summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2002-07-27echo before running sshd; mcbride@countersiege.comTheo de Raadt
2002-07-27Fix two problems.Kenneth R Westerback
1) Even if /sbin/mount_nfs is present, do *not* try to automatically mount nfs file systems. Because not all name resolution mechanisms are present during upgrade (e.g. yp) it is not safe to do so. The user must manually mount nfs file systems if they are the source of any sets. Problem noted by miod@, change seconded by deraadt@. 2) Change initial verbiage so that a brief introduction is displayed, along with the information about !foo or !, before the terminal type is requested. After the terminal type is known display the rest of the introductory text using less -XE. This belts and suspender fix makes both the install and upgrade text fit in 'page' size (<24 lines) chunks and ensures that if the text ever grows beyond a page a more informative prompt is displayed. Problem noted by millert@.
2002-07-27syncTheo de Raadt
2002-07-26request non-executable BSS from uvm (first page of BSS may land insideTheo de Raadt
last page of DATA, and still have executability unless we can find a resolution to the GOT/PLT issues in a.out)
2002-07-26ensure that bss object being used is not ... inside the last page of data.Theo de Raadt
a.out ZMAGIC permits that to happen. so we need to pad it over, for the test to work.
2002-07-26give it some healthy dosage of vitamin constMichael Shalayeff
2002-07-26put to the -Wall an unused variableMichael Shalayeff
2002-07-26build gensysname with HOSTCC.Artur Grabowski
2002-07-26Recognize additional gem pci id, from NetBSD.Dale Rahn
2002-07-26syncDale Rahn
2002-07-26The elusive original author of this program and man page surfaced long enoughPaul Janzen
to state the following, when asked about licensing issues: "It's ok to put any copyright as long as it stays in the public domain. ... Since I did not write it to make any profit or fame in the first place, I guess its use is not limited by any license. Of course I would not object to my name being mentioned." Then he disappeared again. So I've put on a current copyright to him, and labelled the code explicitly as public domain.
2002-07-26More Apple pci devs, from NetBSD.Dale Rahn
2002-07-26add classic mode, closer to the way tetris felt a couple decades ago; ↵Michael Shalayeff
pjanzen@ ok
2002-07-26comma splicePaul Janzen
2002-07-26\a rather than embedded ^GPaul Janzen
2002-07-26Theo pointed out a vsnprintf() whose return value was being usedPaul Janzen
inappropriately.
2002-07-26A few better buffer sizes (from Ian McWilliam, long ago); functions localPaul Janzen
to save.c are no longer in rogue.h; use write() to save files rather than fwrite() since this can end up being called in a signal handler (it's still not clean though).
2002-07-26indentPaul Janzen
2002-07-26We're in curses, so use getch() rather than getchar().Paul Janzen
2002-07-26Zap VM_KMEM_SIZE, not used anymore.Artur Grabowski
Pointed out by francisco <frisco@blackant.net>
2002-07-26bwtwo tested and works, minus one rasops fix coming soon.Miod Vallat
2002-07-26fifo needs 6 entries before we feed it rectangle commandJason Wright
2002-07-26- add hardware support for eraserows and erasecolsJason Wright
- also fix some register names and add yet more of them [one thing this chip isn't lacking is registers...]
2002-07-26- remove duplicate defnsJason Wright
- start at some utility functions for the FBC handling
2002-07-26- Move register defns to their own header file and add a whole bunch of themJason Wright
- map the fbc register set
2002-07-26House cleaning (move reg defns into their own file), and be sure to drainJason Wright
the CG6 before starting operations.
2002-07-26swap log and quick a few timesHenning Brauer
2002-07-26few more $(SHELL); "Aaron Yourk" <ayourk@srt.com>Henning Brauer
2002-07-26make the order of log and quick irrelevant. now bothHenning Brauer
block in log quick all and block in quick log all work. ok dhartmei@, deraadt@
2002-07-26add support for hardware accelerated fill/copy; from NetBSD (committed from ↵Jason Wright
the acceptably fast non-X11 console of a U1)
2002-07-26and hppaTheo de Raadt
2002-07-26regress these only on the architectures where they are expected to workTheo de Raadt
2002-07-26check for non-exec of bssTheo de Raadt
2002-07-26Obey LDSTATIC for the helper binaries.Artur Grabowski
2002-07-25- some KNF.Federico G. Schwindt
- comment out debugging printf for i810. - remove checks for failure if malloc is used with M_WAITOK. mickey@ ok.
2002-07-25love, jasonMichael Shalayeff
2002-07-25missing arg in the call to the pdc_call since some old timesMichael Shalayeff
2002-07-25prettyTheo de Raadt
2002-07-25Avoid time wrap at securelevel 2. ok millert@ fgsch@Thomas Nordin
2002-07-25more snprintf and strlcpy; help from millertTheo de Raadt
2002-07-25strip off annoying local gcc symbols (jason was complaining hangman being ↵Michael Shalayeff
too easy ;)
2002-07-25kill more strcpyTheo de Raadt
2002-07-25cleanupTheo de Raadt
2002-07-25support 486 which lack the pentium_mhz stuff; from rees, kjc okTheo de Raadt
2002-07-25use bus_space_map, not bus_space_map2, and fix failure printf'sJason Wright
2002-07-25Due to changes having already occured, and more being cooked at thisMiod Vallat
very moment, OpenBSD/sparc installation media will not allow upgrades to 3.2.
2002-07-25There's no need to override the default mapping to big endian now that ↵Jason Wright
rasops8 handles byte swapping itself.
2002-07-25If passed RI_BSWAP, swap32 the stampJason Wright
2002-07-25Adapt the framebuffer code to the WSDISPLAY_TYPE constant changes, andMiod Vallat
add the necessary magic for wsemul_sun color handling. Initially based on a patch from jason@, then some tinkering by me.
2002-07-25Rework the color usage in wsemul_sun displays, to match SunOS and SolarisMiod Vallat
text mode colors. This means black text on white background, even for the kernel messages. Initially based on a patch from jason@, then some tinkering by me.