Menu
▾
▴
basilisk-devel
|
From: Geoffrey B. <geo...@in...> - 2010-10-06 23:55:27
|
The following change which was checked in to the repository causes immediate segv on Linux (Linux ... machine name elided .... .6.18-194.17.1.el5 #1 SMP Mon Sep 20 07:16:16 EDT 2010 i686 i686 i386 GNU/Linux RCS file: /home/cvs/cebix/SheepShaver/src/Unix/main_unix.cpp,v retrieving revision 1.88 diff -r1.88 main_unix.cpp 846c846 < #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT --- > #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT && PAGEZERO_HACK Clearly that code was turned off previously Geoffrey |
|
From: Alexei S. <ale...@gm...> - 2010-10-07 00:10:40
|
Which change are you talking about? I don't see any recent activity around that part of main_unix.cpp... Which combination of macros does your config define that causes the problem? -Alexei On Wed, Oct 6, 2010 at 7:55 PM, Geoffrey Brown <geo...@in...> wrote: > The following change which was checked in to the repository causes > immediate segv on Linux > (Linux ... machine name elided .... .6.18-194.17.1.el5 #1 SMP Mon Sep > 20 07:16:16 EDT 2010 i686 i686 i386 GNU/Linux > > > RCS file: /home/cvs/cebix/SheepShaver/src/Unix/main_unix.cpp,v > retrieving revision 1.88 > diff -r1.88 main_unix.cpp > 846c846 > < #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT > --- >> #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT && PAGEZERO_HACK > > > Clearly that code was turned off previously > > Geoffrey > > ------------------------------------------------------------------------------ > Beautiful is writing same markup. Internet Explorer 9 supports > standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. > Spend less time writing and rewriting code and more time creating great > experiences on the web. Be a part of the beta today. > http://p.sf.net/sfu/beautyoftheweb > _______________________________________________ > basilisk-devel mailing list > bas...@li... > https://lists.sourceforge.net/lists/listinfo/basilisk-devel > |
|
From: Geoffrey B. <geo...@in...> - 2010-10-07 00:35:30
|
Ok, that's curious. The sources I was building on Linux previously had this diff diff -r1.88 main_unix.cpp 846c846 < #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT --- > #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT && PAGEZERO_HACK But as you point out there doesn't seem to have been any activity on this recently so I'm not sure how I ended up with that diff -- I went to the web based cvs to see if I could find the change and can't. In any case, without the "PAGEZERO_HACK' flag to make that #if false, things bomb immediately on linux. Adding that "back" I can make forward progress. I have no idea where that came from (a bug fix suggested on a forum?) So, maybe I need to configure for linux differently -- I've been doing ./configure --enable-sdl-audio --enable-sdl-video I'll try to figure out what is going on. Geoffrey On Wed, Oct 6, 2010 at 8:09 PM, Alexei Svitkine <ale...@gm...> wrote: > Which change are you talking about? > > I don't see any recent activity around that part of main_unix.cpp... > > Which combination of macros does your config define that causes the problem? > > -Alexei > > On Wed, Oct 6, 2010 at 7:55 PM, Geoffrey Brown <geo...@in...> wrote: >> The following change which was checked in to the repository causes >> immediate segv on Linux >> (Linux ... machine name elided .... .6.18-194.17.1.el5 #1 SMP Mon Sep >> 20 07:16:16 EDT 2010 i686 i686 i386 GNU/Linux >> >> >> RCS file: /home/cvs/cebix/SheepShaver/src/Unix/main_unix.cpp,v >> retrieving revision 1.88 >> diff -r1.88 main_unix.cpp >> 846c846 >> < #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT >> --- >>> #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT && PAGEZERO_HACK >> >> >> Clearly that code was turned off previously >> >> Geoffrey >> >> ------------------------------------------------------------------------------ >> Beautiful is writing same markup. Internet Explorer 9 supports >> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >> Spend less time writing and rewriting code and more time creating great >> experiences on the web. Be a part of the beta today. >> http://p.sf.net/sfu/beautyoftheweb >> _______________________________________________ >> basilisk-devel mailing list >> bas...@li... >> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >> > > ------------------------------------------------------------------------------ > Beautiful is writing same markup. Internet Explorer 9 supports > standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. > Spend less time writing and rewriting code and more time creating great > experiences on the web. Be a part of the beta today. > http://p.sf.net/sfu/beautyoftheweb > _______________________________________________ > basilisk-devel mailing list > bas...@li... > https://lists.sourceforge.net/lists/listinfo/basilisk-devel > |
