forums.ps2dev.org

nothing special. i had ported one of my projects 'xynth windowing system' to pspsdk. xynth uses sockets for IPC. on PSP that was sockets emulated via pipes + pipe emulated via memory. now it is just sockets emulated via pipe. thats all ;) there is information, sources and binaries of older releases ...

hi, me again ;) - bad file descriptor handling is missing in select(). if we know that, system has no idea about 'fd' which was given in one of the sets 'read/write/except' we should return '-1' and set errno to 'EBADF'. we might increse the count and set exceptfd for that 'fd', but it is safe to ob...

hi, this patch is for _times() function for now ;) i had to generate makefile.in so patch is little big, i can not paste it in here. you can grab it from; # wget http://gsulinux.org/~distch/pspdev/newlib-1.15.0-PSP-syscalls.patch there is no need for pipe.c.orig file anymore, removed it from...

thanks. pipe() emul works fine ;)

hi,
just open pspsdk/doc/html/index.html with a browser. and take a look at http://ps2dev.org/psp/Tutorials for nice pdfs.

anhanguera.

hi, thanks. it was a while i have last played with pspsdk - almost one year- it is superb that pspdev grow so much. thank you all for providing us such good dev kit. you can grab the patch from; # wget http://gsulinux.org/~distch/newlib-1.15.0-PSP.patch anhanguera. ps. i had added comment to...

i would like to pute _times(..) to a seperated file syscalls.c but there is something with my autoconf, version mismatch i think. when i generate .in and configure files with autotools ./configure gives strange errors in newlib.

thanks ;)

hi, - the sceKernel[Send/TrySend/Receive]MsgPipe(...) functions return SCE_KERNEL_ERROR_ILLEGAL_SIZE if the given number of bytes param (size_t len) is greater than the initial size of MsgPipe. in our case the initial size is 'PIPE_BUF' which is declared as 512 bytes in syslimits.h - this is the nor...

hi,

you should set PSPDEV and PATH before running toolchain.sh , if you did not. from toolchain.txt;
anhanguera.

hi, - in __psp_pipe_read (...) it is looking for how many bytes already in the pipe with __psp_pipe_peekmsgsize(...). and if there is no data already in the pipe, returns '-1'. this is correct for 'non-blocking read', for 'blocking read' it should wait until some data cames in to the pipe. - in __ps...

hi, thank you for such a nice tool. it runs perfect on FW.150 using usbhostfs. here is the small patch; Index: usbhostfs/main.c =================================================================== --- usbhostfs/main.c (revision 2234) +++ usbhostfs/main.c (working copy&...

hi, _times() function should be added to a seperate file (syscalls.c ?!). anyway here is a quick and ugly workaround.... Index: newlib-psp/newlib/libc/sys/psp/pipe.c =================================================================== --- newlib-psp/newlib/libc/sys/psp/pipe.c (revisio...

hi, raf, this is really great job. this pipe() support for newlib is very important for me (guess for other developers, too). i was using this (http://svn.sourceforge.net/viewcvs.cgi/*checkout*/xynth/trunk/src/lib/pipe_mem.c) ugly && bugy pipe emulation over memory, and i was tired of workin...

hi, download the psp binary from: http://prdownloads.sourceforge.net/xynth/xynth-0.8.00-psp-1.5.tar.bz2?download or, http://prdownloads.sourceforge.net/xynth/xynth-0.8.00-psp-1.0.tar.bz2?download 1.0 and 1.5 is the psp bios version number. download according to your bios version. below is the same f...

hi,

mkdir tmp
mv psp-1.5.tar.bz2 ./tmp
cd tmp
tar -jxvf psp-1.5.tar.bz2
cd dist
cp -Rf * /psp_root_directory

just that simple... ;)

ps: www.xynth.org seems to be down, http://gsulinux.org/~distch/projects/xynth is the exact addr. and thanks for the replies.

alper "anhanguera" akcan.

