlin 28 o) Disassemly of some more instructions?
166 dpavlin 24 o) Are sll etc 32-bit sign-extending or zero-extending?
167 dpavlin 38 o) Floating point exception correctness.
168 dpavlin 28 o) SPARC v8, v7 etc?
169 dpavlin 38 o) More machine modes and devices.
170 dpavlin 24
171     POWER/PowerPC:
172 dpavlin 34 x) Fix DECR timer speed, so it matches the host.
173 dpavlin 32 x) NetBSD/prep 3.x triggers a possible bug in the emulator:
174     <wdc_exec_command(0xd005e514,0xd60cdd30,0,8,..)>
175     <ata_get_xfer(0,0xd60cdd30,0,8,..)>
176     <0x26c550(&ata_xfer_pool,2,0,8,..)>
177     <0x35c71c(0x3f27000,0,52,8,..)>
178     <ata_exec_xfer(0xd005e4c8,0x3f27000,0,13,..)>
179     <atastart(0xd005e4c8,0x3f27000,0,13,..)>
180     <__wdccommand_start(0xd005e4c8,0x3f27000,0,13,..)>
181     <bsw1(&prep_isa_io_space_tag,0x800001f6,0,176,..)>
182     [ wdc: write to SDH: 0xb0 (sectorsize 2, lba=1, drive 1, head 0) ]
183     <wdcwait(0xd005e4c8,72,64,0xbb8,..)>
184     <0x198120(0xd005e4c8,72,64,0xbb8,..)>
185     <bsr1(&prep_isa_io_space_tag,0,0,0xbb8,..)>
186     <delay(100,0,0,0xbb8,..)>
187     Note: <bsr1(&prep_isa_io_space_tag,0,0,0xbb8,..)>
188 dpavlin 24 x) PPC optimizations; instr combs
189 dpavlin 28 x) 64-bit stuff: either Linux on G5, or perhaps some hobbyist
190     version of AIX? (if there exists such a thing)
191 dpavlin 24 x) macppc: adb controller; keyboard (for framebuffer mode)
192     x) make OpenBSD/macppc work (PCI controller stuff)
193 dpavlin 32 x) Floating point exception correctness.
194     x) Alignment exceptions.
195 dpavlin 24
196 dpavlin 34 PReP:
197 dpavlin 38 x) Clock time! ("Bad battery blah blah")
198 dpavlin 34
199 dpavlin 24 Algor:
200 dpavlin 32 o) Other models than the P5064?
201     o) PCI interrupts... needed for stuff like the tlp NIC?
202 dpavlin 24
203 dpavlin 34 BeBox:
204     o) Interrupts. There seems to be a problem with WDC interrupts
205     "after a short while", although a few interrupts get through?
206     o) Perhaps find a copy of BeOS and try it?
207    
208 dpavlin 44 Malta:
209     o) The Linux/Malta kernel at people.debian.org/~ths/qemu/malta/
210     almost works:
211     ./gxemul -x -o 'rd_start=0x80800000 rd_size=10000000 init=/bin/sh' -C 4KEc
212     -e malta 0x80800000:people.debian.org/~ths/qemu/malta/initrd.gz
213     people.debian.org/~ths/qemu/malta/vmlinux
214     (Remove "init=/bin/sh" to boot into the Debian installer.)
215     There are at least two things that need to be fixed:
216     1. PCI IDE; make Linux oops.
217     2. Implement the NIC.
218    
219 dpavlin 32 HPCmips:
220     x) Mouse/pad support! :)
221     x) A NIC? (As a PCMCIA device?)
222    
223 dpavlin 24 ARM:
224 dpavlin 32 o) See netwinder_reset() in NetBSD; the current "an internal error
225     occured" message after reboot/halt is too ugly.
226 dpavlin 42 o) Generic ARM "wait"-like instruction?
227 dpavlin 34 o) try to get netbsd/evbarm 3.x or 4.x running (iq80321)
228 dpavlin 24 o) make the xscale counter registers (ccnt) work
229     o) make the ata controller usable for FreeBSD!
230 dpavlin 34 o) Debian/cats crashes because of unimplemented coproc stuff.
231 dpavlin 24 fix this?
232    
233 dpavlin 32 Test machines:
234 dpavlin 34 o) dev_fb block fill and copy
235     o) dev_fb draw characters (from the built-in font)?
