/[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

Diff of /trunk/TODO

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 30 by dpavlin, Mon Oct 8 16:20:40 2007 UTC revision 42 by dpavlin, Mon Oct 8 16:22:32 2007 UTC
# Line 1  Line 1 
1  $Id: TODO,v 1.343 2006/08/14 18:46:30 debug Exp $  $Id: TODO,v 1.536 2007/06/15 22:30:17 debug Exp $
2    
3  Hm. This file is in random order, and not all parts of it are up-to-date.  Some things, in no specific order, that I'd like to fix:
4    (Some items in this list are perhaps already fixed.)
5    
6    M88K:
7  Implementation:          o)  Neither NIP nor FIP valid in rte?
8          x)  ARM "wait"-like instruction.          o)  FIP != NIP + 4, in rte! (Simulate delayed branch stuff.)
9          x)  CLOCK FRAMEWORK!          o)  cpu_dyntrans.c: MEMORY_USER_ACCESS implementation for M88K!
10          x)  Mouse support for NetBSD/pmax 4.x!          o)  xmem: Set transaction registers!
11          x)  See netwinder_reset() in NetBSD; the current "an internal error          o)  CMMUs:
12              occured" message after reboot/halt is too ugly.                  o)  Translation invalidations, could be optimized.
13          x)  64-bit ranges in src/cpus/memory_mips_v2p.c                  o)  Move initialization from dev_mvme187 to somewhere
14          x)  Revert the dyntrans page template experiment? Hm.                      more reasonable?
15          x)  Refactor the cpu type detection/initialization/listing.          o)  Instruction trace by using bits of ??IP control regs.
16                  Macro, which can be used as long as the cpu definitions          o)  Interrupts (these are machine dependent, though).
17                  contain a 'name'?          o)  Implement devices etc. for one or more machine modes,
18          x)  Testmachine includes:              to get some guest OS running. OpenBSD/mvme88k on MVME187
19                  + dev_fb block fill and copy              seems to be the smartest path to follow for now.
20                  + dev_fb draw characters (from the built-in font)?                  o)  VME bus device
21                  + dev_fb input device? mouse pointer coordinates and buttons                  o)  PCC2
22                          (allow changes in these to cause interrupts as well?)                  o)  Cirrus Logic serial port controller
23                  + Redefine the halt() function so that it stops "sometimes          o)  Instruction disassembly, and implementation:
24                    soon", i.e. usage in demo code should be:                  o)  See http://www.panggih.staff.ugm.ac.id/download/GCC/info/gcc.i5
25                          for (;;) {                      for some strange cases of when "div" can fail (?)
26                                  halt();                  o)  Floating point stuff
27                          }                  o)  "Graphics" instructions (M88110-specific)
         x)  Continue on SPARC emulation  
                 + Enable it in the configure script as soon as it can  
                   run all the demo programs.  
         x)  Continue on Alpha emulation  (virtual memory, etc). Cleanup.  
         x)  Nicer MIPS status bits in register dumps.  
         x)  Alignment exceptions (MIPS, PPC, ARM?, ...)  
         x)  Rewrite the networking stack; make OpenBSD work better as a guest  
             OS, fix the performance problems, make Linux work with DHCP, etc.  
             Support VDE (vde.sf.net)? Allow SLIP connections, possibly PPP,  
             in addition to ethernet?  
         x)  Implement more ethernet NICs.  
         x)  IOP (I2O) device?  
   
 Documentation:  
         x)  "Install netbsd/pmax first" => only use the install kernel?  
         x)  Rewrite the section about experimental devices, after the  
             framebuffer acceleration has been implemented, and demos  
             written. (Symbolic names instead of numbers; example  
             use cases, etc. Mention demo files that use the various  
             features?)  
         x)  "a very simple linear framebuffer device (for graphics output)"  
             under "which machines does gxemul emulate" ==> better  
             description?  
         x)  Better description on how to set up a cross compiler?  
             Example for MIPS64.  
   
 Long-term design:  
         x)  Instruction combination collisions? How to avoid easily...  
         x)  Think about how to do both SHmedia and SHcompact in a reasonable  
             way!  
         o)  Actually use the settings object, better debugger stuff, etc!  
         o)  Debugger command for enabling/disabling instruction statistics  
             during runtime.   machine.statistics = on|off  
         x)  MAINBUS REDESIGN!  
         x)  PCI redesign... I need to read up on how PCI actually works :)  
         x)  Clock framework! Go through all clock devices, make sure they  
             return correct data, and run at correct speeds!  
         x)  Dyntrans with valgrind-inspired memory checker. (In memory_rw,  
             it would be reasonably simple to add; in each individual fast  
             load/store routine = a lot more work, and it would become  
             kludgy very fast.)  
         x)  Dyntrans with SMP... lots of work to be done here.  
         x)  Dyntrans with cache emulation... lots of work here as well.  
         x)  Reimplement the config file parser from scratch.  
   