hi, I was working on a embedded windowing system (xynth), and ported to pspsdk. give a try... what is xynth; xynth is a portable embedded windowing system, it has nearly all features that you might expect from. anyway here is; web site: www.xynth.org screen shots: http://gsulinux.org/~distch/project...

hi, # export PATH=/usr/local/pspdev/bin:$PATH before patch; # psp-config -p Error, invalid suffix on the end of the path. Should be /bin/psp-config # /usr/local/pspdev/bin/psp-config -p /usr/local/pspdev/psp/sdk after patch; # psp-config -p /usr/local/pspdev/psp/sdk and here is the patch ;) ...

hi again, sceKernelStdin() returns always "0", even after close(0). anyway, here is the new patch; removed if (fd == ...) fd = sceKernelStdin(); ... stuff, and changed initialization. anhanguera. diff -U 3 -H -d -r -N newlib-psp/newlib/libc/sys/psp/libcglue.c newlib-1.13.0/newlib/libc/sys/...

hi, S_IFCHR and isatty() thing is OK with the patch(below). but assumes that user will not open a file called " __PSP_STDIO" there is; if (fd == STDIN_FILENO) { sce_fd = sceKernelStdin(); } else if (fd == STDOUT_FILENO) &...

hi, i confuses me, too ;) but i tried this senorio below; sceIoOpen() -> fd = 3; sceIoOpen() -> fd = 4; sceIoClose(4) sceIoOpen() -> fd = 4 sceIoClose(4) sceIoClose(3) sceIoOpen() -> fd = 3; sceIoOpen() -> fd = 4; as seen here, if we do not open more than 1021 files fd will always smaller then 1024....

hi, t would only affect open(), fstat(), and close(). If we just have a fixed array of pointers, the overhead actually wouldn't be too bad (strdup for open, free for close, table lookup for fstat). Locking shouldn't be necessary. If we support 0<fd<1024, the mem overhead is only 4kb+filenames. this ...

hi, if there is no sceIoFstat or, any other function like char * sceIoFname (SceUID fd) the only(?) choice is to use fd->filename map. and this will slow down(?) all the io; lock map, check/add/del io, unlock map. there is something like int sceIoIoctl (fd, cmd, indata, inlen, outdata, outlen); any ...

hi, open(path, ...) fails if path has "//" ("path//file"), and here is the patch. --- libcglue.org.c Wed Oct 5 13&#58;24&#58;55 2005 +++ libcglue.c Wed Oct 5 14&#58;03&#58;27 2005 @@ -118,7 +118,14 @@ int _open&#40;const char *name, int flags, int mode&#41...

hi, on timeout sceKernelWaitSema() returns SCE_KERNEL_ERROR_WAIT_TIMEOUT, not -1. and (int) SCE_KERNEL_ERROR_WAIT_TIMEOUT is < 0. anyway here is the patch. --- SDL_syssem.org.c Wed Oct 5 10&#58;25&#58;52 2005 +++ SDL_syssem.c Wed Oct 5 10&#58;39&#58;29 2005 @@ -85,7 +85,7 @@ int SDL_...

To answer the original question. The best way to debug the driver is using Naplink and a pl2301 cable or using the Serial IO as provided by mrbrown,etc. I wouldn't bother trying to debug with scr_printf and burning disks, unless you want to spend more dollars on wasted disks than a pl2301 cable or ...

I am very new to ps2 development, and I do need to debug an .irx file. I do not have a prolific cable or smap eth adaptor. I have a klsi eth adaptor, and need to debug it. but I could not manage to print anything.. I cannot manage to use scr_printf inside an irx (iop compiled) .irx file. and printf ...

oh thanks, it is better than doing all by hand..

alper.

thanks pixel, i ll look..

alper.

hi again, here is my learning steps... and a question... first, i checked the iso file given with http://www.thethirdcreation.net/tools/ps2link.zip by using isodump tool. isodump tool says; [3;1H 34 [ 1] 16 2048 02/*. [4;1H 34 [ 1] 16 2048 02/*.. [5;1H 46 [ 1] 19 8661 00/ PS2IPS.IRX;1 [6;1H 46 [ 1] ...