236     o) dev_fb input device? mouse pointer coordinates and buttons
237 dpavlin 32 (allow changes in these to cause interrupts as well?)
238 dpavlin 34 o) Redefine the halt() function so that it stops "sometimes
239     soon", i.e. usage in demo code should be:
240 dpavlin 32 for (;;) {
241     halt();
242     }
243    
244 dpavlin 42 Debugger:
245     o) How does SMP debugging work? Does it simply use "threads"?
246     What if the guest OS (running on an emulated SMP machine)
247     has a usertask running, with userland threads?
248     o) Try to make the debugger more modular and, if possible, reentrant!
249     o) Memory dumps should be able to dump both physical and
250     virtual emulated memory.
251     o) Evaluate expressions within []? That would allow stuff like
252     cpu[x] where x is an expression.
253     o) "pc = pc + 4" doesn't work! Bug. Should work. ("pc=pc+4" works.)
254     o) Settings:
255     x) Special handlers for Write!
256     +) MIPS coproc regs
257     +) Alpha/MIPS/SPARC zero registers
258     +) x86 64/32/16-bit registers
259     x) Value formatter for resulting output.
260     o) Call stack display (back-trace) of emulated programs.
261     o) Nicer looking output of register dumps, floating point registers,
262     etc. Warn about weird/invalid register contents.
263     o) Ctrl-C doesn't enter the debugger on some OSes (HP-UX?)...
264    
265     Dyntrans:
266 dpavlin 44 x) NOTE: ARM etc. that load pc-relative constants, on writes to
267     pages that contain translations, the ENTIRE page must be
268     invalidated, not just the 1/32th that was code.
269 dpavlin 42 x) For 32-bit emulation modes, that have emulated TLBs: tlbindex
270     arrays of mapped pages? Things to think about:
271     x) Only 32-bit mode! (64-bit => too much code)
272     x) One array for global pages, and one array _PER ASID_,
273     for those archs that support that. On M88K, there should
274     be one array for userspace, and one for supervisor, etc.
275     x) Larger-than-4K-pages must fill several bits in the array.
276     x) No TLB search will be necessary.
277     x) Total host space used, for 4 KB pages: 1 MB per table,
278     i.e. 65 MB for 32-bit MIPS, 2 MB for M88K, if one byte
279     is used as the tlb index.
280     x) (The index is actually +1, so that 0 means no hit.)
281     x) "Merge" the cur_physpage and cur_ic_page variables/pointers to
282     one? I.e. change cur_ic_page to cur_physpage.ic_page or something.
283     x) Instruction combination collisions? How to avoid easily...
284 dpavlin 44 x) superh -- no hostpage for e.g. 0x8c000000. devices as ram!
285 dpavlin 42 x) Think about how to do both SHmedia and SHcompact in a reasonable
286     way! (Or AMD64 long/protected/real, for that matter.)
287     x) 68K emulation; think about how to do variable instruction
288     lengths across page boundaries.
289     x) Dyntrans with valgrind-inspired memory checker. (In memory_rw,
290     it would be reasonably simple to add; in each individual fast
291     load/store routine = a lot more work, and it would become
292     kludgy very fast.)
293     x) Dyntrans with SMP... lots of work to be done here.
294     x) Dyntrans with cache emulation... lots of work here as well.
295     x) Remove the concept of base RAM completely; it would be more
296     generic to allow RAM devices to be used "anywhere".
297     o) dev_mp doesn't work well with dyntrans yet
298     o) In general, IPIs, CAS, LL/SC etc must be made to work with dyntrans
299     x) Redesign/rethink the delay slot mechanism used for e.g. MIPS,
300     so that it caches a translation (that is, an instruction
301     word and the instr_call it was translated to the last
302     time), so that it doesn't need to do slow
303     to_be_translated for each end of page?
304     x) Program Counter statistics:
305     Per machine? What about SMP? All data to the same file?
306     A debugger command should be possible to use to enable/
307     disable statistics gathering.
308     Configuration file option!
309     x) Breakpoints:
310     o) Physical vs virtual addresses!
311     o) 32-bit vs 64-bit sign extension for MIPS, and others?