|
From: Alexei S. <ale...@gm...> - 2010-10-07 00:43:39
|
How does it crash? -Alexei On Wed, Oct 6, 2010 at 8:35 PM, Geoffrey Brown <geo...@in...> wrote: > Ok, that's curious. The sources I was building on Linux previously > had this diff > > diff -r1.88 main_unix.cpp > 846c846 > < #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT > --- >> #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT && PAGEZERO_HACK > > But as you point out there doesn't seem to have been any activity on > this recently > so I'm not sure how I ended up with that diff -- I went to the web based cvs to > see if I could find the change and can't. > > In any case, without the "PAGEZERO_HACK' flag to make that #if false, things > bomb immediately on linux. Adding that "back" I can make forward progress. > I have no idea where that came from (a bug fix suggested on a forum?) > > So, maybe I need to configure for linux differently -- I've been doing > > ./configure --enable-sdl-audio --enable-sdl-video > > I'll try to figure out what is going on. > > Geoffrey > > > > > On Wed, Oct 6, 2010 at 8:09 PM, Alexei Svitkine > <ale...@gm...> wrote: >> Which change are you talking about? >> >> I don't see any recent activity around that part of main_unix.cpp... >> >> Which combination of macros does your config define that causes the problem? >> >> -Alexei >> >> On Wed, Oct 6, 2010 at 7:55 PM, Geoffrey Brown <geo...@in...> wrote: >>> The following change which was checked in to the repository causes >>> immediate segv on Linux >>> (Linux ... machine name elided .... .6.18-194.17.1.el5 #1 SMP Mon Sep >>> 20 07:16:16 EDT 2010 i686 i686 i386 GNU/Linux >>> >>> >>> RCS file: /home/cvs/cebix/SheepShaver/src/Unix/main_unix.cpp,v >>> retrieving revision 1.88 >>> diff -r1.88 main_unix.cpp >>> 846c846 >>> < #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT >>> --- >>>> #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT && PAGEZERO_HACK >>> >>> >>> Clearly that code was turned off previously >>> >>> Geoffrey >>> >>> ------------------------------------------------------------------------------ >>> Beautiful is writing same markup. Internet Explorer 9 supports >>> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >>> Spend less time writing and rewriting code and more time creating great >>> experiences on the web. Be a part of the beta today. >>> http://p.sf.net/sfu/beautyoftheweb >>> _______________________________________________ >>> basilisk-devel mailing list >>> bas...@li... >>> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >>> >> >> ------------------------------------------------------------------------------ >> Beautiful is writing same markup. Internet Explorer 9 supports >> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >> Spend less time writing and rewriting code and more time creating great >> experiences on the web. Be a part of the beta today. >> http://p.sf.net/sfu/beautyoftheweb >> _______________________________________________ >> basilisk-devel mailing list >> bas...@li... >> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >> > > ------------------------------------------------------------------------------ > Beautiful is writing same markup. Internet Explorer 9 supports > standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. > Spend less time writing and rewriting code and more time creating great > experiences on the web. Be a part of the beta today. > http://p.sf.net/sfu/beautyoftheweb > _______________________________________________ > basilisk-devel mailing list > bas...@li... > https://lists.sourceforge.net/lists/listinfo/basilisk-devel > |
|
From: Geoffrey B. <geo...@in...> - 2010-10-07 00:46:49
|