 -------------------------------------------------------------------------------  
   
 Simple Valgrind-like checks?  
         o)  Mark every address with bits which tell whether or not the address  
             has been written to.  
         o)  What should happen when programs are loaded?  Text/data, bss (zero  
             filled). But stack space and heap is uninitialized.  
         o)  Uninitialized local variables:  
                 A load from a place on the stack which has not previously  
                 been stored to => warning. Increasing the stack pointer using  
                 any available means should reset the memory to uninitialized.  
         o)  If calls to malloc() and free() can be intercepted:  
                 o)  Access to a memory area after free() => warning.  
                 o)  Memory returned by malloc() is marked as not-initialized.  
                 o)  Non-passive, but good to have: Change the argument  
                     given to malloc, to return a slightly larger memory  
                     area, i.e.  margin_before + size + margin_after,  
                     and return the pointer  + margin_before.  
                     Any access to the margin_before or _after space results  
                     in warnings. (free() must be modified to free the  
                     actually allocated address.)  
   
 SMP:  
         o)  dev_mp doesn't work well with dyntrans yet  
         o)  In general, IPIs, CAS, LL/SC etc must be made to work with dyntrans  
28    
29  MIPS:  MIPS:
30          +)  Some more work on opcodes.          o)  Nicer MIPS status bits in register dumps.
31                  x) The "wait" instruction. How to implement this functionality?          o)  Floating point exception correctness.
32                          (SMP, non-MIPS, interrupt correctness, host idling, ...)          o)  Fix this? Triggered by NetBSD/sgimips? Hm:
33                    to_be_translated(): TODO: unimplemented instruction:
34                    000000000065102c: 00200800 (d)  rot_00  at,zr,0
35            o)  Some more work on opcodes.
36                  x) MIPS64 revision 2.                  x) MIPS64 revision 2.
37                          o)  Find out which actual CPUs implement the rev2 ISA!                          o)  Find out which actual CPUs implement the rev2 ISA!
38                            o)  DINS, DINSM, DINSU etc
39                            o)  DROTR32 and similar MIPS64 rev 2 instructions,
40                                which have a rotation bit which differs from
41                                previous ISAs.
42                  x) _MAYBE_ TX79 and R5900 actually differ in their                  x) _MAYBE_ TX79 and R5900 actually differ in their
43                     opcodes? Check this carefully!                     opcodes? Check this carefully!
44          o)  Dyntrans: Count register updates are probably not 100% correct yet.          o)  Dyntrans: Count register updates are probably not 100% correct yet.
45          o)  Refactor code for performance and readability/maintainability.          o)  Refactor code for performance and readability/maintainability.
         o)  DROTR32 and similar MIPS64 rev 2 instructions, which have  
             a rotation bit which differs from previous ISAs.  
         o)  EI and DI instructions for MIPS64/32 rev 2. NOTE: These are  
             _NOT_ the same as for R5900!  
46          o)  (Re)implement 128-bit loads/stores for R5900.          o)  (Re)implement 128-bit loads/stores for R5900.
47            o)  Coprocessor 1x (i.e. 3) should cause cp1 exceptions, not 3?
48                    (See http://lists.gnu.org/archive/html/qemu-devel/2007-05/msg00005.html)
49          o)  R4000 and others:          o)  R4000 and others:
50                  x)  watchhi/watchlo exceptions, and other exception                  x)  watchhi/watchlo exceptions, and other exception
51                      handling details                      handling details
52            o)  MIPS 5K* have 42 physical address bits, not 40/44?
53          o)  R10000 and others:  (R12000, R14000 ?)          o)  R10000 and others:  (R12000, R14000 ?)
54                    x)  The code before the line
55                            /*  reg[COP0_PAGEMASK] = cpu->cd.mips.coproc[0]->tlbs[0].mask & PAGEMASK_MASK;  */
56                        in cpu_mips.c is not correct for R10000 according to
57                        Lemote's Godson patches for GXemul. TODO: Go through all
58                        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
59                        and make sure everything works with R10000.
60                        Then test with OpenBSD/sgi?
61                    x)  Entry LO mask (as above).
62                  x)  memory space, exceptions, ...                  x)  memory space, exceptions, ...
63                  x)  use cop0 framemask for tlb lookups                  x)  use cop0 framemask for tlb lookups
64                      (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)                      (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)
65    
66  Dyntrans:  SuperH:
67          x)  Redesign/rethink the delay slot mechanism used for e.g. MIPS,          x)  Auto-generation of loads/stores! This should get rid of at least
68                  so that it caches a translation (that is, an instruction              the endianness check in each load/store.
69                  word and the instr_call it was translated to the last          x)  Experiment with whether or not correct ITLB emulation is
70                  time), so that it doesn't need to do slow              actually needed. (20070522: I'm turning it off today.)
71                  to_be_translated for each end of page?          x)  SH4 interrupt controller:
72          x)  Program Counter statistics:                  x)  MASKING should be possible!
73                  Per machine? What about SMP? All data to the same file?          x)  SH4 DMA (0xffa00000)
74                  A debugger command should be possible to use to enable/          x)  SH4 UBC (0xff200000)
75                  disable statistics gathering.          x)  Store queues can copy 32 bytes at a time, there's no need to
76                  Configuration file option!              copy individual 32-bit words. (Performance improvement.)
77          x)  Common fatal_abort() function, which drops into the debugger          x)  SH4 BSC (Bus State Controller)
78                  without continuing.          x)  Instruction tracing should include symbols for branch targets,
