Log into edg.uchicago.edu (an e-shop Linux machine)
export CVSROOT=:ext:edg.uchicago.edu:/designs/cdf_l2/hdl_repo
export CVS_RSH=ssh
cvs checkout hitfinder (or maybe just cvs update if you already have hitfinder checked out)
Log into an e-shop Windows PC. Start Max+Plus II V10.1. (Probably any other recent version will do. I've never seen a Hit Finder problem that could be traced to using the wrong version of Max+Plus.) The project you want should be on drive Z, which is your e-shop UNIX home directory. So the project is typically one of 'Z:\hitfinder\boot\boot', 'Z:\hitfinder\dad\dad', 'Z:\hitfinder\hitman\hitman', 'Z:\hitfinder\merge\merge', 'Z:\hitfinder\mop\mop'. (There is also 'xtrpemu', but I hope that by the time you are reading this, it is obsolete and no longer needed for CDF.)
Be sure to save your changes with cvs commit so that you have a good record of what you changed from day to day (combining CVS with a logbook is a very good way to understand what happened when you are confused a few months from now)
I recommend not archiving Altera compiler output in CVS, except for the programming files needed to load the boards. Typically I archive only these kinds of files: .tdf (source files), .acf (pin assignments, compiler settings, etc.), .vec (simulation test vectors), .mif (power-up memory images), .fram (flash RAM images of compiled firmware for dad, hitman, and mop chips, which can be loaded in situ in the SVT VME crates), and the .pof files (compiled firmware) for the boot and merge chips (which must be reloaded by plugging a programming cable into each board's front panel).
To create flash RAM memory images for dad, hitman, and mop chips, to back to edg, after compiling the code on the PC, and
cd hitfinder/tools
cc ttfconv.c -o ttfconv
cd ../dad
gzip -9f dad.ttf
../tools/ttfconv dad.ttf.gz dad.fram
cd ../hitman
gzip -9f hitman.ttf
../tools/ttfconv hitman.ttf.gz hitman.fram
cd ../mop
gzip -9f mop.ttf
../tools/ttfconv mop.ttf.gz mop.fram
Now you can move these files to dad.fram, hit.fram (hit, not hitman, oops), and mop.fram in directory b0dap30:/cdf/code-common/cdfonline/svt_config/hf and then update all Hit Finders' firmware by executing the command svtvme_hf_fill_fram(-1) on each of the SVT crates that contains Hit Finders. You can replace '-1' with a slot number to do just a single Hit Finder. '-1' does slots 4,5,6,13,14,15. Note that if you do something dumb with the .fram files, e.g. mixing up the 'hit' and 'mop' programs, then it is possible that you will fry all $200K worth of Hit Finder boards. So I'd recommend trying this on one board before rewriting the whole set. (If we had thought of this, we probably would have selected the 'mop' and 'hitman' chips from different families to avoid this problem. This can also happen, of course, if you do something dumb like erasing the pin assignments for a chip.) The program reload won't take effect until you "reboot" each Hit Finder, either by cycling the power or by writing a '3' to the "brain" register (VME address '0') on each board, e.g. with ~cdf_svt/bin/vme b0svt00 write 13 0 3 for slot 13 of crate 0.
To download the 'boot' and 'merge' chips, you need the .pof files, an Altera programming cable, and the Altera Stand-Alone Programmer (or the full Max+Plus software). The 'merge' chip is a single-chip JTAG chain. The 'boot' chip (10K50VBC356-2) and its EPC2 make a two-chip JTAG chain; *************??????************ comes first. Leave the 10K50V's programming file, and load the .pof into the EPC2. You need to "initialize device from JTAG chain" to get the EPC2 to reload the 'boot' chip itself, or else cycle the power.
Last updated on $Date: 2003/04/29 01:03:48 $ (UTC)