312     x) INVALIDATION should cause translations in _all_ cpus to be
313     invalidated, e.g. on a write to a write-protected page
314     (containing code)
315 dpavlin 44 x) 16-bit encodings? (MIPS16, ARM Thumb, etc)
316 dpavlin 42 x) Lots of other stuff: see src/cpus/README_DYNTRANS
317 dpavlin 44 x) Native code generation backends... think carefully about this.
318 dpavlin 42
319     Simple Valgrind-like checks?
320     o) Mark every address with bits which tell whether or not the address
321     has been written to.
322     o) What should happen when programs are loaded? Text/data, bss (zero
323     filled). But stack space and heap is uninitialized.
324     o) Uninitialized local variables:
325     A load from a place on the stack which has not previously
326     been stored to => warning. Increasing the stack pointer using
327     any available means should reset the memory to uninitialized.
328     o) If calls to malloc() and free() can be intercepted:
329     o) Access to a memory area after free() => warning.
330     o) Memory returned by malloc() is marked as not-initialized.
331     o) Non-passive, but good to have: Change the argument
332     given to malloc, to return a slightly larger memory
333     area, i.e. margin_before + size + margin_after,
334     and return the pointer + margin_before.
335     Any access to the margin_before or _after space results
336     in warnings. (free() must be modified to free the
337     actually allocated address.)
338    
339 dpavlin 32 Better CD Image file support:
340     x) Support CD formats that contain more than 1 track, e.g.
341     CDI files (?). These can then contain a mixture of e.g. sound
342     and data tracks, and booting from an ISO filesystem path
343     would boot from [by default] the first data track.
344     (This would make sense for e.g. Dreamcast CD images, or
345     possibly other live-CD formats.)
346    
347     Networking:
348 dpavlin 38 x) Redesign of the networking subsystem, at least the NAT translation
349     part. The current way of allowing raw ethernet frames to be
350     transfered to/from the emulator via UDP should probably be
351     extended to allow the frames to be transmitted other ways as
352     well.
353     x) Also adding support for connecting ttys (either to xterms, or to
354     pipes/sockets etc, or even to PPP->NAT or SLIP->NAT :-).
355     x) Documentation updates (!) are very important, making it easier to
356     use the (already existing) network emulation features.
357 dpavlin 32 x) Fix performance problems caused by only allowing a
358     single TCP packet to be unacked.
359     x) Don't hardcode offsets into packets!
360     x) Test with lower than 100 max tcp/udp connections,
361     to make sure that reuse works!
362     x) Make OpenBSD work better as a guest OS!
363     x) DHCP? Debian doesn't actually send DHCP packets, even
364     though it claims to? So it is hard to test.
365     x) Multiple networks per emulation, and let different
366     NICs in machines connect to different networks.
367     x) Support VDE (vde.sf.net)? Easiest/cleanest (before a
368     redesign of the network framework has been done) is
369     probably to connect it using the current (udp) solution.
370     x) Allow SLIP connections, possibly PPP, in addition to
371     ethernet?
372    
373 dpavlin 24 Cache simulation:
374 dpavlin 28 o) Command line flags for:
375     o) CPU endianness?
376     o) Cache sizes? (multiple levels)
377 dpavlin 24 o) Separate from the CPU concept, so that multi-core CPUs sharing
378     e.g. a L2 cache can be simulated (?)
379     o) Instruction cache emulation is easiest (if separate from the
380     data cache); similar hack as the S;I; hack in cpu_dyntrans.c.
381     NOTE: if the architecture has a delay slot, then an instruction
382     slot can actually be executed as 2 instructions.
383     o) Data cache emulation = harder; each arch's load/store routines
384     must include support? running one instruction at a time and
385     having a cpu-dependant lookup function for each instruction
386     is another option (easier to implement, but very very slow).
387    
388     Documentation:
389 dpavlin 42 x) Update the documentation regarding the testmachine interrupts.
390 dpavlin 32 x) Note about sandboxing/security:
391     Not all emulated instructions fail in the way they would
392     do on real hardware (e.g. a userspace program writing to
393     a system register might work in GXemul, but it would
394     fail on real hardware). Sandbox = contain from the
395     host OS. But the emulated programs will run "less
396     securely".
397     x) Try NetBSD/arc 4.x! (It seems to work with disk images!)
398     x) NetBSD/pmax 4 install instructions: xterm instead of vt100!
