2 A: Access to the latest CVS development tree:
3 cvs -d :pserver:pserver@cvs.jankratochvil.net:/cvs login
4 Just hit ENTER (empty password)
5 cvs -d :pserver:pserver@cvs.jankratochvil.net:/cvs -z3 checkout -d captive priv/captive
7 cvs -d :pserver:pserver@cvs.jankratochvil.net:/cvs -z3 checkout -d reactos -r captive -kk priv/reactos
9 The final mapping of directories to CVS repository after the step above:
10 captive -> priv/captive
11 captive/macros -> macros
12 captive/reactos -> priv/reactos
14 Current diff of captive changes against reactos (it may have backlog):
15 cvs -d :pserver:pserver@cvs.jankratochvil.net:/cvs -z3 rdiff -u -R -r bp_captive -r captive priv/reactos
17 Official reactos website:
18 http://www.reactos.com/
20 Access to the original reactos source tree (mirrored to my "priv/reactos"):
21 cvs -d :pserver:cvsanon@mok.lvcm.com:/CVS/ReactOS login
22 Type the password "cvsanon" (without quotes)
23 cvs -d :pserver:cvsanon@mok.lvcm.com:/CVS/ReactOS -z3 checkout reactos
25 You may also like ViewCVS web access to the repository:
26 http://cvs.jankratochvil.net/viewcvs/priv/captive/
27 http://cvs.jankratochvil.net/viewcvs/priv/reactos/?only_with_tag=captive
28 http://cvs.jankratochvil.net/viewcvs/macros/
30 Finally you can compile the package from the "captive" directory:
34 The package gets preconfigured with --enable-maintainer-mode option
35 automatically by './autogen.pl'. You can later just update the "captive"
36 subdirectory if it is your currentdir:
41 Although './autogen.pl' should not be needed do not bugreport if it wan't
42 rerun as some dependencies may not be respected by the package.
44 Q: Why it is based on static structures, no multithreading possibility?
45 A: reactos has also static design therefore there would be no possibility
46 to use the whole libcaptive in multithreaded way.
47 Another question is whether to support multithreading in the scope of one
48 reactos machine - this would be possible but it is not yet implemented
49 in libcaptive. All such points should be marked by "TODO:thread"
51 Q: Prefer during libcaptive coding the GNU stylee or reactos (=W32) coding style?
52 A: Hard decision as captive connects these two worls together.
53 It was decided to prefer coding in GNU style in the Gnome specific way.
55 Full function replaces apparently must use the same calling conventions.
56 Coding of the full-replace should be done in GNU style if coded from scratch
57 although it is permitted to leave the original function code if only minor
58 modifications were performed (decide whether it is not better to patch
59 reactos source files instead to keep the code unified).
60 Any other code besides function replace should use GNU style calling conventions.
62 All the normal (GNU) world uses 32-bit wchar_t but reactos (=W32) uses
63 16-bit (unsigned short) wchar_t. This would need GCC -fshort-wchar option.
64 It has effect on L"literal-string" or L'c' character-constants.
65 Unfortunately they expect libc-named functions such as wcsrchr(3) to also
66 follow this broken 16-bit way. FIXME: reactos-compat-includes
68 Q: How is the libcaptive project trustworthy?
69 A: We have to run foreign native closed-source W32 code which is certainly NOT
70 trust trustworthy in any way. We could close it to some separate address
71 space and give it some safe gateways to invoce our code for the public world
72 (read "UNIX processes") interaction. It would be too much complicated
73 without any gain thus we choose to open our libcaptive-using process to any
74 attacks by W32 but the whole libcaptive process MUST be closed by a sandbox.
75 Of course all the communication with libcaptive must be taken with care by
78 Q: How to check valid input arguments? Invalid processing state assumptions?
79 A: libcaptive uses uses various Gnome macros out of <glib/gmessages.h>.
80 The common final still should run safely when all theses checks are disabled
81 as the bug-free code should never invoce them. This is the same for
82 libcaptive with a little excuse for foreign native closed-source W32 code
83 beign run which can <a href="FIXME:How is the libcaptive project trustworthy?"
84 >never be trusted</a>. As we glued all the libcaptive code
85 together with the W32 code we no longer can protect all the borders to the
86 enemy land inside the one tainted process. Therefore even fully trusted
87 libcaptive code can never safely run with <glib/gmessages.h> checks
88 disabled. YMMV if your trust W32 code. Be aware that W32 code has to also
89 handle correctly all the corrupted disk data structure to comply with such
90 trusted relationship. I really dunno if it complies but I don't care as
93 Q: Is it safe to captive_cleanup() the process and captive_init() it again?
94 A: It should be possible to fully cleanup the state of libcaptive even after
95 <a href="FIXME:How to check valid input arguments? Invalid processing state assumptions?"