79          x)  INVALIDATION should cause translations in _all_ cpus to be              and so on, to make the output more human readable.
80              invalidated, e.g. on a write to a write-protected page          x)  SH3-specific devices: Pretty much everything!
81              (containing code)          x)  NetBSD/evbsh3, hpcsh! Linux?
82          x)  16-bit encodings? (MIPS16, ARM Thumb, SH3, ...)          x)  Floating point speed!
83          x)  Lots of other stuff: see src/cpus/README_DYNTRANS          x)  Floating point exception correctness.
84          x)  true recompilation backend? think carefully about this,          x)  NetBSD HEAD (as of April 2007) hangs during bootup, because it
85              experiment in a separate project (not in GXemul)              turns on/off interrupts in an unfortunately synchronized way
86                  o) First test would be to just implement a simple              with dyntrans. This needs to be fixed.
87                     instruction such as MIPS' addiu or lui, on AMD64          x)  Exceptions for unaligned load/stores. OpenBSD/landisk uses
88                     hosts...              this mechanism for its reboot code (machine_reset).
89          x)  Idle loop detection? (Depends on target.) Could be turned          x)  Think carefully about how to implement SH5/SH64 (for evbsh5).
90              into usleep(1) or similar on the host... except when doing  
91              e.g. SMP emulation. Then it becomes trickier.  Landisk SH4:
92            x)  When NetBSD/landisk 4.0 has been released, make sure it works
93  Transputer:              in the emulator. (Update documentation, etc.)
94          x)  Implement support for Helios binaries.  
95          x)  Stack and register contents at startup?  Dreamcast:
96          x)  Figure out how to boot an entire Helios distribution.          x)  G2 DMA
97          x)  Implement all instructions. :)          x)  LAN adapter (dev_mb8696x.c).  NetBSD root-on-nfs.
98            x)  PVR:  Lots of stuff. See dev_pvr.c.
99            x)  Better GDROM support
100            x)  Modem
101            x)  PCI bridge/bus?
102            x)  Maple bus:
103                    x)  Correct controller input
104                    x)  Mouse input
105            x)  Software emulation of BIOS calls:
106                    x)  GD-ROM emulation: Use the GDROM device.
107                    x)  Use the VGA font as a fake ROM font. (Better than
108                        nothing.)
109            x)  Make as many as possible of the KOS examples run!
110            x)  More homebrew demos/games.
111            x)  SPU: Sound emulation (ARM cpu).
112            x)  VME processor emulation? "(Sanyo LC8670 "Potato")" according to
113                Wikipedia, LC86K87 according to Comstedt's page. See
114                http://www.maushammer.com/vmu.html for a good description of
115                the differences between LC86104C and the one used in the VME.
116    
117  Alpha:  Alpha:
118          o)  Virtual memory (tlbs etc)          x)  OSF1 PALcode, Virtual memory support.
119          o)  Get {NetBSD,OpenBSD,Linux}/alpha booting. :)          x)  PALcode replacement! PAL1E etc opcodes...?
120            x)  Interrupt/exception/trap handling.
121  SPARC:          x)  Floating point exception correctness.
122          o)  Load/stores to alternate address spaces!          x)  More work on bootup memory and register contents.
123          o)  Save/restore register windows etc!          x)  More Alpha machine types, so it could work with
124                OpenBSD, FreeBSD, and Linux too?
125    
126    SPARC (both the ISA and the machines):
127            o)  Implement Adress space identifiers; load/stores etc.
128            o)  Exception/trap/interrupt handling.
129            o)  Save/restore register windows etc! Both v9 and pre-v9!
130          o)  Finish the subcc and addcc flag computation code.          o)  Finish the subcc and addcc flag computation code.
131          o)  Add more registers (floating point, control regs etc)          o)  Add more registers (floating point, control regs etc)
132          o)  Disassemly of some more instructions?          o)  Disassemly of some more instructions?
133          o)  Are sll etc 32-bit sign-extending or zero-extending?          o)  Are sll etc 32-bit sign-extending or zero-extending?
134          o)  Finish the GDB register stuff.          o)  Floating point exception correctness.
135          o)  SPARC v8, v7 etc?          o)  SPARC v8, v7 etc?
136            o)  More machine modes and devices.
 Debugger:  
         o)  How does SMP debugging work? Does it simply use "threads"?  
                 What if the guest OS (running on an emulated SMP machine)  
                 has a usertask running, with userland threads?  
         o)  Try to make the debugger more modular and, if possible, reentrant!  
         o)  Remove the emul command? (But show network info if showing  
                 machines?)  
         o)  Generalize the expression evaluator. (debugger_expr.c?)  
                 settable variables      ("show nr of instructions on average")  
                 emul[x]                 defaults to current emul  
                 machine[x]              defaults to current machine  
                 cpu[x]                  defaults to currently focused cpu  
                 registers               cpu arch dependent (#-prefix)  
                 symbols                 @-prefix  
                 numeric constants       decimal, hex, and octal ($-prefix)  
                 boolean                 yes,no, true,false  
                 operators (+ - * / % & | ^ !)  
                 parentheses for grouping subexpressions  
                 NOTE: the change from % to # for register prefix!  
                 examples:  
                                 emul[0].machine[2].cpu[0].pc  
                                 machine[test2].cpu[1].ra = main  
                                 settings.show_trace_tree = yes  
   
                 Settings:  
                         o)  Remove a setting.  
                         o)  Read/write a setting given a name. (Read as  
                             string and/or int64_t simultaneously?)  
                         o)  Warnings when exiting the emulator, if the  
                             settings have not been removed exactly in  
                             the same way as they were added? This would  
                             improve code cleanliness in the long term.  
                             (I.e. require a corresponding _destroy()  
                             function for all _new functions... machine_  
                             cpu_ etc.)  
   