sigsegv right after the gtk gui. I was able (barely) to step in gdb until the crash occurred and if I recall correctly it was exactly at the point of executing that ifdefed code. I was trying to build a fresh checkout to see if my vhd stuff breaks on linux. Later, I looked to see what had "changed" and found that diff. Geoffrey On Wed, Oct 6, 2010 at 8:42 PM, Alexei Svitkine <ale...@gm...> wrote: > How does it crash? > > -Alexei > > On Wed, Oct 6, 2010 at 8:35 PM, Geoffrey Brown <geo...@in...> wrote: >> Ok, that's curious. The sources I was building on Linux previously >> had this diff >> >> diff -r1.88 main_unix.cpp >> 846c846 >> < #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT >> --- >>> #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT && PAGEZERO_HACK >> >> But as you point out there doesn't seem to have been any activity on >> this recently >> so I'm not sure how I ended up with that diff -- I went to the web based cvs to >> see if I could find the change and can't. >> >> In any case, without the "PAGEZERO_HACK' flag to make that #if false, things >> bomb immediately on linux. Adding that "back" I can make forward progress. >> I have no idea where that came from (a bug fix suggested on a forum?) >> >> So, maybe I need to configure for linux differently -- I've been doing >> >> ./configure --enable-sdl-audio --enable-sdl-video >> >> I'll try to figure out what is going on. >> >> Geoffrey >> >> >> >> >> On Wed, Oct 6, 2010 at 8:09 PM, Alexei Svitkine >> <ale...@gm...> wrote: >>> Which change are you talking about? >>> >>> I don't see any recent activity around that part of main_unix.cpp... >>> >>> Which combination of macros does your config define that causes the problem? >>> >>> -Alexei >>> >>> On Wed, Oct 6, 2010 at 7:55 PM, Geoffrey Brown <geo...@in...> wrote: >>>> The following change which was checked in to the repository causes >>>> immediate segv on Linux >>>> (Linux ... machine name elided .... .6.18-194.17.1.el5 #1 SMP Mon Sep >>>> 20 07:16:16 EDT 2010 i686 i686 i386 GNU/Linux >>>> >>>> >>>> RCS file: /home/cvs/cebix/SheepShaver/src/Unix/main_unix.cpp,v >>>> retrieving revision 1.88 >>>> diff -r1.88 main_unix.cpp >>>> 846c846 >>>> < #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT >>>> --- >>>>> #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT && PAGEZERO_HACK >>>> >>>> >>>> Clearly that code was turned off previously >>>> >>>> Geoffrey >>>> >>>> ------------------------------------------------------------------------------ >>>> Beautiful is writing same markup. Internet Explorer 9 supports >>>> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >>>> Spend less time writing and rewriting code and more time creating great >>>> experiences on the web. Be a part of the beta today. >>>> http://p.sf.net/sfu/beautyoftheweb >>>> _______________________________________________ >>>> basilisk-devel mailing list >>>> bas...@li... >>>> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >>>> >>> >>> ------------------------------------------------------------------------------ >>> Beautiful is writing same markup. Internet Explorer 9 supports >>> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >>> Spend less time writing and rewriting code and more time creating great >>> experiences on the web. Be a part of the beta today. >>> http://p.sf.net/sfu/beautyoftheweb >>> _______________________________________________ >>> basilisk-devel mailing list >>> bas...@li... >>> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >>> >> >> ------------------------------------------------------------------------------ >> Beautiful is writing same markup. Internet Explorer 9 supports >> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >> Spend less time writing and rewriting code and more time creating great >> experiences on the web. Be a part of the beta today. >> http://p.sf.net/sfu/beautyoftheweb >> _______________________________________________ >> basilisk-devel mailing list >> bas...@li... >> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >> > > ------------------------------------------------------------------------------ > Beautiful is writing same markup. Internet Explorer 9 supports > standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. > Spend less time writing and rewriting code and more time creating great > experiences on the web. Be a part of the beta today. > http://p.sf.net/sfu/beautyoftheweb > _______________________________________________ > basilisk-devel mailing list > bas...@li... > https://lists.sourceforge.net/lists/listinfo/basilisk-devel > |
|
From: Michael S. <msb...@me...> - 2010-10-07 01:14:24
|
FYI... If I recall correctly, what is happening in this section of the code is this: In REAL_ADDRESSING mode, the addresses in the VM (Mac OS) are the same as in the host system (e.g. Linux or Mac OS). The problem is Mac OS wants a low memory global area, which means it needs an allocation at address zero in the host address space. First it tries to allocate the entire RAM/ROM area starting at address zero in the host, assuming that it can arrange for the SheepShaver module to be loaded at some other address (HAVE_LINKER_SCRIPT). If that succeeds, we're in good shape. Otherwise, on Darwin (OS X) hosts it can use a trick to get around the problem (PAGEZERO_HACK). The auto-config is supposed to test if that works. The hack means it doesn't have to do a low memory allocation. If it can't do PAGEZERO_HACK then it tries to allocate just the small low memory area it needs, at zero. On some Linux systems, you must do sudo sysctl -w vm.mmap_min_addr=0 to allow it to allocate at address zero. But if you don't allow it, you should get an error, not a segv. And on non-OS X hosts, I can't figure out how it would work to not allocate the address zero block. On Oct 6, 2010, at 7:46 PM, Geoffrey Brown wrote: > sigsegv right after the gtk gui. I was able (barely) to step in gdb until > the crash occurred and if I recall correctly it was exactly at the point > of executing that ifdefed code. I was trying to build a fresh > checkout to see if my vhd stuff breaks on linux. Later, I > looked to see what had "changed" and found that diff. > > Geoffrey > > On Wed, Oct 6, 2010 at 8:42 PM, Alexei Svitkine > <ale...@gm...> wrote: >> How does it crash? >> >> -Alexei >> >> On Wed, Oct 6, 2010 at 8:35 PM, Geoffrey Brown <geo...@in...> wrote: >>> Ok, that's curious. The sources I was building on Linux previously >>> had this diff >>> >>> diff -r1.88 main_unix.cpp >>> 846c846 >>> < #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT >>> --- >>>> #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT && PAGEZERO_HACK >>> >>> But as you point out there doesn't seem to have been any activity on >>> this recently >>> so I'm not sure how I ended up with that diff -- I went to the web based cvs to >>> see if I could find the change and can't. >>> >>> In any case, without the "PAGEZERO_HACK' flag to make that #if false, things >>> bomb immediately on linux. Adding that "back" I can make forward progress. >>> I have no idea where that came from (a bug fix suggested on a forum?) >>> >>> So, maybe I need to configure for linux differently -- I've been doing >>> >>> ./configure --enable-sdl-audio --enable-sdl-video >>> >>> I'll try to figure out what is going on. >>> >>> Geoffrey >>> >>> >>> >>> >>> On Wed, Oct 6, 2010 at 8:09 PM, Alexei Svitkine >>> <ale...@gm...> wrote: >>>> Which change are you talking about? >>>> >>>> I don't see any recent activity around that part of main_unix.cpp... >>>> >>>> Which combination of macros does your config define that causes the problem? >>>> >>>> -Alexei >>>> >>>> On Wed, Oct 6, 2010 at 7:55 PM, Geoffrey Brown <geo...@in...> wrote: >>>>> The following change which was checked in to the repository causes >>>>> immediate segv on Linux >>>>> (Linux ... machine name elided .... .6.18-194.17.1.el5 #1 SMP Mon Sep >>>>> 20 07:16:16 EDT 2010 i686 i686 i386 GNU/Linux >>>>> >>>>> >>>>> RCS file: /home/cvs/cebix/SheepShaver/src/Unix/main_unix.cpp,v >>>>> retrieving revision 1.88 >>>>> diff -r1.88 main_unix.cpp >>>>> 846c846 >>>>> < #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT >>>>> --- >>>>>> #if REAL_ADDRESSING && HAVE_LINKER_SCRIPT && PAGEZERO_HACK >>>>> >>>>> >>>>> Clearly that code was turned off previously >>>>> >>>>> Geoffrey >>>>> >>>>> ------------------------------------------------------------------------------ >>>>> Beautiful is writing same markup. Internet Explorer 9 supports >>>>> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >>>>> Spend less time writing and rewriting code and more time creating great >>>>> experiences on the web. Be a part of the beta today. >>>>> http://p.sf.net/sfu/beautyoftheweb >>>>> _______________________________________________ >>>>> basilisk-devel mailing list >>>>> bas...