Commit Graph

111125 Commits

Author SHA1 Message Date
thorpej
b5154d96e4 Build a selection of install kernels. 2003-04-10 02:22:06 +00:00
christos
1af07f5803 Add MIIF_FORCEANEG from Martin Husemann 2003-04-10 01:58:21 +00:00
thorpej
24f1d87951 Descend into the gzboot_ADI_BRH_flash_0x00140000 subdir. 2003-04-10 01:50:28 +00:00
thorpej
a10453bb40 Add support for using cached mbuf and cluster phys addrs. 2003-04-09 23:32:03 +00:00
nathanw
cc1db07645 Just use PAGE_SHIFT from <machine/vmparam.h> instead of going through
hoops to read and cache the value from the crash dump header.
2003-04-09 22:46:39 +00:00
thorpej
9c04149fde If ${MKPIC} is not "no", then depend on the _pic versions of libbfd
and libopcodes.  Otherwise, depend on the regular library.
2003-04-09 22:38:24 +00:00
matt
dd1424e7ac Add POOL_VTOPHYS. Change vtophys to return -1 if pmap_extract fails.
(callers of vtophys should always supply a valid VA so that
pmap_extract should never fail).
2003-04-09 22:37:32 +00:00
matt
c72503e1bd Add some debug printf's. 2003-04-09 22:35:21 +00:00
thorpej
a63a02e3b9 If MKPIC is "no", then we need to provide *something* for BFD-using
programs to link against, so reset MKLINKLIB to "yes" in that case.
2003-04-09 22:33:26 +00:00
dsl
c4dba16f34 Add support for i386 boot code - for new bootxx stuff 2003-04-09 22:30:59 +00:00
matt
b4c7fb17b2 Cleanup dmamap_sync a bit and add a few more comments. Add support for
cached physaddr's in mbufs.
2003-04-09 22:28:56 +00:00
nathanw
e9ab31d743 POOL_VTOPHYS: Cast argument to ALPHA_K0SEG_TO_PHYS() to vaddr_t to
prevent gcc complaining about bitwise operations on pointers.
2003-04-09 22:14:31 +00:00
dsl
f14ef8a1e3 Add support for 'new' format i386 bootxx code 2003-04-09 22:14:27 +00:00
thorpej
9bd03b1f69 Cast the arg to MIPS_KSEG0_TO_PHYS() in POOL_VTOPHYS() (thanks, nathan!). 2003-04-09 22:10:58 +00:00
thorpej
03befad98b In uvm_map_clean(), only call pgo_put if the object has one.
From Quentin Garnier <quatriemek.com!netbsd>.
2003-04-09 21:39:29 +00:00
kleink
27d8588746 Change the strerror() string for ERANGE consistently to "Result too
large or too small" and reword its intro(2) description to reflect
the present use for non-numerical out-of-range errors; resolves
Andrew Brown's PR standards/21004, the discussion of which was
dramatically sped up by repeated whipping from Perry Metzger. :-)
2003-04-09 21:36:28 +00:00
jmmv
33b8c350b6 Add a Catalan libc message catalog file (plus everything else needed to
track the file).  Closes my own PR lib/19498.
2003-04-09 20:17:06 +00:00
thorpej
4115b37b40 Note pool cache paddr, mbuf, and bus_dma changes. 2003-04-09 19:31:22 +00:00
thorpej
df25ffa39a Bump to 1.6R -- mbuf and pool changes. 2003-04-09 19:28:06 +00:00
jdolecek
5b8f6703cc update rxvt entry to form as shipped with current (2.7.3) rxvt
most notably this adds \017 to me, which fixes PR misc/9361

entry retains F10-F12 + END key fixes from rev. 1.67 of termcap.src
2003-04-09 19:21:46 +00:00
thorpej
744fb366d9 Update for cached physical addresses in the pool caches. 2003-04-09 19:02:29 +00:00
jdolecek
d135e24e2d fix couple more simple lock issues on smbfs_close() code path:
* vinvalbuf needs to be called without simplelocks held
* need to release the lock in opencount > 0 case
* need to release the lock before smbfs_findclose(), since that
  can send a request to SMB server and attempt to pool_get() a request buffer
problem path found & testing by Martin Husemann, fix adresses PR kern/21067
2003-04-09 18:57:29 +00:00
thorpej
bcea7d5f28 Use cached physical addresses for mbufs and clusters to save having
to extract the physical address from the virtual.

