summaryrefslogtreecommitdiffstats
path: root/lib/libc/stdlib/malloc.c (unfollow)
Commit message (Collapse)AuthorFilesLines
2013-04-03Set the stream orientation in open_{,w}memstream().guenther3-3/+19
Check it in the regress test ok mpi@
2013-04-03Handle big time_tguenther1-2/+2
ok deraadt@
2013-04-02add comment for how EINPROGRESS is handled there.eric1-1/+6
2013-04-02Use __off_t instead of off_t to make this usable on its own even in ourkettenis1-3/+3
default compilation environment. ok espie@, millert@, guenther@
2013-04-02make sure we always point before the current offset.eric1-2/+2
spotted by matthew@
2013-04-02Revert rev 1.21 to fix a race condition where multiple mkdir -p's tryingnaddy1-11/+7
to create overlapping paths in parallel could error out. ok millert@
2013-04-02Use macros sotoinpcb() and intotcpcb() instead of casts. Use NULLbluhm7-24/+23
instead of 0 for pointers. No binary change. OK mpi@
2013-04-02better implementation for tcp_read() that can get the packet length ineric2-23/+35
multiple read. prodded by deraadt@
2013-04-02- Have the tht(4) ioctl handler for SIOCSIFFLAGS set ENETRESETbrad1-4/+6
instead of calling tht_iff() directly. - Add a comment in tht_up() about the use of tht_iff(). - Have tht_iff() set IFF_ALLMULTI when in promisc mode. ok dlg@
2013-04-02Set the IFF_ALLMULTI flag as appropriate.brad1-6/+7
ok dlg@
2013-04-02Get rid of SYSCALL_G7RFLAG. iIt's been deprecated for a long time already,kettenis2-7/+6
and we've started using %g7 as a per-thread register now. If you have any binaries left that use this flag they'll probably crash and burn!
2013-04-02Remove Evergreen ids people have incorrectly added to radeondrm.jsg1-5/+1
There is currently no support for Evergreen/R800. ok mpi@
2013-04-02change log format to fix warninggilles1-3/+3
2013-04-02Instead of storing the link-level address of every interface in a globalmpi4-34/+20
array indexed by interface numbers, add a new field to the interface descriptor pointing to it. claudio@ and todd@ like it, ok mikeb@
2013-04-02Let this work on gcc < 4, which don't accept /dev/null as a valid input file.miod1-10/+28
2013-04-02syncderaadt1-0/+2
2013-04-02Handle big time_tguenther2-6/+6
ok deraadt@
2013-04-02handle large time_tderaadt1-2/+5
2013-04-02Pull in <sys/_types.h> instead of <machine/_types.h> in prep for the futureguenther2-4/+4
ok deraadt@
2013-04-02for parsing time values, fall back to atoll() for time_tderaadt1-2/+2
ok guenther
2013-04-02kill useless castderaadt1-2/+2
2013-04-02handle larger time_tderaadt1-2/+2
ok guenther
2013-04-02handle large time_t in DEBUG codederaadt1-2/+2
2013-04-02handle time_t which is not u_int32_tderaadt1-3/+5
ok guenther
2013-04-02Use internal types instead of the old BSD u_int#_t types in theguenther1-5/+5
standard portion of the header. Most of the diff from James Turner (james (at) calminferno.net)
2013-04-02Remove "/IPv4" from the comments of the M_{TCP,UDP}_CSUM_IN_* flags,lteo1-5/+5
since they are also used for TCP or UDP checksums over IPv6 (originally pointed out by naddy@ when I was updating the mbuf(9) man page). ok bluhm claudio
2013-04-02tv_sec is not neccessarily the same type (or size) as time_tderaadt1-8/+6
ok guenther
2013-04-02Stop assuming time_t is longguenther2-8/+8
ok deraadt@
2013-04-02handle large time_t at least until 2038. then chaos breaks out.deraadt1-1/+3
ok guenther note to tedu: 2038 -> rwhod
2013-04-02Use a time_t variable with ctime()guenther1-2/+2
ok deraadt@
2013-04-02handle long time_t; ok guentherderaadt1-3/+5
2013-04-02Don't assume sizeof(time_t)==4guenther1-3/+4
ok krw@
2013-04-02recognize and ignore power change event from newer thinkpadsjcs1-28/+22
group and sort ignored events ok deraadt
2013-04-01Allow raw IPv6 sockets for IPsec protocols. IPv4 does the same.bluhm1-7/+7
Ok claudio@
2013-04-01There is no IPv6 default scope in OpenBSD. Remove leftovers.bluhm1-28/+1
No binary change. OK claudio@
2013-04-01If more than one lookup line is found in resolv.conf, the latest oneeric1-18/+9
takes precedence. Simplify code while there. suggested by deraadt@
2013-04-01properly check for domain name truncation at various places and faileric3-20/+30
if that happens. prodded by deraadt@
2013-04-01When there are multiple groups of functions with their own #includesguenther14-46/+55
listed, the #includes in each group are independent, so that users don't need to guess at possible carry-over. general idea ok otto@ millert@ actual diff ok jmc@ schwarze@
2013-04-01Make ugen_get_cdesc() a generic function by renaming it to usbd_get_cdesc()mglocker3-45/+43
and move it to usb_subr.c. OK mpi@
2013-04-01Build mips kernels with -G 0, to disable use of so-called `small' data andmiod3-6/+6
bss sections. The current kernel linker script is not gp-friendly enough for that; and while gas 2.15 was not attempting to output gp-relative relocations for variables which might have ended up in .sdata or .sbss, gas 2.17 will, and the kernel will fail to link. To be improved eventually with a better kernel ld script putting the gp-addressable sections close enough to gp... and making sure kernel gp is reloaded in all the userland->kernel code paths which might need gp in the kernel.
2013-04-01Some firmware revisions on UltraSPARC T2 machines apparently insist on havingkettenis1-3/+24
"rngs" and "rng" nodes in the Hypervisor machine description.
2013-04-01Do not expose the trap T_* values to userland, since they invade thederaadt7-17/+14
namespace a bit much. These are the names for the value encoded in core.c_ucode or siginfo.si_trapno, but userland never looks directly and only cares about the abstract signal code. ok guenther