andrew@eleceng.bradford.ac.uk (Andrew G. Minter) (09/28/90)
I'm using Bruce Evan's 32-bit Minix and booting with shoelace. After upgrading my RAM to 4Mb, I tried increasing the size of the cache to 1Mb, but now shoelace locks up whebn trying to load the fs. Does anyone out there have any idea what might be wrong? Cheers, Andrew -- ############################################################################## # Andrew G. Minter, # JANET: andrew@bradford # # Lecturer in Information Systems, # EARN/BITNET: andrew@bradford.ac.uk # # Dept. of Electrical Engineering, # # # University of Bradford, # Phone: +44.274.733466 ext 347 or 326 # # Bradford, West Yorkshire, BD7 1DP # Telex: 51309 UNIBFD G # # United Kingdom # Fax: +44.274.391521 <-- just changed # ############################################################################## # Soon moving to STC Technology Ltd, London Road, Harlow, Essex # ##############################################################################
brucee@runxtsa.runx.oz.au (Bruce Evans) (09/30/90)
In article <31783@nigel.ee.udel.edu> andrew@eleceng.bradford.ac.uk (Andrew G. Minter) writes: >After upgrading my RAM to 4Mb, I tried increasing the size of the cache to >1Mb, but now shoelace locks up whebn trying to load the fs. The cache has to fit below 640K because the system is loaded into the real-mode address space :(. Shoelace should detect that the big FS would overwrite shoelace and abort the boot. However, the check for this has a bug. It fails for small overwrites. I'm not sure what happens for large ones. -- Bruce Evans (evans@syd.dit.csiro.au)
hays@isc.intel.com (Kirk Hays) (10/06/90)
In article <2324@runxtsa.runx.oz.au> brucee@runxtsa.runx.oz.au (Bruce Evans) writes: |In article <31783@nigel.ee.udel.edu> andrew@eleceng.bradford.ac.uk (Andrew G. Minter) writes: |>After upgrading my RAM to 4Mb, I tried increasing the size of the cache to |>1Mb, but now shoelace locks up whebn trying to load the fs. | |The cache has to fit below 640K because the system is loaded into the |real-mode address space :(. Shoelace should detect that the big FS would |overwrite shoelace and abort the boot. However, the check for this has a |bug. It fails for small overwrites. I'm not sure what happens for large |ones. |-- |Bruce Evans (evans@syd.dit.csiro.au) Sounds like time for a two stage boot: 1. boot a "small" 386 minix using shoelace 2. run a protected mode shoelace that runs in the low 640K of memory, loading the "big" 386 minix into high RAM. -- Kirk Hays - NRA Life. Dulce et Decorum est, pro patra mori. "The way of the samurai is found in death. It is this simple. If you can accept it then you will fight as though you are already dead." Tsunemoto Yamamoto, _Hagakure_
MDOELL%DOSUNI1.BITNET@cunyvm.cuny.edu (Magnus Doell) (10/08/90)
Couldn't we avoid the cache to be build into the image? As we have vm with the 386, why don't use it to fill the memory used by FS' cache during initialization? If we make the initialization a little bit trickier, we could avoid to build in nearly all the bss by just copying the data regions a little bit. By the way, I had read about a project with a 2-stage FS-cache a while ago (I guess they wanted the second cache dynamically allocated outside FS' address space). What's about this? All this makes the image a little bit smaller, the loading a little bit faster and the initialization a little bit trickier and slower (but you can copy a lot of memory before reading one sector :-). Ma.D.