aarch64: fix mod/ref emulation in the MI pmap support code
Correct the initial PTE settings as per the table below
ref | mod | AF | AP | OS_MODEMUL
emul | emul | | |
-----+------+-----+---------+------------
no | no | set | RO/RW | not set
yes | no | clr | RO/RW | not set
no | yes | set | RO | set
yes | yes | clr | RO | set
[*] a write fault should be fixed up, and allowed.
and detect the modified emulation using OS_MODEMUL in pmap_fault fixup.
TODO:
- MI PMAP pmap_clear_reference needs to look more like pmap_clear_modify
and actually change PTE entries.
- MI PMAP needs support for HW updated modified and referenced bits.
aarch64: the MI pmap doesn't support HW modified / referenced bits.
Disable ARMV81_HAFDBS because the MI pmap doesn't handle HW supported
modified and referenced bits yet.
PR bin/60248 Various fixes to shutdown
First, fix (in a slightly different way than suggested)
the incorrect use of (1 minutes) sometimes.
While here, add more error checking to the shutdown time
parsing, and fix -k to be what it was originally intended
to be, which was nothing like "kick the users off" (which
it didn't do, unless all the "shutdown coming" messages
scared them away), all it did was install /etc/nologin so
no-one could log in (from 5 mins before the scheduled time).
That's pointless, if you want to install a /etc/nologin, just
install it (cat >/etc/nologin).
The -k flag was intended to assist with debugging, "k" was for
"kidding" - shutdown goes though all the motions, writing
messages, writing to syslog, etc - but that's all it does,
no manipulation of /etc/nologin at all, and no actual shutting
[6 lines not shown]
PR lib/60249 add some casts, and #include <stdlib.h>
Casts needed for assigning size_t value to int variable.
bit_alloc() uses calloc under the hood, and is a macro, so
anything calling it needs to have <stdlib.h> in scope. The
doc doesn't indicate it is required of the application, so
include it from <bitstring.h>
arm/generic: use generic setstatclockrate
This allows us to get rid of the weak symbol hack in a9tmr since GENERIC is the only configuration with both a9tmr and gtmr active in one configuration. Configurations without __HAVE_GENERIC_SETSTATCLOCKRATE still get a setstatclockrate from the ifdef.
arm/fdt: introduce a generic setstatclockrate function
These changes lay the groundwork for cleaning up setstatclockrate implementations across arm timers.
Pull up the following, requested by christos in ticket #283:
external/ibm-public/postfix//dist/README_FILES/NON_BERKELEYDB_README up to
external/ibm-public/postfix//dist/README_FILES/REQUIRETLS_README up to
external/ibm-public/postfix//dist/conf/postfix-non-bdb-script up to
external/ibm-public/postfix//dist/html/NON_BERKELEYDB_README.html up to
external/ibm-public/postfix//dist/html/REQUIRETLS_README.html up to
external/ibm-public/postfix//dist/html/nbdb_reindexd.8.html up to
external/ibm-public/postfix//dist/html/postfix-non-bdb.1.html up to
external/ibm-public/postfix//dist/man/man1/postfix-non-bdb.1 up to
external/ibm-public/postfix//dist/man/man8/nbdb_reindexd.8 up to
external/ibm-public/postfix//dist/mantools/check-proxy-type-table up to
external/ibm-public/postfix//dist/proto/NON_BERKELEYDB_README.html up to
external/ibm-public/postfix//dist/proto/REQUIRETLS_README.html up to
external/ibm-public/postfix//dist/src/cleanup/cleanup_message_test.c up to
external/ibm-public/postfix//dist/src/global/ehlo_mask_test.c up to
external/ibm-public/postfix//dist/src/global/nbdb_clnt.c up to
external/ibm-public/postfix//dist/src/global/allowed_prefix.c up to
external/ibm-public/postfix//dist/src/global/allowed_prefix.h up to
[575 lines not shown]
Explicitly cast IOCPARM_MAX to size_t - it can never be negative.
Hopefully avoids build breakage by a signed vs. unsigned comparision warning
on mips*64*.
risc-v: bump some parameter values on riscv64
Match all other 64bit platforms for
- PAGER_MAP_DEFAULT_SIZE
- UBC_WINSHIFT
- UBC_NWINS
Spotted by thorpej
Allow (p)redefinitions of symbols in <machine/param.h>
sun3 (via machine/param3.h) defines UBC_NWINS UBC_WINSHIFT & MAXEXEC
before including <m68k/param.h> == and defines UBC_NWINS differently.
Allow for that (hopefully fix broken sun3 build).
[ Why sun3 has __mc680010__ defined is left as an unexplained mystery ]