@li... >>>>> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >>>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Beautiful is writing same markup. Internet Explorer 9 supports >>>> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >>>> Spend less time writing and rewriting code and more time creating great >>>> experiences on the web. Be a part of the beta today. >>>> http://p.sf.net/sfu/beautyoftheweb >>>> _______________________________________________ >>>> basilisk-devel mailing list >>>> bas...@li... >>>> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >>>> >>> >>> ------------------------------------------------------------------------------ >>> Beautiful is writing same markup. Internet Explorer 9 supports >>> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >>> Spend less time writing and rewriting code and more time creating great >>> experiences on the web. Be a part of the beta today. >>> http://p.sf.net/sfu/beautyoftheweb >>> _______________________________________________ >>> basilisk-devel mailing list >>> bas...@li... >>> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >>> >> >> ------------------------------------------------------------------------------ >> Beautiful is writing same markup. Internet Explorer 9 supports >> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >> Spend less time writing and rewriting code and more time creating great >> experiences on the web. Be a part of the beta today. >> http://p.sf.net/sfu/beautyoftheweb >> _______________________________________________ >> basilisk-devel mailing list >> bas...@li... >> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >> > > ------------------------------------------------------------------------------ > Beautiful is writing same markup. Internet Explorer 9 supports > standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. > Spend less time writing and rewriting code and more time creating great > experiences on the web. Be a part of the beta today. > http://p.sf.net/sfu/beautyoftheweb > _______________________________________________ > basilisk-devel mailing list > bas...@li... > https://lists.sourceforge.net/lists/listinfo/basilisk-devel |
|
From: Geoffrey B. <geo...@in...> - 2010-10-07 01:31:20
|
I rebuilt with a fresh checkout of basilisk and sheepshaver. Here's how I configured ./configure --without-gtk --with-sdl-video This is pretty tough to debug, but the crash occurs at vm_alloc.cpp line 294. if (mmap((caddr_t)addr, size, VM_PAGE_DEFAULT, the_map_flags, fd, 0) =\ = (void *)MAP_FAILED) causes a seg fault addr = 0 size = 0x4000000 the_map_flags = 0 Clearly the odd diff I had prevented this being executed. Geoffrey |
|
From: Alexei S. <ale...@gm...> - 2010-10-07 01:45:01
|
If its crashing in mmap() itself, that really sounds like a bug in your OS. It should either succeed or fail and return an error code. -Alexei On Wed, Oct 6, 2010 at 9:30 PM, Geoffrey Brown <geo...@in...> wrote: > I rebuilt with a fresh checkout of basilisk and sheepshaver. Here's > how I configured > > ./configure --without-gtk --with-sdl-video > > > This is pretty tough to debug, but the crash occurs > at vm_alloc.cpp line 294. > > if (mmap((caddr_t)addr, size, VM_PAGE_DEFAULT, the_map_flags, fd, 0) =\ > = (void *)MAP_FAILED) > > > causes a seg fault > > addr = 0 > size = 0x4000000 > the_map_flags = 0 > > > Clearly the odd diff I had prevented this being executed. > Geoffrey > > ------------------------------------------------------------------------------ > Beautiful is writing same markup. Internet Explorer 9 supports > standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. > Spend less time writing and rewriting code and more time creating great > experiences on the web. Be a part of the beta today. > http://p.sf.net/sfu/beautyoftheweb > _______________________________________________ > basilisk-devel mailing list > bas...@li... > https://lists.sourceforge.net/lists/listinfo/basilisk-devel > |
|
From: Geoffrey B. <geo...@in...> - 2010-10-07 11:43:19
|
I agree, however, this pretty vanilla red hat 5 which is why it's surprising. I wonder if there's a 32 bit/ 64 bit thing going on here. Geoffrey On Wed, Oct 6, 2010 at 9:44 PM, Alexei Svitkine <ale...@gm...> wrote: > If its crashing in mmap() itself, that really sounds like a bug in your OS. > > It should either succeed or fail and return an error code. > > -Alexei > > On Wed, Oct 6, 2010 at 9:30 PM, Geoffrey Brown <geo...@in...> wrote: >> I rebuilt with a fresh checkout of basilisk and sheepshaver. Here's >> how I configured >> >> ./configure --without-gtk --with-sdl-video >> >> >> This is pretty tough to debug, but the crash occurs >> at vm_alloc.cpp line 294. >> >> if (mmap((caddr_t)addr, size, VM_PAGE_DEFAULT, the_map_flags, fd, 0) =\ >> = (void *)MAP_FAILED) >> >> >> causes a seg fault >> >> addr = 0 >> size = 0x4000000 >> the_map_flags = 0 >> >> >> Clearly the odd diff I had prevented this being executed. >> Geoffrey >> >> ------------------------------------------------------------------------------ >> Beautiful is writing same markup. Internet Explorer 9 supports >> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >> Spend less time writing and rewriting code and more time creating great >> experiences on the web. Be a part of the beta today. >> http://p.sf.net/sfu/beautyoftheweb >> _______________________________________________ >> basilisk-devel mailing list >> bas...@li... >> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >> > > ------------------------------------------------------------------------------ > Beautiful is writing same markup. Internet Explorer 9 supports > standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. > Spend less time writing and rewriting code and more time creating great > experiences on the web. Be a part of the beta today. > http://p.sf.net/sfu/beautyoftheweb > _______________________________________________ > basilisk-devel mailing list > bas...@li... > https://lists.sourceforge.net/lists/listinfo/basilisk-devel > |
|
From: Geoffrey B. <geo...@in...> - 2010-10-07 12:15:54
|
Here's a bit more info from strace: [pid 29510] mmap2(NULL, 67108864, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0 [pid 29510] --- SIGSEGV (Segmentation fault) @ 0 (0) --- [pid 29510] --- SIGSEGV (Segmentation fault) @ 0 (0) --- On Thu, Oct 7, 2010 at 7:43 AM, Geoffrey Brown <geo...@in...> wrote: > I agree, however, this pretty vanilla red hat 5 which is why it's > surprising. I wonder if there's > a 32 bit/ 64 bit thing going on here. > > Geoffrey > > On Wed, Oct 6, 2010 at 9:44 PM, Alexei Svitkine > <ale...@gm...> wrote: >> If its crashing in mmap() itself, that really sounds like a bug in your OS. >> >> It should either succeed or fail and return an error code. >> >> -Alexei >> >> On Wed, Oct 6, 2010 at 9:30 PM, Geoffrey Brown <geo...@in...> wrote: >>> I rebuilt with a fresh checkout of basilisk and sheepshaver. Here's >>> how I configured >>> >>> ./configure --without-gtk --with-sdl-video >>> >>> >>> This is pretty tough to debug, but the crash occurs >>> at vm_alloc.cpp line 294. >>> >>> if (mmap((caddr_t)addr, size, VM_PAGE_DEFAULT, the_map_flags, fd, 0) =\ >>> = (void *)MAP_FAILED) >>> >>> >>> causes a seg fault >>> >>> addr = 0 >>> size = 0x4000000 >>> the_map_flags = 0 >>> >>> >>> Clearly the odd diff I had prevented this being executed. >>> Geoffrey >>> >>> ------------------------------------------------------------------------------ >>> Beautiful is writing same markup. Internet Explorer 9 supports >>> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >>> Spend less time writing and rewriting code and more time creating great >>> experiences on the web. Be a part of the beta today. >>> http://p.sf.net/sfu/beautyoftheweb >>> _______________________________________________ >>> basilisk-devel mailing list >>> bas...@li... >>> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >>> >> >> ------------------------------------------------------------------------------ >> Beautiful is writing same markup. Internet Explorer 9 supports >> standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. >> Spend less time writing and rewriting code and more time creating great >> experiences on the web. Be a part of the beta today. >> http://p.sf.net/sfu/beautyoftheweb >> _______________________________________________ >> basilisk-devel mailing list >> bas...@li... >> https://lists.sourceforge.net/lists/listinfo/basilisk-devel >> > |
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X