                 Help command should have subsections! One for "expressions",  
                 mirrored in the documentation, but the internal help should  
                 be the one that should be considered correct.  
         o)  see src/debugger.c for more  
137    
138  POWER/PowerPC:  POWER/PowerPC:
139            x)  Fix DECR timer speed, so it matches the host.
140            x)  NetBSD/prep 3.x triggers a possible bug in the emulator:
141                <wdc_exec_command(0xd005e514,0xd60cdd30,0,8,..)>
142                  <ata_get_xfer(0,0xd60cdd30,0,8,..)>
143                    <0x26c550(&ata_xfer_pool,2,0,8,..)>
144                    <0x35c71c(0x3f27000,0,52,8,..)>
145                  <ata_exec_xfer(0xd005e4c8,0x3f27000,0,13,..)>
146                    <atastart(0xd005e4c8,0x3f27000,0,13,..)>
147                      <__wdccommand_start(0xd005e4c8,0x3f27000,0,13,..)>
148                        <bsw1(&prep_isa_io_space_tag,0x800001f6,0,176,..)>
149                    [ wdc: write to SDH: 0xb0 (sectorsize 2, lba=1, drive 1, head 0) ]
150                        <wdcwait(0xd005e4c8,72,64,0xbb8,..)>
151                          <0x198120(0xd005e4c8,72,64,0xbb8,..)>
152                            <bsr1(&prep_isa_io_space_tag,0,0,0xbb8,..)>
153                            <delay(100,0,0,0xbb8,..)>
154                Note: <bsr1(&prep_isa_io_space_tag,0,0,0xbb8,..)>
155          x)  PPC optimizations; instr combs          x)  PPC optimizations; instr combs
156          x)  64-bit stuff: either Linux on G5, or perhaps some hobbyist          x)  64-bit stuff: either Linux on G5, or perhaps some hobbyist
157                  version of AIX? (if there exists such a thing)                  version of AIX? (if there exists such a thing)
         x)  find and fix the bug which causes NetBSD/macppc to fail after  
             an install!  
158          x)  macppc: adb controller; keyboard (for framebuffer mode)          x)  macppc: adb controller; keyboard (for framebuffer mode)
159          x)  make OpenBSD/macppc work (PCI controller stuff)          x)  make OpenBSD/macppc work (PCI controller stuff)
160            x)  Floating point exception correctness.
161            x)  Alignment exceptions.
162    
163    PReP:
164            x)  Clock time! ("Bad battery blah blah")
165    
166  Algor:  Algor:
167          PCI interrupts... needed or stuff like the tlp NIC?          o)  Other models than the P5064?
168            o)  PCI interrupts... needed for stuff like the tlp NIC?
169    
170    BeBox:
171            o)  Interrupts. There seems to be a problem with WDC interrupts
172                "after a short while", although a few interrupts get through?
173            o)  Perhaps find a copy of BeOS and try it?
174    
175    HPCmips:
176            x)  Mouse/pad support! :)
177            x)  A NIC? (As a PCMCIA device?)
178    
179  ARM:  ARM:
180          o)  try to get netbsd/evbarm 3.x running (iq80321)          o)  See netwinder_reset() in NetBSD; the current "an internal error
181                occured" message after reboot/halt is too ugly.
182            o)  Generic ARM "wait"-like instruction?
183            o)  try to get netbsd/evbarm 3.x or 4.x running (iq80321)
184          o)  make the xscale counter registers (ccnt) work          o)  make the xscale counter registers (ccnt) work
185          o)  make the ata controller usable for FreeBSD!          o)  make the ata controller usable for FreeBSD!
186          o)  zaurus for openbsd...          o)  Debian/cats crashes because of unimplemented coproc stuff.
         o)  debian/cats crashes because of unimplemented coproc stuff.  
187              fix this?              fix this?
188    
189    Test machines:
190            o)  dev_fb block fill and copy
191            o)  dev_fb draw characters (from the built-in font)?
192            o)  dev_fb input device? mouse pointer coordinates and buttons
193                    (allow changes in these to cause interrupts as well?)
194            o)  Redefine the halt() function so that it stops "sometimes
195                soon", i.e. usage in demo code should be:
196                    for (;;) {
197                            halt();
198                    }
199    
200    Debugger:
201            o)  How does SMP debugging work? Does it simply use "threads"?
202                What if the guest OS (running on an emulated SMP machine)
203                has a usertask running, with userland threads?
204            o)  Try to make the debugger more modular and, if possible, reentrant!
205            o)  Remove the emul command? (But show network info if showing
206                machines?)
207            o)  Memory dumps should be able to dump both physical and
208                virtual emulated memory.
209            o)  Evaluate expressions within []? That would allow stuff like
210                cpu[x] where x is an expression.
211            o)  "pc = pc + 4" doesn't work! Bug. Should work. ("pc=pc+4" works.)
212            o)  Settings:
213                    x)  Special handlers for Write!
214                            +)  MIPS coproc regs
215                            +)  Alpha/MIPS/SPARC zero registers
216                            +)  x86 64/32/16-bit registers
217                    x)  Value formatter for resulting output.
218            o)  Call stack display (back-trace) of emulated programs.
219            o)  Nicer looking output of register dumps, floating point registers,
220                etc. Warn about weird/invalid register contents.
221            o)  Ctrl-C doesn't enter the debugger on some OSes (HP-UX?)...
222    
223    Dyntrans:
224            x)  For 32-bit emulation modes, that have emulated TLBs: tlbindex
225                arrays of mapped pages? Things to think about:
226                    x)  Only 32-bit mode! (64-bit => too much code)
227                    x)  One array for global pages, and one array _PER ASID_,
228                        for those archs that support that. On M88K, there should
229                        be one array for userspace, and one for supervisor, etc.
230                    x)  Larger-than-4K-pages must fill several bits in the array.
231                    x)  No TLB search will be necessary.
232                    x)  Total host space used, for 4 KB pages: 1 MB per table,
233                        i.e. 65 MB for 32-bit MIPS, 2 MB for M88K, if one byte
234                        is used as the tlb index.
235                    x)  (The index is actually +1, so that 0 means no hit.)
236            x)  "Merge" the cur_physpage and cur_ic_page variables/pointers to
237                one? I.e. change cur_ic_page to cur_physpage.ic_page or something.
238            x)  Instruction combination collisions? How to avoid easily...
239            x)  Think about how to do both SHmedia and SHcompact in a reasonable
240                way! (Or AMD64 long/protected/real, for that matter.)
241            x)  68K emulation; think about how to do variable instruction
242                lengths across page boundaries.
243            x)  Dyntrans with valgrind-inspired memory checker. (In memory_rw,
244                it would be reasonably simple to add; in each individual fast
245                load/store routine = a lot more work, and it would become
246                kludgy very fast.)
247            x)  Dyntrans with SMP... lots of work to be done here.
248            x)  Dyntrans with cache emulation... lots of work here as well.
249            x)  Remove the concept of base RAM completely; it would be more
250                generic to allow RAM devices to be used "anywhere".
251            o)  dev_mp doesn't work well with dyntrans yet
252            o)  In general, IPIs, CAS, LL/SC etc must be made to work with dyntrans
253            x)  Redesign/rethink the delay slot mechanism used for e.g. MIPS,
254                    so that it caches a translation (that is, an instruction
255                    word and the instr_call it was translated to the last
256                    time), so that it doesn't need to do slow
257                    to_be_translated for each end of page?
258            x)  Program Counter statistics:
259                    Per machine? What about SMP? All data to the same file?
260                    A debugger command should be possible to use to enable/
261                    disable statistics gathering.
262                    Configuration file option!
263            x)  Breakpoints:
264                    o) Physical vs virtual addresses!
265                    o) 32-bit vs 64-bit sign extension for MIPS, and others?
266            x)  INVALIDATION should cause translations in _all_ cpus to be
267                invalidated, e.g. on a write to a write-protected page
268                (containing code)
269            x)  16-bit encodings? (MIPS16, ARM Thumb, 32-bit SH on SH64)
270            x)  Lots of other stuff: see src/cpus/README_DYNTRANS
271            x)  Native code generation backends:
272                    o)  calculate at runtime whether or not chunks of emulated
273                        (physical) memory are worth translating to native code
274                        (it is assumed that it has high overhead)
275                    o)  experiment with calling the host's cc and ld externally;
276                        extremely high overhead, but could be interesting none-
277                        theless.
278                    o)  experiment with using LLVM, or GNU Lightning?
279                    o)  Important cases to think about:
280                            x)  loads/stores
281                            x)  delay branches
282                            x)  other kinds of calls, branches
283                    o)  branches to already translated code blocks can
284                        link the blocks together (block-chaining), although
285                        I'll probably want to wait with this until other
286                        things work.
287                    o)  The first tests should be done with "testm88k", because
288                        that does not affect other modes.
289    
290    -------------------------------------------------------------------------------
291    
292    Performance comparison when emulating the QEMU_MIPS machine (QEMU's default
293    MIPS machine mode):
294    
295    mips-test-0.2:
296    --------------
297    
298    1. while true; do ls -l > /dev/null; echo -n .; done, 80x36 dots
299    2. while true; do /usr/bin/md5sum /usr/bin/* > /dev/null; echo -n .; done, 80 dots
300    3. while true; do grep hej lib/libc.so.6  > /dev/null; echo -n .; done, 80 dots
301    
302                            Test 1          Test 2          Test 3
303                            ------          ------          ------
304    QEMU 0.9.0:             2 min 20 sec    45 sec          4 min 41 seconds
305    GXemul-20070608:        1 min 59 sec    3 min 18 sec    18 min 10 seconds  [A]
306    
307    
308    [A] = Normal portable dyntrans, no native code generation.
309    
310    -------------------------------------------------------------------------------
311    
312    
313    Simple Valgrind-like checks?
314            o)  Mark every address with bits which tell whether or not the address
315                has been written to.
316            o)  What should happen when programs are loaded?  Text/data, bss (zero
317                filled). But stack space and heap is uninitialized.
318            o)  Uninitialized local variables:
319                    A load from a place on the stack which has not previously
320                    been stored to => warning. Increasing the stack pointer using
321                    any available means should reset the memory to uninitialized.
322            o)  If calls to malloc() and free() can be intercepted:
323                    o)  Access to a memory area after free() => warning.
324                    o)  Memory returned by malloc() is marked as not-initialized.
325                    o)  Non-passive, but good to have: Change the argument
326                        given to malloc, to return a slightly larger memory
327                        area, i.e.  margin_before + size + margin_after,
328                        and return the pointer  + margin_before.
329                        Any access to the margin_before or _after space results
330                        in warnings. (free() must be modified to free the
331                        actually allocated address.)
332    
333    Better CD Image file support:
334            x)  Support CD formats that contain more than 1 track, e.g.
335                CDI files (?). These can then contain a mixture of e.g. sound
336                and data tracks, and booting from an ISO filesystem path
337                would boot from [by default] the first data track.
338                (This would make sense for e.g. Dreamcast CD images, or
339                possibly other live-CD formats.)
340    
341    Networking:
342            x)  Redesign of the networking subsystem, at least the NAT translation
343                    part. The current way of allowing raw ethernet frames to be
344                    transfered to/from the emulator via UDP should probably be
345                    extended to allow the frames to be transmitted other ways as
346                    well.
347            x)  Also adding support for connecting ttys (either to xterms, or to
348                    pipes/sockets etc, or even to PPP->NAT or SLIP->NAT :-).
349            x)  Documentation updates (!) are very important, making it easier to
350                    use the (already existing) network emulation features.
351            x)  Fix performance problems caused by only allowing a
352                single TCP packet to be unacked.
353            x)  Don't hardcode offsets into packets!
354            x)  Test with lower than 100 max tcp/udp connections,
355                to make sure that reuse works!
356            x)  Make OpenBSD work better as a guest OS!
357            x)  DHCP? Debian doesn't actually send DHCP packets, even
358                    though it claims to? So it is hard to test.
359            x)  Multiple networks per emulation, and let different
360                NICs in machines connect to different networks.
361            x)  Support VDE (vde.sf.net)? Easiest/cleanest (before a
362                redesign of the network framework has been done) is
363                probably to connect it using the current (udp) solution.
364            x)  Allow SLIP connections, possibly PPP, in addition to
365                ethernet?
366    
367  Cache simulation:  Cache simulation:
368          o)  Command line flags for:          o)  Command line flags for:
369                  o)  CPU endianness?                  o)  CPU endianness?
# Line 242  Cache simulation: Line 380  Cache simulation:
380              is another option (easier to implement, but very very slow).              is another option (easier to implement, but very very slow).
381    
382  Documentation:  Documentation:
383          o)  machines, cpus, devices.          x)  Update the documentation regarding the testmachine interrupts.
384          o)  Automagic documentation generation:          x)  Note about sandboxing/security:
385                    Not all emulated instructions fail in the way they would
386                    do on real hardware (e.g. a userspace program writing to
387                    a system register might work in GXemul, but it would
388                    fail on real hardware).  Sandbox = contain from the
389                    host OS. But the emulated programs will run "less
390                    securely".
391            x)  Try NetBSD/arc 4.x! (It seems to work with disk images!)
392            x)  NetBSD/pmax 4 install instructions: xterm instead of vt100!
393            x)  BETTER DEVICE EXAMPLES!
394                    o)  Move away from technical.html to somewhere new.
395                    o)  DEVICE_TICK
396                    o)  Implement example devices using interrupts, dyntrans
397                        memory access, etc.?
398            x)  Document the dyntrans core?
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            x)  Better description on how to set up a cross compiler?
408                Example for MIPS64.
409            o)  Automagic documentation generation?
410                    x)  machines, cpus, devices.
411                  x)  REMEMBER that several machines/devices can be in                  x)  REMEMBER that several machines/devices can be in
412                          the same source file!                          the same source file!
413          o)  Try to rewrite the install instructions for those machines          o)  Try to rewrite the install instructions for those machines
414              that use 3MAX into using CATS? (To remove the need to a raw              that use 3MAX into using CATS or hpcmips? (To remove the need
415              ffs partition using up all of the disk image.)              to use a raw ffs partition, using up all of the disk image.)
   
