summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2001-05-13PAGE_MASK is (PAGE_SIZE - 1), *not* (PAGE_SHIFT - 1)Todd C. Miller
2001-05-13Initial support for Broadcom 5820, which is very much like the 5805 exceptJason Wright
that the packet context structure for ipsec has changed (added two fields and, annoyingly, rearranged several of them). The MCR2 operations (only RNG is used at this point) are supported, too.
2001-05-13PAGE_SIZE is a constant now. I should know that, I'm guilty asMiod Vallat
far as this port is concerned.
2001-05-13Honor the ddb.console sysctl.Miod Vallat
2001-05-12Do not use curproc before it gets initialized.Miod Vallat
From niklas@.
2001-05-12A number of buggy BIOS implementations leave the routerMichael Shalayeff
entry as 000:00:0, which is typically not the correct device/function. If the router device address is set to this value, and the compatible router entry is undefined (zero is the correct value to indicate undefined), then we work on the basis it is most likely an error, and search the entire device-space of bus 0 (but obviously starting with 000:00:0, in case that really is the right one). from Dave Sainty <dave@dtsp.co.nz>
2001-05-12always print the icu we've foundMichael Shalayeff
2001-05-12Less verbose; angelos@ okAaron Campbell
2001-05-12Move bzero() after test for correct allocation (jj@wabbitt.org)Angelos D. Keromytis
2001-05-12syncTheo de Raadt
2001-05-12more cyrix devsTheo de Raadt
2001-05-11Kill remaining vm_offset_t and vm_size_t; NetBSD used as a guideTodd C. Miller
2001-05-11syncNathan Binkert
2001-05-11New National Semiconductor gigabit phyNathan Binkert
2001-05-11Check m_pullup() and m_pullup2() return for NULL, not 0; itojun@ okAaron Campbell
2001-05-11I think this will match and work on a hifn 7951Theo de Raadt
2001-05-11syncTheo de Raadt
2001-05-11some moreTheo de Raadt
2001-05-11MAP_COPY -> MAP_PRIVATEArtur Grabowski
2001-05-11Bye bye.Artur Grabowski
2001-05-11Use genassym.cfArtur Grabowski
2001-05-11Add a genassym.cf that uses the standard scripts.Artur Grabowski
We loose the dependencies and all strange bugs that appear because genassym.c is a user program with _KERNEL defined.
2001-05-11more rd merge (headcolds suck)Theo de Raadt
2001-05-11Don't ask C for more syntactic sugar than it can provide.Miod Vallat
2001-05-11Revert previous untested change, until a fix is ready.Miod Vallat
2001-05-11unify rd(4) even moreTheo de Raadt
2001-05-11oh, i forgot to put this in ;-)Theo de Raadt
2001-05-11unifyTheo de Raadt
2001-05-11kmemstats, nselcoll, forkstat, and nchstats structures throughAngelos D. Keromytis
sysctl. deraadt@ ok
2001-05-11sysctl entries for kmemstats, nselcoll, nchstats, and forkstat.Angelos D. Keromytis
2001-05-11check for 'fastroute' condition (like ip_input) in bridge_filter() andJason Wright
deal with it.
2001-05-10Simplify probe routine.Aaron Campbell
2001-05-10Provide PAGE_{MASK,SHIFT,SIZE} constants.Miod Vallat
2001-05-10Don't play with MCR_IEN, it breaks serial console input on the 425e.Todd C. Miller
If it turns out that this causes problems on other 4xx models this can be added back for the non-425e models.
2001-05-10Nuke the DCA check. ACPI can only be console on a 425e; thorpej@netbsd.orgTodd C. Miller
2001-05-10Add missing #include <sys/resourcevar.h> so these can compile.Todd C. Miller
2001-05-10In apciparam() in the ospeed == 0 case, actually set the cfcr to the newTodd C. Miller
value; bad@netbsd.org
2001-05-10Skip "serial 1" on non-425e models. It's mapped to DCA at 9 on everyTodd C. Miller
other 4xx model, and the "not configured" could be pretty annoying. This means we can nuke apcicheckdca() and just key off being a 425e. Also, ia->ia_addr is a physical address, not a kernel virtual address; thorpej@netbsd.org
2001-05-10dmamem_unmap the memory before dmamem_freeig it.Michael Shalayeff
only affects unaligned transfers (such as disklabel, fsck and dump). we used to get away w/ this due to older pmap, uvm or both. be more precise w/ double-buffer on command failure as well. once at it, allow more than 1 segment for double-buffer to be allocated. also fix a couple of endianesses and command timeouts.
2001-05-10fix typo introduced in non-UVM case yesterdayTodd C. Miller
2001-05-10More sync to NetBSD.Artur Grabowski
The highlight is some more advices to madvise(2). o MADV_DONTNEED will deactive the pages in the given range giving a quicker reuse. o MADV_FREE will garbage-collect the pages and swap resources causing the next fault to either page in new pages from backing store (mapped vnode) or allocate new zero-fill pages (anonymous mapping).
2001-05-10Hasn't been used for ages. ZAP!Artur Grabowski
2001-05-10remove the pt{1,23}_{alloc,free} functions now that pool works withoutArtur Grabowski
UVM and UVM is required anyway.
2001-05-10Don't crash in pgt_page_alloc if uvm_km_kmemalloc fails.Artur Grabowski
(we'll crash somewhere else anyway, but that's a later problem).
2001-05-10UVM is no longer optional on sparc.Artur Grabowski
2001-05-10Switch sparc to PMAP_NEW. I've been running with that code for ages.Artur Grabowski
2001-05-10oops.Artur Grabowski
2001-05-10Some locking protocol fixes and better enforcement of wiring limits.Artur Grabowski
From NetBSD.
2001-05-10When acpi is used as the console, set cn_tab appropriately so that aTodd C. Miller
serial console on the 425e gets attached properly. The hp300 port should really be changed such that drivers don't touch cn_tab themselves but this turns out to be a bit tricky so I'm going to punt on that for now.
2001-05-10Move the timeout_set() call up a bitTodd C. Miller
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-27 02:07:39 +0000