charles@hpcvca.CV.HP.COM (Charles Brown) (10/05/89)
Some friends of mine use an Amiga 2000HD in their work. (They are in video.) They have had some problems with the (Rodime 40MB) hard disk. The disk was giving lots of read errors. They brought it back to the dealer that had configured it. It was returned the software had been completely reconfigured. Most of the changes seemed unimportant. But there is one line in the s:startup-sequence which I don't understand. runback c:patch2090a My friends asked me to find out what this command does. They are worried that the dealer has used an unclean solution to the disk problems. I brought path2090a back to my Amiga 2000 and tried it. It reported Patch2090 V1.0 By: Khalid Aldoreri Unable to find Devices. Please reread Docs files. Naturally I don't have the docs files. Can you (anybody) tell me what patch2090 does? Where does it come from? Is it safe? -- Charles Brown charles@cv.hp.com or charles%hpcvca@hplabs.hp.com or hplabs!hpcvca!charles or "Hey you!" Not representing my employer. "The guy sure looks like plant food to me." Little Shop of Horrors
lphillips@lpami.wimsey.bc.ca (Larry Phillips) (10/05/89)
In <5660062@hpcvca.CV.HP.COM>, charles@hpcvca.CV.HP.COM (Charles Brown) writes: > Patch2090 V1.0 By: Khalid Aldoreri > Unable to find Devices. Please reread Docs files. >Naturally I don't have the docs files. Can you (anybody) tell me what >patch2090 does? Where does it come from? Is it safe? I don't have the docs handy, but here's the scoop... The 2090(a) suffers from problems of overrunning its FIFO during periods of intense DMA (hi-res, 4 bitplanes, etc.). When this happens, the disk operation needs to be retried. This causes a slowdown, and if there are too nmany retries, a read/write error. Unfortunately the entire operation is retried, meaning that 5 retries of a 100K transfer could try to transfer 500K. Patch 2090 attempts to alleviate this problem by watching for heavy DMA by checking the screen parameters. If appropriate, it changes the driver's MaxTransfer value to something lower (somewhere between 512 bytes and 4K I think). This means that the driver has less data to retry, and more chance of getting the request satisfied. It is, in a sense, an 'unclean solution', though at this time, it is the only solution short of putting a low MaxTransfer value into your mountlist and suffering a slowdown all the time, whether it is needed or not. Khalid did go 'outside the rules' to implement the method. Since the MaxTransfer value that is actually consulted by the file system is in the driver's data area, he could not access it directly. To find it, he uses a special MaxTransfer value in the mountlist, and then searches for that value, so that he knows where to change it. In spite of his bending of the rules, it is pretty slick, performing as advertised. It may or may not break under any revision of the file system, driver software, or OS rev. I think the dealer in question has done your friends (and Khalid) a disservice by supplying Patch2090 without docs. Khalid has written a substantial number of programs, all freely distributable, for the Amiga, and when they bend the rules, he says so. You might want to hit the dealer up for the docs. Meanwhile, I will see if I can dig up the docs for you. -larry -- The Mac? Oh, that's just like a computer, only slower. +-----------------------------------------------------------------------+ | // Larry Phillips | | \X/ lphillips@lpami.wimsey.bc.ca -or- uunet!van-bc!lpami!lphillips | | COMPUSERVE: 76703,4322 -or- 76703.4322@compuserve.com | +-----------------------------------------------------------------------+
) (10/06/89)
In article <5660062@hpcvca.CV.HP.COM>, charles@hpcvca.CV.HP.COM (Charles Brown) writes... >there is one line in the s:startup-sequence which I don't understand. > runback c:patch2090a >Naturally I don't have the docs files. Can you (anybody) tell me what >patch2090 does? Where does it come from? Is it safe? This pgm appeared in 1 of the Amiga magazines. It limits the DMA xfers to a smaller size (defined in mountlist) when it detects that a 4-color hi-res screen is being displayed - this normally caused the DMA device to crash i think (no more bus for the DMA xfers). It req'd some device info to start up - i don't have my docs handy...(in AmigoTimes #5 maybe) Mark Kaye | | | Box 172, Munster Hamlet | 613-838-3580 | kaye@fscore.dec.com | Ontario, Canada K0A 3P0 | | DEC fscore::kaye or kaye @kao |
