Re: [Freedos-devel] Kernel 2035a HD format problem (Arkady's unstable version)
BIOS setting is CHS (translation), FDISK shows FAT32 (not FAT32 L). So, FAT32 LBA works but we have a problem with FAT32 CHS. Unfrtunately I can't test this here. The hard disk parameter is auto-detected by BIOS, I've no way to change it (16xx cylinder). I've tried: 1) FreeDOS FDISK (delete and create primary partition) 2) FreeDOS Format 0.91r (FORMAT D: /U) I did a few more test and found that only 2035a-UNSTABLE will report BAD SECTOR, switch back to 2035 can format successfully ... But 2035 will crash M$-DOS's EMM386 (SB Emulation 2.0 still need it). Well... no kernel is perfect. Perhaps the problem is some of the Arkady's optimisations in DSK.C?! --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 ___ Freedos-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/freedos-devel
[Freedos-devel] Re: Idea about Virtual PC compatibility / A20 handling
Hi, I had some problems with receiving mail, but reading the online list archives gives me the idea that there is a misunderstanding here... - Tyler found a small tool, and the readme of that tool states that it is important for him (the author of the tool) to keep the tool 0.5k size. - I checked for optimality of the binary, and found that it starts with mov ah,30 / int 21, followed directly by code to release environment from RAM (so the returned DOS version is not used). - So I asked the author of the tool why he had not optimzed away the int 21.30 call, and after a while, he remembered that he once had a program which would crash if it started with using XMS services right away. It did that on a Virtual PC system. His solution was to make the program start with int 21.30. - We had a problem with Virtual PC in emm386/himem recently, and I remembered that there was code to disable A20 between exec and the first int 21 call by the exec-ed program. Conclusion: I wrote to the list, suggesting that this might be an A20 related problem, and suggesting to re-add Tom's invention, the refuse to actually disable A20, to HIMEM (or EMM386: at least MS EMM386 locks the real A20 to on and uses the paged memory system to SIMULATE a (fast) switchable A20 for DOS), but not (as for Tom's version) as a compile time option but as a command line switch. And of course I would like some comments whether my speculations about why use int 21.30 before XMS to avoid hangs?. Maybe that particular HIMEM or EMM386 crashed in XMS handling on that particular virtual hardware if the A20 was not enabled explicitly before calling the XMS handler? And of course I am interested in the state of LOADFIX for FreeDOS. I think it once was part of FreeCOM, but 0.82pl3 does not have it compiled in? There is no separate LOADFIX listed in the LSM database either...!? Eric --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 ___ Freedos-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/freedos-devel
Re: [Freedos-devel] Re: Idea about Virtual PC compatibility / A20 handling
At 08:02 PM 8/23/2004 +0200, Eric Auer wrote: Conclusion: I wrote to the list, suggesting that this might be an A20 related problem, and suggesting to re-add Tom's invention, the refuse to actually disable A20, to HIMEM (or EMM386: at least MS EMM386 locks the real A20 to on and uses the paged memory system to SIMULATE a (fast) switchable A20 for DOS), but not (as for Tom's version) as a compile time option but as a command line switch. The only new options and features I will add to EMM386/HIMEM at this point in the FreeDOS release schedule are those which are know to correct an error or fix a compatibility problem. Known means exactly that. It does not mean: theorized, guessed, posited, hypothecated, assumed, suggested, or run up a flagpole to see what salutes. The last two features to EMM386 were SB to give it SoundBlaster compatibility -- which DOES work with SoundBlast emulation by the way -- and ALTBOOT to fix an incompatibility with BreadBox/GEOS's Ensemble. If, and only if, a problem occurs with an application that requires a new feature as verified by me will I personally add such a feature. Naturally, you may convince Tom or someone else to add the feature in the absence of such verification, just not me. Verification needs to be either by my testing here or via a reasonably detailed explanation by a reliable source as to why the feature is desired or necessary. Alternatively, if you have incontrovertible proof that a new EMM386 option is vital to prevent space aliens from irradiating the earth with their king-sized jello ray and converting all life forms to gelatinous cubes, I'd like to see that. I still won't add the option, but looking at the proof should be a hoot. And finally, should one of us send me a private e-mail explaining why my position is incorrect, how I am Not Doing The Right Thing, and why FreeDOS suffers mightily under my personal development restriction, I feel certain that I shall develop a mysterious case of hysterical blindness and fail to read its content. Fortunately, I am sure the condition will clear up soon thereafter. --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 ___ Freedos-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/freedos-devel