96 ><glib/gmessages.h> failures</a>. It is currently not implemented and it
97 may get dropped at all in the future as libcaptive process MUST be sandboxed
98 anyway. You are recommended (currently required) to terminate the sandboxed
99 process and initialize the new one instead of captive_cleanup().
101 Q: I am missing symbol although it is already implemented by reactos.
102 A: First look at $(top_srcdir)/src/libcaptive/ke/exports.captivesym
103 where it may be enough to add such symbol to the list of exported symbols.
104 If it is not compiled at all find such symbol definition in reactos.
105 If such file or directory isn't listed in captive reactos compilation part
106 you may need to add it to $(top_srcdir)/src/libcaptive/reactos/ according
107 to the template files around - you will really only add the file to
108 Makefile.am, no real file gets copied there (just symlinked automatically).
110 Usuaully you will also like to edit the appended file and simplify it by lines
113 #endif /* LIBCAPTIVE */
114 as too much (useless for you now) code would have dependencies on other
115 missing code and you would get into never ending race of solving missing ones.
116 If the feature being implemented is some feature not needed or not suitable
117 for the emulation by libcaptive you should implement the emulation
118 below $(top_srcdir)/src/libcaptive/ directly. The filename should conform
119 to the reactos directory tree standard if applicable.
121 Always rather implement emply emulation function than patching the referring
122 code to prevent code development conflicts with the mainstream reactos.
124 Q: Why not Wine instead of ReactOS?
125 A: Although <a href="http://www.winehq.com/">Wine</a> has better development
126 activity than <a href="http://www.reactos.com/">ReactOS</a> the decision
127 is strictly technical. We need to run W32 file system driver which always
128 lives in kernel-land in W32 system design.
129 guest-OS host-OS implements core library correspondence
130 Wine GNU/Linux user ntdll
131 ReactOS i386 hardware kernel+user ntoskrnl
132 (ReactOS is going to reuse Wine for its 'user' part.)
133 FIXME: href in the table title:
134 host-OS: http://www.vmware.com/support/reference/common/glossary/#hostos
135 guest-OS: http://www.vmware.com/support/reference/common/glossary/#guestos
136 While ReactOS is nice to provide us the wanted kernel emulation we also
137 need to run the guest-OS in GNU/Linux. Initially I wanted to extend Wine
138 to the kernel emulation effort but fortunately
139 <a href="mailto:Steven_Ed4153@yahoo.com">Steven Edwards<a> pointed me to
140 the ReactOS project which better suits the needs of captive.
142 The host-OS problem unfortunately hits us on ReactOS but this isn't
143 much hard task. W32 is designed to be hardware-independent using its <code>hal.dll</code>.
144 Unfortunately ReactOS doesn't follow this design and thus we have to patch
145 and replace various parts of hardware-dependent code even inside <code>ntoskrnl.exe</code>.
147 Some functions are provided both by <code>ntdll.dll</code> and
148 <code>ntoskrnl.exe</code> in W32. Suggested by
149 <a href="mailto:chorns@users.sourceforge.net">Casper Hornstrup</a> this has
150 to be differentiated as <code>ntdll.dll</code> lives in the userland
151 (low addressspace) and <code>ntoskrnl.exe</code> in the kernelland
152 (high addressspace). Unfortunately these two guys are very far each other
153 and they would have serious problems to correctly hear themselves as the
154 addresspace differentiates at 0x80000000. For completeness it may also
155 differentiate at 0xC0000000 address instead depending on whether /3GB was
156 specified during startup. BTW reactos always uses 0xC0000000 point (macro
157 <code>KERNEL_BASE</code>) although this address doesn't apply in any way on
158 reactos running in host-OS GNU/Linux when compiled for libcaptive.
160 Q: What is Ordinal in the W32 library symbols import/export scope?
161 A: Each exported symbol has its Ordinal specified - this number specifies the offset
162 in the library symbol table. During the symbol import from a different module
163 it can specify either the full symbol name or just this number - Ordinal.
164 I dunno why it was ever created as the symbol name string size is very tiny.
165 Fortunately the current versions of W32 binaries no longer utilize this Ordinal
166 feature thus captive doesn't provide them in its export.
168 Q: Why I don't see debug info for imported binary W32 modules? Do you use W32 symbol files?
169 Q: W32 symbol files have different debugging info structure. We would have to
170 convert it and I am not aware Wine or ReactOS already done this task.
171 Not yet implemented and I currently do not plan to do it.
173 Q: Any competiting projects?
174 A: Depending on your demands you may be interested in one of:
175 <a href="http://www.powerquest.com/partitionmagic/">PowerQuest PartitionMagic</a>:
177 <a href="http://www.acronis.com/products/osselector/">Acronis OS Selector</a>:
180 Q: Do you differentiate cdecl/stdcall/fastcall functions?
182 <a href="http://msdn.microsoft.com/library/en-us/vclang/html/_core_argument_passing_and_naming_conventions.asp"
183 >three different types of function calls</a>.