 More generic out_of_memory error reporting, and check everywhere!  
         Causes: OpenBSD has low default limits for normal users.  
                 Host is 32-bit? (32-bit hosts are limited to 4 GB or less  
                 of userspace memory.)  
                 You are actually low on RAM. (As trivial as this might sound,  
                 Unix systems usually allow processes to allocate virtual  
                 memory beyond the amount of RAM in the machine.)  
   
 Breakpoints: 32-bit vs 64-bit sign extension for MIPS, warnings, etc.  
         Use the debugger's symbolic name stuff. (which will have to be  
         extended soon to support stuff like  "2*x + symbol + y" etc. cool  
         stuff)  
416    
417  The Device subsystem:  The Device subsystem:
418          x)  allow devices to be moved and/or changed in size (down to a          x)  allow devices to be moved and/or changed in size (down to a
419              minimum size, etc, or up to a max size)              minimum size, etc, or up to a max size); if there is a collision,
420                return false. It is up to the caller to handle this situation!
421            x)  NOTE: Translations must be invalidated, both for
422                registering new devices, and for moving existing ones.
423                cpu->invalidate translation caches, for all CPUs that
424                are connected to a specific memory.
425          x)  keep track of interrupts and busses? actually, allowing any device          x)  keep track of interrupts and busses? actually, allowing any device
426              to be a bus might be a nice idea.              to be a bus might be a nice idea.
427          x)  turn interrupt controllers into devices? :-)          x)  turn interrupt controllers into devices? :-)
428          x)  refactor various clocks/nvram/cmos into one device?          x)  refactor various clocks/nvram/cmos into one device?
429    
 Clocks:  
         x)  General framework for automagic clock adjustment for _all_  
             kinds of clocks and timers. (Which should be possible to turn  
             off, of course, like the way DECstation emulation works now.)  
   
