summaryrefslogtreecommitdiff
path: root/sys/dev/pci/noct.c
AgeCommit message (Collapse)Author
2004-05-04Remove useless ``elm'' argument from the SIMPLEQ_REMOVE_HEAD macro.Alexander Yurchenko
This matches our SLIST behaviour and NetBSD's SIMPLEQ as well. ok millert krw deraadt
2004-02-03advertise features of our crypto chips better; ok tdevalTheo de Raadt
jason is being a slacker
2004-01-09remove uvm_extern.hBrad Smith
tested on alpha, i386, powerpc, sparc64, m68k. ok miod@
2003-06-02nuke clase 3 & 4Jason Wright
2003-02-18convert to new registration schemeJason Wright
2002-07-21- Deal with the fact that nsp2k's bridge can't deal with burst read/writeJason Wright
by detecting conditions that could generate one from a sequential access on the parent bridge and generating a harmless nop (scratch registers are just plain nice to have) - also, commit work in progress for dealing with modexp properly, one parameter still left to compute, tho.
2002-07-17Missing splx, and rearrange a bit.Jason Wright
2002-07-16Add support for DES and 3DES singleton operationsJason Wright
2002-07-16- Add bus_dmamap_sync() in appropriate places for EAJason Wright
- move the input workq stuff into it's own function (it's going to be subdivided further to handle non-hash operations).
2002-07-16- Add a kthread for handling EA requests so that we have a context where weJason Wright
can call bus_dmamem_{map/unmap} safely, this is because this chip has (you'll love this...) NO support for scatter-gather - implement enough of the crypto(9) api to support the simple/stateless operations: SHA1 and MD5 simple hash - remove debugging EA queue handling
2002-06-28Whoops, don't actually register for MODEXP yetJason Wright
2002-06-28Implement MODEXP as MODMUL for debugging purposes until I understand theJason Wright
args for MODEXP.
2002-06-21skeleton queue handling for the enc/auth piece of the chip (fed with nops viaJason Wright
timeout for now).
2002-06-21skeleton for queue handling on the pk side of the chip. For now it justJason Wright
enqueues nops and handles them.
2002-06-04limit number of rng's passed up per tick to 32; suggested by mickey.Jason Wright
2002-06-02Skeleton driver for the NetOctave NSP2000 (only supports RNG at the moment)Jason Wright