184 captive is completely compiled with the GNU standard function calls "cdecl"
185 - _even_ the functions from reactos source files with different design
186 function call type (=calltype) are all compiled as GNU standard "cdecl" ones and thus
187 the whole project uses one unified calltype. There is always one 'relaying'
188 function generated inside libcaptive/ke/exports.c file named
189 functionname_calltype (such as functionname_fastcall etc.).
191 Relaying function are generated as:<dl>
192 <dt>cdecl</dt><dd>No relaying function generated, just a simple #define.
193 You won't see this function in the gdb backtrace.
194 We cannot generate it by C code as we do not know the number of its arguments.
195 We require the host OS GCC feature "__attribute__((__cdecl__))"
196 although it should be the default function call.
198 <dt>stdcall</dt><dd>Relaying stub generated, all arguments passed.
199 We require the host OS GCC feature "__attribute__((__stdcall__))".
201 <dt>fastcall</dt><dd>W32 puts first two arguments to ECX and EDX.
202 GCC puts the arguments in the order EAX, EDX, ECX, ...
203 Although GCC can change this order by "-mreg-alloc=REGS" commandline
204 option such option is not supported by stock GNU/Linux GCC compilers.
205 Therefore we pretend an unused first argument and swap the order of
206 second vs. third function argument to get the right parameter pass.
207 We require the host OS GCC feature "__attribute__((__regparm__(3)))".
208 We require the host OS GCC feature "__attribute__((__stdcall__))".
210 32-bit return value in EAX is always passed back although it may be vain.
211 Please see the documentation of "captivesym.pl" for more information.
213 Unfortutely the situation is dense during calling from the standard
214 calltype forced captive/reactos to the foreign binary W32 code as this
215 code we cannot recompile in any way. All function prototypes leading
216 to the W32 binary code are declared by CAPTIVE_CDECL or CAPTIVE_STDCALL
217 macros from "captive/calltype_reactos.h". These macros have to be
218 #ifdef-ed inside reactos code as they are not available during standalong
219 reactos compilation. CAPTIVE_FASTCALL is not provided but it should not be
220 needed as the non-core kernel code inside foreign W32 binaries usually
221 does not provide any FASTCALL functions. There may be problem if some
222 internal reactos code gets compiled while being called through
223 CAPTIVE_CDECL or CAPTIVE_STDCALL declared protypes - there is currently
226 Q: What happens if the calltype differs than expected by caller/callee?
227 A: If it is fastcall/non-fastcall type you will usually find it easily
228 as the passed arguments will differ. stdcall/cdecl types is much harder
229 as the only effect will be unexpected ESP returned which gets automatically
230 corrected during the caller function return from EBP.
232 We may do some sanity checking in libcaptive/ps/signal.c SIGSEGV handler
233 but any such solution is just a perusing of expected W32 binary code
234 stack frame handling code. It is not possible to correctly do such sanity
235 check as it would have to be coded directly into the W32 binary code.
237 Q: Which registers get saved during various function calltypes?
238 A: EDI, ESI, EBX is always saved on stack. It is stored on the stack
239 in this particular order from bottom to top addresses (therefure
240 it is stored by 'push EBX', 'push ESI', 'push EDI').
242 Fortunately GNU/Linux GCC AFAIK has the same register saving behaviour.
243 If you get some register corruption please check
244 <a href="What happens if the calltype differs than expected by caller/callee?">function calltypes</a>.
246 Q: Why do you call the platform "W32"?
247 A: http://mail.gnu.org/pipermail/libtool/2000-September/000110.html
249 Q: What captive-specific W32 registry entries do you use?
250 A: "\captive\filesystem" as the base path for the loaded W32 filesystem
252 Q: Why do you use reactos instead of native <code>ntoskrnl.exe</code>?
253 A: Although the reactos implementation is not native and therefore it will
254 always have has some incompatibilities with native W32 platform it is
255 much easier to debug anything with source-available code.
256 If there will be serious compatibility problems we may start replacing
257 reactos implementation with native <code>ntoskrnl.exe</code> code
258 but it should be considered as last resort - it is enough pain to debug
259 closed source filesystem drivers behaviour while all the other system
260 components are still opensourced.
262 Not using <code>ntoskrnl.exe</code> has no license gain as you already
263 need valid W32 licence to use native filesystem driver. Any known W32
264 license covering W32 native filesystem driver also covers <code>ntoskrnl.exe</code>.
266 Eric Kohl, reactos developer:
267 ..." and ReactOS cannot run on Linux!"
268 "ReactOS won't run on Linux because of direct hardware access!"