430  PCI:  PCI:
431            x)  Pretty much everything related to runtime configuration, device
432                slots, interrupts, etc must be redesigned/cleaned up. The current
433                code is very hardcoded and ugly.
434            o)  Allow cards to be added/removed during runtime more easily.
435            o)  Allow cards to be enabled/disabled (i/o ports, etc, like
436                NetBSD needs for disk controller detection).
437            o)  Allow devices to be moved in memory during runtime.
438            o)  Interrupts per PCI slot, etc. (A-D).
439            o)  PCI interrupt controller logic... very hard to get right,
440                because these differ a lot from one machine to the next.
441          x)  last write was ffffffff ==> fix this, it should be used          x)  last write was ffffffff ==> fix this, it should be used
442              together with a mask to get the correct bits. also, not ALL              together with a mask to get the correct bits. also, not ALL
443              bits are size bits! (lowest 4 vs lowest 2?)              bits are size bits! (lowest 4 vs lowest 2?)
444          x)  add support for address fixups          x)  add support for address fixups
445          x)  generalize the interrupt routing stuff (lines etc). this should          x)  generalize the interrupt routing stuff (lines etc)
             be per machine? or per bus, that's better  
         x)  add a "pcn" NIC (AMD PCnet32 Lance 79c970 (PCI 1022:2000)),  
             could be useful for several machine modes (Malta, Algor, evbarm,  
             hp700?, macppc, etc.)  
   
 Network layer:  
         o)  DHCP (for Debian and BSD installers :-)  
         o)  increase performance  
         o)  don't rely on NetBSD-ish usage  
         o)  Multiple networks per emulation, and let different  
             NICs in machines connect to different networks.  
         o)  many other issues: see src/net.c  
446    
447  Busses:  Clocks and timers:
448          o)  Redesign the entire "mainbus" concept!          x)  Fix the PowerPC DECR interrupt speed! (MacPPC and PReP speed, etc.)
449          o)  Busses should be placed in a hierarchical tree!          x)  DON'T HARDCODE 100 HZ IN cpu_mips_coproc.c!
450          o)  Easily configurable interrupt routing in SMP systems.          x)  NetWinder timeofday is incorrect! Huh? grep -R for ta_rtc_read in
451          o)  Specific clock/bus speeds, cpu speeds etc.              NetBSD sources; it doesn't seem to be initialized _AT ALL_?!
452          o)  Synchronization over network? or at least in dyntrans within          x)  Cobalt TOD is incorrect!
453              one emulated machine          x)  Go through all other machines, one by one, and fix them.
         o)  dev->bus: TurboChannel, PCMCIA, ADB?  
454    
455  Config file parser:  Config file parser:
456          o)  Rewrite it from scratch!          o)  Rewrite it from scratch!
457          o)  Usage of any expression available through the debugger          o)  Usage of any expression available through the debugger
458            o)  Allow interrupt controllers to be added! and interrupts
459                to be used in more ways than before
460          o)  Support for running debugger commands (like the -c          o)  Support for running debugger commands (like the -c
461              command line option)              command line option)
462    
463  Floating point layer:  Floating point layer:
464          o)  make it common enough to be used by _all_ emulation modes          o)  make it common enough to be used by _all_ emulation modes
465          o)  implement more stuff          o)  implement correct error/exception handling and rounding modes
466            o)  implement more helper functions (i.e. add, sub, mul...)
467          o)  non-IEEE modes (i.e. x86)?          o)  non-IEEE modes (i.e. x86)?
468    
469  Userland emulation:  Userland emulation:
470          x)  Lots of stuff; freebsd and netbsd (and linux?) syscalls.          x)  Try to prefix "/emul/mips/" or similar to all filenames,
471          x)  Dynamic linking? Hm.              and only if that fails, try the given filename.
472                Read this setting from an environment variable, and only
473                if there is none, fall back to hardcoded string.
474            x)  File descriptor (0,1,2) assumptions? Find and fix these?
475            x)  Dynamic linking!
476            x)  Lots of stuff; freebsd, netbsd, linux, ... syscalls.
477            x)  Initial register/stack contents (environment, command line args).
478            x)  Return value (from main).
479            x)  mmap emulation layer
480            x)  errno emulation layer
481            x)  struct conversions for many syscalls
482    
483  Sound:  Sound:
484          x)  generic sound framework          x)  generic sound framework
485          x)  add one or more sound cards as devices          x)  add one or more sound cards as devices; add a testmachine
486                sound card first?
487            x)  Dreamcast sound? Generic PCI sound cards?
488    
489  ASC SCSI controller:  ASC SCSI controller:
490          x)  NetBSD/arc 2.0 uses the ASC controller in a way which GXemul          x)  NetBSD/arc 2.0 uses the ASC controller in a way which GXemul
491              cannot yet handle. (NetBSD 1.6.2 works ok.) (Possibly a problem              cannot yet handle. (NetBSD 1.6.2 works ok.) (Possibly a problem
492              in NetBSD itself, http://mail-index.netbsd.org/source-changes/              in NetBSD itself, http://mail-index.netbsd.org/source-changes/
493              2005/11/06/0024.html suggests that.)              2005/11/06/0024.html suggests that.)
494                NetBSD 4.x seems to work? :)
495    
496  Caches / memory hierarchies: (this is mostly MIPS-specific)  Caches / memory hierarchies: (this is mostly MIPS-specific)
497          o)  src/memory*.c: Implement correct cache emulation for          o)  src/memory*.c: Implement correct cache emulation for
# Line 346  Caches / memory hierarchies: (this is mo Line 510  Caches / memory hierarchies: (this is mo
510              possible.              possible.
511    
512  File/disk/symbol handling:  File/disk/symbol handling:
513            o)  Make sure that disks can be added/removed during runtime!
514                (Perhaps this needs a reasonably large re-write.)
515          o)  Remove some of the complexity in file format guessing, for          o)  Remove some of the complexity in file format guessing, for
516                  Ultrix kernels that are actually disk images?                  Ultrix kernels that are actually disk images?
517          o)  Better handling of tape files          o)  Better handling of tape files      
518          o)  Read function argument count and types from binaries? (ELF?)          o)  Read function argument count and types from binaries? (ELF?)
519          o)  Better demangling of C++ names. Note: GNU's C++ differs from e.g.          o)  Better demangling of C++ names. Note: GNU's C++ differs from e.g.
520              Microsoft's C++, so multiple schemes must be possible. See              Microsoft's C++, so multiple schemes must be possible. See
# Line 357  File/disk/symbol handling: Line 523  File/disk/symbol handling:
523  Userland ABI emulation:  Userland ABI emulation:
524          o)  see src/useremul.c          o)  see src/useremul.c
525    
 Terminal/console:  
         o)  allow emulated serial ports to be connected to the outside  
             world in a more generic way, or even to other emulated  
             machines(?)  
   