399     x) Rewrite the section about experimental devices, after the
400     framebuffer acceleration has been implemented, and demos
401     written. (Symbolic names instead of numbers; example
402     use cases, etc. Mention demo files that use the various
403     features?)
404     x) "a very simple linear framebuffer device (for graphics output)"
405     under "which machines does gxemul emulate" ==> better
406     description?
407 dpavlin 24
408 dpavlin 22 The Device subsystem:
409     x) allow devices to be moved and/or changed in size (down to a
410 dpavlin 32 minimum size, etc, or up to a max size); if there is a collision,
411     return false. It is up to the caller to handle this situation!
412     x) NOTE: Translations must be invalidated, both for
413     registering new devices, and for moving existing ones.
414     cpu->invalidate translation caches, for all CPUs that
415     are connected to a specific memory.
416 dpavlin 22
417     PCI:
418 dpavlin 38 x) Pretty much everything related to runtime configuration, device
419     slots, interrupts, etc must be redesigned/cleaned up. The current
420     code is very hardcoded and ugly.
421     o) Allow cards to be added/removed during runtime more easily.
422     o) Allow cards to be enabled/disabled (i/o ports, etc, like
423     NetBSD needs for disk controller detection).
424     o) Allow devices to be moved in memory during runtime.
425     o) Interrupts per PCI slot, etc. (A-D).
426     o) PCI interrupt controller logic... very hard to get right,
427     because these differ a lot from one machine to the next.
428 dpavlin 24 x) last write was ffffffff ==> fix this, it should be used
429     together with a mask to get the correct bits. also, not ALL
430     bits are size bits! (lowest 4 vs lowest 2?)
431 dpavlin 22 x) add support for address fixups
432 dpavlin 32 x) generalize the interrupt routing stuff (lines etc)
433 dpavlin 22
434 dpavlin 32 Clocks and timers:
435 dpavlin 38 x) Fix the PowerPC DECR interrupt speed! (MacPPC and PReP speed, etc.)
436 dpavlin 32 x) DON'T HARDCODE 100 HZ IN cpu_mips_coproc.c!
437 dpavlin 42 x) NetWinder timeofday is incorrect! Huh? grep -R for ta_rtc_read in
438     NetBSD sources; it doesn't seem to be initialized _AT ALL_?!
439 dpavlin 32 x) Cobalt TOD is incorrect!
440     x) Go through all other machines, one by one, and fix them.
441 dpavlin 22
442     Config file parser:
443 dpavlin 24 o) Rewrite it from scratch!
444 dpavlin 22 o) Usage of any expression available through the debugger
445 dpavlin 34 o) Allow interrupt controllers to be added! and interrupts
446     to be used in more ways than before
447 dpavlin 22 o) Support for running debugger commands (like the -c
448     command line option)
449    
450     Floating point layer:
451     o) make it common enough to be used by _all_ emulation modes
452 dpavlin 32 o) implement correct error/exception handling and rounding modes
453     o) implement more helper functions (i.e. add, sub, mul...)
454 dpavlin 22 o) non-IEEE modes (i.e. x86)?
455    
456 dpavlin 14 Userland emulation:
457 dpavlin 42 x) Try to prefix "/emul/mips/" or similar to all filenames,
458     and only if that fails, try the given filename.
459     Read this setting from an environment variable, and only
460     if there is none, fall back to hardcoded string.
461     x) File descriptor (0,1,2) assumptions? Find and fix these?
462 dpavlin 38 x) Dynamic linking!
463     x) Lots of stuff; freebsd, netbsd, linux, ... syscalls.
464     x) Initial register/stack contents (environment, command line args).
465     x) Return value (from main).
466     x) mmap emulation layer
467     x) errno emulation layer
468 dpavlin 44 x) ioctl emulation layer for all devices :-[
469 dpavlin 42 x) struct conversions for many syscalls
470 dpavlin 12
471 dpavlin 22 Sound:
472     x) generic sound framework
473 dpavlin 32 x) add one or more sound cards as devices; add a testmachine
474     sound card first?
475 dpavlin 34 x) Dreamcast sound? Generic PCI sound cards?
476 dpavlin 12
477 dpavlin 24 ASC SCSI controller:
478     x) NetBSD/arc 2.0 uses the ASC controller in a way which GXemul
479     cannot yet handle. (NetBSD 1.6.2 works ok.) (Possibly a problem
480     in NetBSD itself, http://mail-index.netbsd.org/source-changes/
481     2005/11/06/0024.html suggests that.)
482 dpavlin 32 NetBSD 4.x seems to work? :)
483 dpavlin 24
484 dpavlin 22 Caches / memory hierarchies: (this is mostly MIPS-specific)
485     o) src/memory*.c: Implement correct cache emulation for
486     all CPU types. (currently only R2000/R3000 is implemented)
487 dpavlin 28 (per CPU, multiple levels should be possible, associativity etc!)
488 dpavlin 22 o) R2000/R3000 isn't _100%_ correct, just almost correct :)
489     o) Move the -S (fill mem with random) functionality into the
490     memory.c subsystem, not machine.c or wherever it is now
491     o) ECC stuff, simulation of memory errors? (Machine dependent)
492     o) More than 4GB of emulated RAM, when run on a 32-bit host?
493     (using manual swap-out of blocks to disk, ugly)
494     o) A global command line option should be used to turn
495     cache emulation on or off. When off, caches should be
496     faked like they are right now. When on, caches and
497     memory latencies should be emulated as correctly as
498     possible.
499 dpavlin 2
500 dpavlin 22 File/disk/symbol handling:
501 dpavlin 34 o) Make sure that disks can be added/removed during runtime!
502     (Perhaps this needs a reasonably large re-write.)
503 dpavlin 28 o) Remove some of the complexity in file format guessing, for
504     Ultrix kernels that are actually disk images?
505 dpavlin 44 o) Remove temporary files (/tmp/gxemul.blahblah) if loading fails
506     for some reason (unrecognized file, etc).
507 dpavlin 34 o) Better handling of tape files
508 dpavlin 22 o) Read function argument count and types from binaries? (ELF?)
509 dpavlin 24 o) Better demangling of C++ names. Note: GNU's C++ differs from e.g.
510     Microsoft's C++, so multiple schemes must be possible. See
511     URL at top of src/symbol_demangle.c for more info.
512 dpavlin 2
513 dpavlin 22 Userland ABI emulation:
514     o) see src/useremul.c
515 dpavlin 12
516 dpavlin 22 Better framebuffer and X-windows functionality:
517 dpavlin 44 o) Do a complete rewrite of the framebuffer/console stuff, so that:
518     1) It does not rely on X11 specifically.
519     2) It is possible to interact with emulated framebuffers
520     and consoles "remotely", e.g. via a web page which
521     controls multiple virtualized machines.
522     3) It is possible to run on (hypothetical) non-X11
523     graphics systems.
524 dpavlin 32 o) Generalize the update_x1y1x2y2 stuff to an extend-region()
525     function...
526 dpavlin 22 o) -Yx sometimes causes crashes.
527     o) Simple device access to framebuffer_blockcopyfill() etc,
528     and text output (using the built-in fonts), for dev_fb.
529     o) CLEAN UP the ugly event code
530     o) Mouse clicks can be "missed" in the current system; this is
531     not good. They should be put on a stack of some kind.
532     o) More 2D and 3D framebuffer acceleration.
533     o) Non-resizable windows? Or choose scaledown depending
534     on size (and center the image, with a black border).
535     o) Different scaledown on different windows?
536 dpavlin 32 o) Non-integral scale-up? (E.g. 640x480 -> 1024x768)
537 dpavlin 22 o) Switch scaledown during runtime? (Ala CTRL-ALT-plus/minus)
538 dpavlin 24 o) Bug reported by Elijah Rutschman on MacOS with weird
539     keys (F5 = cursor down?).
540 dpavlin 22 o) Keyboard and mouse events:
541     x) Do this for more machines than just DECstation
542     x) more X11 cursor keycodes
543     x) Keys like CTRL, ALT, SHIFT do not get through
544     by themselves (these are necessary for example
545     to change the font of an xterm in X in the
546     emulator)
547     o) Generalize the framebuffer stuff by moving _ALL_ X11
548 dpavlin 44 specific code to a separate module.
549 dpavlin 2
550 dpavlin 42 -------------------------------------------------------------------------------
551    

  ViewVC Help
Powered by ViewVC 1.1.26