On the ARM, also use the "read-only at MMU" indication to avoid a
redundant cache clean operation.

Other platforms should use these two as examples of how to use these
new pool/mbuf features to improve network performance.  Note this requires
a platform to provide a working POOL_VTOPHYS().

Part 3 in a series of simple patches contributed by Wasabi Systems
to improve network performance.
2003-04-09 18:51:35 +00:00
jdolecek
1f29707d43 move ntfs_remove() to be together with the other dummies - easier to c&p
eventually
2003-04-09 18:46:47 +00:00
jdolecek
6810b56243 add dummy link, rename, mkdir, rmdir vnode ops, which release the
appropriate vnodes before failing with EOPNOTSUPP - this is necessary
to not deadlock later
2003-04-09 18:41:05 +00:00
thorpej
f775de9f61 * Use a pool_cache constructor to record the physical address of mbufs
in the mbuf header.
* Use the new cached paddr feature of the pool_cache API to record
  the physical address of mbuf clusters.  (We cannot use a ctor for
  clusters, since clusters have no constructed form; they are merely
  buffers).

Bus_dma back-ends may use the cached physical addresses to save having to
extract the physical address from virtual.

* Provide space in m_ext recording the vm_page *'s for an SOSEND_LOAN_CHUNK-
  sized non-cluster external buffer.  Use this in the sosend_loan code to
  save having to extract the physical address from virtual and then look
  up the vm_page *'s.

* Provide an indication that an external buffer is mapped read-only at
  the MMU.  Set this flag for the external buffer in the sosend_loan
  case, since loaned pages are always mapped read-only.  Bus_dma back-ends
  may use this information to save cache flushing, since a cache flush of
  a read-only mapping is redundant on some architectures (the cache would
  have already been flushed when making the mapping read-only).

Part 2 in a series of simple patches contributed by Wasabi Systems
to improve network performance.
2003-04-09 18:38:01 +00:00
thorpej
a0aee79a1d Add the ability for pool caches to cache the physical address of
objects.  Clients of the pool_cache API must consistently use
the "paddr" variants or not, otherwise behavior is undefined.

Enable this on Alpha, ARM, MIPS, and x86.  Other platforms must
define POOL_VTOPHYS() in the appropriate manner in order to enable
the feature.

Part 1 of a series of simple patches contributed by Wasabi Systems
to improve network performance.
2003-04-09 18:22:13 +00:00
thorpej
fe57487140 Remove something accidentally included in last commit. 2003-04-09 18:17:34 +00:00
provos
7685de2195 use LOGIN_NAME_MAX instead of MAXLOGNAME; closes pr/21048 2003-04-09 17:50:08 +00:00
jmmv
c1b9eafbfd Add supported kernel options to the SYNOPSIS section. Somebody more qualified
should describe what they do...  Closes my own PR misc/17648.
2003-04-09 17:41:28 +00:00
jmmv
46337f6c93 Add a SYNOPSIS section with all kernel options; bump date.
Closes my own PR misc/17647.
2003-04-09 17:30:04 +00:00
jmmv
a6f4d1b666 Document the COM_HAYESP option; bump date. Closes my own PR misc/17640. 2003-04-09 17:21:48 +00:00
jmmv
737d93d096 Document the PMS_DISABLE_POWERHOOK option, and bump date.
Closes my own PR misc/17639.
2003-04-09 17:11:55 +00:00
jmmv
3634aeb3d8 ... and bump date. 2003-04-09 17:04:40 +00:00
jmmv
91f7948df4 Move description of IPSEC* related options from options(4) to ipsec(4) and
add them to the SYNOPSIS section.  Closes my own PR misc/17634.
2003-04-09 16:44:46 +00:00
thorpej
7360657293 Tweak the way the pagesize-related variables are set:
* Remove DEFAULT_PAGE_SIZE.  We don't use PAGE_SIZE the way Mach did.
* In uvm_setpagesize(), if we are called with uvmexp.pagesize == 0,
  then assert that PAGE_SIZE != 0 (i.e. a constant), and set uvmexp.pagesize
  accordingly.
* Provide defaults for MIN_PAGE_SIZE and MAX_PAGE_SIZE if not defined
  by <machine/vmparam.h>.  If PAGE_SIZE is not a constant, MIN_PAGE_SIZE
  and MAX_PAGE_SIZE must be provided.
* If MIN_PAGE_SIZE and MAX_PAGE_SIZE are not equal (i.e. PAGE_SIZE may
  not be a constant in all configurations), then ensure that PAGE_SIZE
  and friends expand to variable references for LKMs.
2003-04-09 16:34:10 +00:00
thorpej
d154229cf2 Don't define PAGE_SIZE, etc. in terms of NBPG, rather use the same
logic as used by <machine/param.h> to define constant PAGE_SIZE if
possible.

Also define the minimum and maximum PAGE_SIZE that might appear in
any given configuration (4K and 8K, respectively).
2003-04-09 16:27:06 +00:00
aymeric
a069ac021f Fix a bug where window(1) would (appear to) miss a character when two
characters are typed fast one after the other (<10 ms).
The visual effect under vi(1) was even more impressive. ;-)
This bug was actually introduced inadvertantly in 1995 in rev 1.4.
2003-04-09 16:25:10 +00:00
thorpej
efdd83f440 Move CPU configuration predicates to <machine/cpuconf.h>. 2003-04-09 16:22:33 +00:00
jdolecek
011c1952fd #if 0 some no longer used macros; keep around for refence 2003-04-09 16:18:56 +00:00
jdolecek
af5ced0d31 replace VOP__UNLOCK(), VN_LOCK(), VGET() macros with the real thing, to
imporove readability
g/c some non-NetBSD code for same reason
2003-04-09 16:18:17 +00:00
jdolecek
81d7a920df print some potentially useful stuff in ntfs_print()
replace the VOP__UNLOCK() macros with VOP_UNLOCK() directly - it just
  obfuscates the code
similarily for VN_LOCK()->vn_lock()
unlock dvp before ntvattrget call in ntfs_lookup() in '..' case, not after
fix problem in ntfs_lookup() where PDIRUNLOCK was not set in one code path
  after unlocking parent directory vnode
2003-04-09 16:12:18 +00:00
jdolecek
a0e0f99677 add real ntfs remove vnode op, which releases vnode locks on parent directory
and the file vnode before returning EOPNOTSUPP
fixes PR kern/19595 by Erik Berls
2003-04-09 16:02:18 +00:00
matt
4daebcbff5 Make the imask_t typedef a struct rather than a naked array. The attribute
on the array typedef was contaminating other u_int32_t naked arrays and
affecting their alignment.
2003-04-09 15:44:26 +00:00
yamt
65e3d14a8a rename a very confusing variable name.
(must_commit -> stalewriteverf)
2003-04-09 14:30:30 +00:00
yamt
b90af31ce6 when commit failed and fall to write, re-set 'off' and 'cnt'
because it can be changed in 'needcommit' path.
2003-04-09 14:27:58 +00:00
yamt
2cca2b5641 update a comment to follow the previous change. 2003-04-09 14:24:32 +00:00
yamt
9b96b4ab78 make per-iod datas together. 2003-04-09 14:22:33 +00:00
yamt
0dff798bde rename nm_verf to nm_writeverf because it's confusing with nm_verf{str,len}. 2003-04-09 14:21:24 +00:00
drochner
e0fb4a2b6b -in bindtextdomain(), don't crash if a NULL dirname is passed - return the
current path setting in this case (or default)
-fix return value
-don't copy a string to itself
2003-04-09 14:11:33 +00:00