stan@teroach.UUCP (Stan Fisher) (10/06/89)
In article <5205@shlump.nac.dec.com> kaye@FSCORE.dec.com (Mark Kaye - He who dies with the most toys - is dead!) writes: >In article <5660062@hpcvca.CV.HP.COM>, charles@hpcvca.CV.HP.COM (Charles Brown) writes... >>there is one line in the s:startup-sequence which I don't understand. >> runback c:patch2090a >>Naturally I don't have the docs files. Can you (anybody) tell me what >>patch2090 does? Where does it come from? Is it safe? > >This pgm appeared in 1 of the Amiga magazines. It limits the DMA xfers >to a smaller size (defined in mountlist) when it detects that a 4-color >hi-res screen is being displayed - this normally caused the DMA device to >crash i think (no more bus for the DMA xfers). It req'd some device >info to start up - i don't have my docs handy...(in AmigoTimes #5 maybe) >Mark Kaye | | | >Box 172, Munster Hamlet | 613-838-3580 | kaye@fscore.dec.com | >Ontario, Canada K0A 3P0 | | DEC fscore::kaye or kaye @kao I've had this program for awhile and decided to try it out last night after being reminded of it by this thread of discussion. I followed the docs to the tee, but.... she no worky for me! If I runback the program, I get no indication from patch2090 that it's done anything. No errors NOR activation info to the screen. I go into DpaintIII with hi-res 16 color, and loads and saves still take FOREVER. Has anyone actually had this work? Did Dpaint load and save (alot) quicker than without it? If I just run patch2090 then it complains that it couldn't find any devices and that I should re-read the docs. Could it be that I have SetCPU running with the fastrom option and the 2090a has the 'CardROM' option going? Here's the hardware: Rev 4.5 2000 2090A 2620 ProRam SuperAgnus etc. I added the MaxTransfer line to both my FFS partitions and have the startup-sequence runback patch2090 right after the second partition has been mounted, no difference in operation though. :^( Stan Fisher - stan@teroach.phx.mcd.mot.com - asuvax!mcdphx!teroach!stan Motorola Microcomputer Division, Tempe, Arizona - Voice (602) 438-3228 Call our User Group BBS "M.E.C.C.A." running Atredes 1.1 @ (602) 893-0804
bjc@pollux.UUCP (Betty J. Clay) (10/12/89)
In article <5660062@hpcvca.CV.HP.COM> charles@hpcvca.CV.HP.COM (Charles Brown) writes: >Some friends of mine use an Amiga 2000HD in their work. (They are in >video.) They have had some problems with the (Rodime 40MB) hard disk. >The disk was giving lots of read errors. They brought it back to the (some deleted) >My friends asked me to find out what this command does. They are > Patch2090 V1.0 By: Khalid Aldoreri > Unable to find Devices. Please reread Docs files. >Naturally I don't have the docs files. Can you (anybody) tell me what >patch2090 does? Where does it come from? Is it safe? >-- You can trust Khalid Aldoseri's work, folks. Khalid is an outstanding programmer who lives in Bahrain, in the Persian Gulf. He writes many very fine utilities for the Amiga, and he posts them on CompuServe, where he is one of the most regular participants. This patch, if I remember correctly, was to correct some contention when using oversize screens or overscan. Watch for Khalid's SCRIPIT program. It has been released now, and provides a scripting language (complete with AREXX port) that can do many of the things AREXX can do, but without requiring an AREXX port in the program being controlled. It is quite a program. And Khalid is quite a programmer. I might add that programming is Khalid's hobby. He is involved in several business ventures over there, such as selling electronics and the like. Betty ================================================================== Betty Clay ......texbell\!pollux\!bjc CompuServe 76702,337 ====================================================================