Hello Jerome,
> A while back, a user posted a bug report > https://sourceforge.net/p/freedos/bugs/235/ > that claims two separate programs (Dillo and MPlayer) > destroyed their FAT32 partition during usage. > > Does anyone have any information on if this is a bug? Or, do you > think it was a problem specifically related to that single user and > possible caused by hardware running them from USB? Or, something else > entirely? > > Thanks for any feedback you can provide in that bug report. I can only give an educated guess here, but: Dillo and Mplayer both are apps which use DOS extenders and protected mode. I do not expect Mplayer to write to disk, but Dillo probably writes to a cache on disk? It would be good to know whether disk writes are necessary to cause the problem. Would using a RAMDISK as the cache or TEMP directory for Dillo avoid the issue? It could be that USB BIOS disk writes somehow get messed up when DOS apps running in protected mode access the disk, or to be more exact, ask the DOS kernel to access the disk. This can depend on many factors, such as which drivers are loaded in which order. According to Jack, loading UHDD before EMM386 (JEMMEX, JEMM386) can help to reduce DMA issues by letting UHDD do the DMA instead of a possibly less stable BIOS, for example. Of course USB is a different story, but I can easily imagine that whether you do or do not load EMM386 at all will make a difference for the problem described in bug ticket 235. In few cases, a BIOS will not allow USB writes at all after booting from USB, but because the bug specifically mentions Dillo and Mplayer, my best guess would be "this is related to protected mode apps trying to access USB disks". Eric :-) _______________________________________________ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel