204349 Commits

Author SHA1 Message Date
jmcneill
f95d0b07fa don't use kmem from softint context 2011-08-23 15:56:12 +00:00
reinoud
6613ea22ef Implement pmap_create() 2011-08-23 15:35:53 +00:00
reinoud
6ef1d57a0c Fix zero page cleaning. The supplied `pa' is a physical address and no page
number so don't multiply with PAGE_SIZE
2011-08-23 15:12:31 +00:00
jmcneill
36013536d8 host and userkernel timespec might differ in size (because of time_t) so
instead of thunk_clock_getres() filling in a timespec, use instead
thunk_clock_getres_monotonic() that returns the resolution as a long
2011-08-23 14:37:50 +00:00
reinoud
e7b3bf46ad Clean up chatty pmap.c and fix important bug that claimed too little kmem
space. Thanks go to jmcneill for finding it!
2011-08-23 14:23:08 +00:00
reinoud
a0ace6b461 OOps... there is no aprintf_debug() 2011-08-23 13:37:32 +00:00
reinoud
98f35557ac Braindead implementation of pmap_zero_page() 2011-08-23 13:35:57 +00:00
christos
28ab15e044 don't use lwp_setprivate in fork, but copy the private lwp member directly
because userland might have messed with the TLS register without letting
the kernel know. This fixes fork() on amd64. Thanks chuq!
2011-08-23 13:01:25 +00:00
pgoyette
81d8330c12 Remove left-over debug line 2011-08-23 12:55:50 +00:00
pgoyette
99deda018d Module build glue for if_rum and if_axe 2011-08-23 12:54:05 +00:00
pgoyette
75c5b19872 Update for modular build 2011-08-23 12:53:29 +00:00
reinoud
f19d57b776 The pv_tables and the kernel pmap structure are not managed by UVM and thus
shouldn't be returned in pmap_virtual_space() for UVM.
2011-08-23 12:46:58 +00:00
jmcneill
e5faa0dc10 get rid of uvmhist stuff 2011-08-23 12:36:20 +00:00
pgoyette
3016e04310 In preparation for building this as a module, clean-up some gcc
-Wsign-compare issues so it will build with WARNS=4
2011-08-23 12:33:50 +00:00
jmcneill
616594aca0 pull in GENERIC.local if it exists 2011-08-23 12:27:01 +00:00
reinoud
1fba2c8574 Initial implementation of pmap_remove()
Also cleaned up some debug output to be aprint_debug()'d.
2011-08-23 12:06:14 +00:00
reinoud
9ae5eb8d83 Implement pv_alloc() and pv_free() and add various printf's to see what
functions are called that might have side effects due to not being
implemented!
2011-08-23 11:36:11 +00:00
christos
0404783632 document another non-literal format string 2011-08-23 10:47:06 +00:00
reinoud
9b209ce7da Cleanup pv_get() before enabling multiple-mappings 2011-08-23 10:41:47 +00:00
christos
b182f6a794 document non-literal formats 2011-08-23 10:26:07 +00:00
christos
69a4e2ee5b PR/45269: Andreas Gustafsson: Instead of falling off the edge when eating trailing newlines
if the block has moved, arrange so that trailing newlines are never placed in the string
in the first place, by accumulating them and adding them only after we've encountered a
non-newline character. This allows also for more efficient appending since we know how much
we need beforehand. From FreeBSD.
2011-08-23 10:04:39 +00:00
christos
d452d7e758 - add pid to the trace file so that we don't keep overwriting ourselves
- use va_copy to print the trace arguments so that we don't deplete it for the real printf
2011-08-23 10:01:32 +00:00
christos
819193d44d add more gcc printf format attributes 2011-08-23 09:59:20 +00:00
oki
a69c375f9d make compile with options RT3050. 2011-08-23 08:10:08 +00:00
hannken
fc2e6c60c4 When consuming only part of a path in rump_vop_lookup():
- Make sure to consume complete path components.
- Consume trailing slashes too.
- Do not clear REQUIREDIR.

Test rump/modautoload/t_modautoload now passes.
2011-08-23 07:40:32 +00:00
hannken
f937bd97b7 Stop abusing relookup() to prepare the creation of new nodes
in the upper layer.
Replace union_relookup() with union_do_lookup() that prepares
a component, calls VOP_LOOKUP() and does the EEXIST test.
2011-08-23 07:39:37 +00:00
christos
875018bf8a add -Wformat=2 2011-08-23 05:22:25 +00:00
oki
32a22fa000 make compile without VMSWAP. no functional change. 2011-08-23 03:00:35 +00:00
jmcneill
dc155b3bd5 pmap_enter: fix pmap_do_enter parameters 2011-08-23 00:52:33 +00:00
enami
75f2eef8f6 Fix the name of syscall which takes timespec as argument. 2011-08-22 22:14:46 +00:00
enami
39dd5517da Remove return statement which can't be reached. 2011-08-22 22:12:34 +00:00
enami
9d6dd8515a When both nanoseconds fields of futimens/utimensat call are set
to UTIMES_NOW, act as if NULL is passed to second argument, i.e.,
do same permission check and set exactly same value to both access
and modification time.
2011-08-22 22:09:07 +00:00
reinoud
9e34284f74 Dirty patch fixing the assert triggered on i386. On this arch the sbrk(0) HAS
been moved upward and will trigger the assert. When this situation is
encountered the kmem_data_end will be moved upward equaly.
2011-08-22 21:59:09 +00:00
jmcneill
19f8eb48c8 build fixes for netbsd-5 and i386 2011-08-22 21:45:38 +00:00
jym
cec3d05bda Arch-specific rc.conf files are not really autonomous rc.conf
files, they are appended to the end of etc/defaults/rc.conf.

So rename them to rc.conf.append for clarity, as suggested by mrg@. Adapt
Makefile accordingly.
2011-08-22 20:48:38 +00:00
jym
f56903492e Modify etc/defaults/Makefile so that architectures can specify an additional
rc.conf file. This one should reside under etc/etc.${MACHINE}/, and will
get automatically appended to etc/defaults/rc.conf at build time if present.

This is used by i386 and amd64 to append a small MD rc.conf(5) configuration
at the end of the defaults/rc.conf file, so that powerd(8) can be started
by default when we are running in a Xen environment. This is needed to support
save/restore functions for domains.

From all the alternatives proposed to fix that issue (from /etc/rc.conf
parsing in postinstall to etc/defaults/rc.conf arch-hooks) I believe
this one will appease everyone because it:
- does not touch etc/defaults/rc.conf template file,
- patches it at build time for MD hooks only when required,
- does not need to parse/modify a user-specified file like /etc/rc.conf (which
is a complex, error-prone operation),
- only enables powerd(8) by default when conditions are met (Xen environment)
while still allowing root to shoot himself in the foot if he wants to
override this manually in /etc/rc.conf.

See also http://mail-index.netbsd.org/tech-userlevel/2011/07/25/msg005246.html
2011-08-22 18:54:05 +00:00
reinoud
778142a2ba Implement pmap_extract() and print pmap_zero_page() pa address. 2011-08-22 16:22:16 +00:00
reinoud
3b8975f0ed Start NetBSD/usermode's pmap. Its using a temp file as a physical memory
backup and that should be documented in the code. A physical address is thus a
file offset(!) and a virtual address is a `normal' accesible address.

Still to do: various misc functions and pmap_extract() in special.

Credits also go to Ben Harris for his work on the Acorn26 pmap that followed
the Daemon Book recommendation for systems without real page tables on wich
this implementation is modelled after.
2011-08-22 15:36:23 +00:00
reinoud
a436c9264e Add thunk_munmap() to NetBSD/usermode's thunk 2011-08-22 15:30:16 +00:00
reinoud
fca46a08f4 Change standard settings in sys/arch/usermode/conf/GENERIC so to get a better
debugging environment.
2011-08-22 15:28:34 +00:00
reinoud
e778fd293b Don't override -O options... we might want to build with another -O setting in
the config file.
2011-08-22 15:27:32 +00:00
reinoud
0fe675bc9f Bump NKMEMPAGES from 512 to 2048 for NetBSD/usermode. This will be sufficient
for a 128 MB amd64 machine. Might better be parameterized one day.
2011-08-22 15:26:55 +00:00
mrg
f63becec07 disable mmx/sse here too. hopefully fixes amd64 /boot issues.
certainly changes the output in ways that gcc 4.1 doesn't.
2011-08-22 09:43:08 +00:00
he
d6e8830ca4 Hm, upstream maintainer says loop needs to start from 0, make it so. 2011-08-22 09:19:51 +00:00
wiz
a223481bf4 file-5.08 out. 2011-08-22 08:33:24 +00:00
he
8d79df0020 Replace needless use of double math just to compare colour distances
with integer arithmetic.  Reported to and discussed with upstream maintainer.
2011-08-22 06:52:35 +00:00
dholland
ce24bfd93f Until someone figures out a fix for PR 40599, document that MKKERBEROS=no
will break the system without either MKPAM=no or a customized PAM config.
As suggested by Ian D. Leroux on current-users and in PR 45263, but with
different text.
2011-08-22 02:37:15 +00:00
dholland
75fea7f548 Add patch from Klaus Klein in PR 45258 to note that popcount32/64
require <stdint.h>.
2011-08-22 01:36:04 +00:00
dholland
112cae6f8e Don't pollute the namespace with stdint.h. (This file is included from
string.h so it can't blithely do whatever it feels like.) PR 45258.
2011-08-22 01:24:15 +00:00
dholland
ef311a878c Requires stdint.h. 2011-08-22 00:33:16 +00:00