 Save state of the whole emulated machine, to be able to load it back  
         in later?  (Memory, all device's states, all registers and  
         so on.  Like taking a snapshot. (SimOS seems to do this,  
         according to its website.))  
   
526  Better framebuffer and X-windows functionality:  Better framebuffer and X-windows functionality:
527            o)  Generalize the update_x1y1x2y2 stuff to an extend-region()
528                function...
529          o)  -Yx sometimes causes crashes.          o)  -Yx sometimes causes crashes.
530          o)  Simple device access to framebuffer_blockcopyfill() etc,          o)  Simple device access to framebuffer_blockcopyfill() etc,
531              and text output (using the built-in fonts), for dev_fb.              and text output (using the built-in fonts), for dev_fb.
# Line 378  Better framebuffer and X-windows functio Line 536  Better framebuffer and X-windows functio
536          o)  Non-resizable windows?  Or choose scaledown depending          o)  Non-resizable windows?  Or choose scaledown depending
537                  on size (and center the image, with a black border).                  on size (and center the image, with a black border).
538          o)  Different scaledown on different windows?          o)  Different scaledown on different windows?
539            o)  Non-integral scale-up? (E.g. 640x480 -> 1024x768)
540          o)  Switch scaledown during runtime? (Ala CTRL-ALT-plus/minus)          o)  Switch scaledown during runtime? (Ala CTRL-ALT-plus/minus)
541          o)  Bug reported by Elijah Rutschman on MacOS with weird          o)  Bug reported by Elijah Rutschman on MacOS with weird
542              keys (F5 = cursor down?).              keys (F5 = cursor down?).
# Line 391  Better framebuffer and X-windows functio Line 550  Better framebuffer and X-windows functio
550          o)  Generalize the framebuffer stuff by moving _ALL_ X11          o)  Generalize the framebuffer stuff by moving _ALL_ X11
551                  specific code to src/x11.c!                  specific code to src/x11.c!
552    
553  Statistics:  (this could be interesting)  -------------------------------------------------------------------------------
         o)  Save to file and show graphics. It should be possible to  
             run gxemul after a simulation to just show the graphics,  
             or convert to a .ppm or .tga or similar.  
         o)  memory accesses (to measure cache efficiency and  
                 page coloring efficiency)  
         o)  nr of simultaneous ASIDs in use in the TLB, for MIPS  
         o)  percentage of time spent in different "states", such as  
             running userland code, kernel code, or idling (for CPUs  
             that have such an instruction, or whenever the PC is  
             inside a specific idle-function (address range)).  
             Possible additional state (for example on R3000): caches  
             disabled.  
         o)  position of read/write on (SCSI) disks  
554    
Legend:
Removed from v.30  
changed lines
  Added in v.42
  ViewVC Help
Powered by ViewVC 1.1.26