Attilio Rao 86d2e48c22 Per their definition, atomic instructions used in conjuction with ... memory barriers should also ensure that the compiler doesn't reorder paths where they are used. GCC, however, does that aggressively, even in presence of volatile operands. The most reliable way GCC offers for avoid instructions reordering is clobbering "memory" even if that is theoretically an heavy-weight operation, flushing the content of all the registers and forcing reload of them (We could rely, however, on gcc DTRT by just understanding the purpose as this is a well-known pattern for many modern operating-systems). Not all our memory barriers, right now, clobber memory for GCC-like compilers. The most notable cases are IA32 and amd64 where the memory barrier are treacted the same as normal atomic instructions. Fix this by offering the possibility to implement atomic instructions with memory barriers separately from the normal version and implement the GCC-like specific one using memory clobbering. Thanks to Chris Lattner (@apple) for his discussion on llvm specifics. Reported by: jhb Reviewed by: jhb Tested by: rdivacky, Giovanni Trematerra <giovanni dot trematerra at gmail dot com>		2009-10-06 13:45:49 +00:00
..
amd64	Per their definition, atomic instructions used in conjuction with	2009-10-06 13:45:49 +00:00
arm	- Drop unused pmap_use_l1 function and comment out currently unused	2009-10-05 10:08:58 +00:00
boot	lindev(4) [1] is supposed to be a collection of linux-specific pseudo	2009-09-26 12:45:28 +00:00
bsm	Add audit events for process descriptor system calls, which will appear in	2009-09-29 21:25:59 +00:00
cam	Report SATA 3.x devices.	2009-09-27 20:50:54 +00:00
cddl	Return EOPNOTSUPP instead of EINVAL when doing chflags(2) over an old	2009-10-01 18:58:26 +00:00
compat	Make sure that the primary native brandinfo always gets added	2009-10-03 11:57:21 +00:00
conf	cxgb(4) updates, including:	2009-10-05 20:21:41 +00:00
contrib	Use __FBSDID to embed RCS ID.	2009-09-21 08:37:57 +00:00
crypto
ddb	Put square backets ([]) around process names for system processes to patch	2009-10-01 19:12:14 +00:00
dev	cxgb(4) updates, including:	2009-10-05 20:21:41 +00:00
fs	Fix a bug that causes the fsx test case of mmap'ed page being out of sync	2009-10-04 10:38:04 +00:00
gdb
geom	- Improve error message consistency and wording.	2009-10-05 08:44:31 +00:00
gnu	Fix the build by using proper format.	2009-06-25 16:48:13 +00:00
i386	Per their definition, atomic instructions used in conjuction with	2009-10-06 13:45:49 +00:00
ia64	Make sure that the primary native brandinfo always gets added	2009-10-03 11:57:21 +00:00
isa	Tweak the way that the ACPI and ISA bus drivers match hint devices to	2009-08-24 21:51:46 +00:00
kern	Fix build on amd64, where sysctl arg1 is a pointer.	2009-10-05 22:23:12 +00:00
kgssapi	When the KOBJMETHOD() macro was updated, it resulted in the	2009-06-14 17:33:46 +00:00
libkern	done method is supposed to return int.	2009-06-22 22:09:18 +00:00
mips	Make sure that the primary native brandinfo always gets added	2009-10-03 11:57:21 +00:00
modules	cxgb(4) updates, including:	2009-10-05 20:21:41 +00:00
net	Put #ifdef INET around parts of the FLOWTABLE code, to unbreak	2009-10-03 10:56:03 +00:00
net80211	Update 802.11s mesh support to draft 3.03. This includes a revised frame	2009-09-22 18:18:14 +00:00
netatalk	Reverse misordered unlock and lock in at_control for netatalk phase I	2009-08-12 10:44:13 +00:00
netgraph	Get those pesky RFCOMM RPM data bits right. This is likely a noop.	2009-09-10 23:30:13 +00:00
netinet	In tcp_input(), we acquire a global write lock at first only if a	2009-10-05 22:24:13 +00:00
netinet6	Enable adding a link-local address even if ND6_IFF_IFDISABLED.	2009-10-02 07:00:20 +00:00
netipsec	Changed an IPSEC_ASSERT to a simple test, as such invalid packets	2009-10-01 15:33:53 +00:00
netipx	Use queue(9) instead of hand-crafted link lists for the global IPX	2009-06-24 20:57:50 +00:00
netnatm
netncp
netsmb	Don't print out a message on loading a module. 'kldload -v' and 'kldstat	2009-09-10 18:33:08 +00:00
nfs	Revert rev 192323 (nfs_common.c only):	2009-07-12 03:53:52 +00:00
nfsclient	Reverting the previous change for now. Some users reports the patch	2009-09-15 22:09:42 +00:00
nfsserver	Ensure that tv_sec is between INT32_MIN and INT32_MAX, so ZFS won't object.	2009-09-26 18:23:16 +00:00
nlm	unifdef NFSCLIENT because the nlm depends on the nfsclient even if NFSCLIENT	2009-10-03 12:22:12 +00:00
opencrypto	If crypto operation is finished with EAGAIN, don't repeat operation from	2009-09-04 09:48:18 +00:00
pc98	MFi386: revision 197653	2009-10-01 10:46:22 +00:00
pci	intpm/sb700: force polling mode if configured interrupt is SMI	2009-09-19 08:56:28 +00:00
powerpc	Make sure that the primary native brandinfo always gets added	2009-10-03 11:57:21 +00:00
rpc	Set the prison in NFS anon and GSS SVC creds (as I indended to in r197581).	2009-09-28 18:55:29 +00:00
security	Having thrown the cat out of the house, add a necessary include.	2009-09-08 13:24:36 +00:00
sparc64	Make sure that the primary native brandinfo always gets added	2009-10-03 11:57:21 +00:00
sun4v	Add a new sysctl for reporting all of the supported page sizes.	2009-09-18 17:04:57 +00:00
sys	Reserve numbers for XScale.	2009-10-02 11:14:12 +00:00
teken	Add support for VT200-style mouse input.	2009-09-27 18:19:41 +00:00
tools	- Increase dynamic range of filter coefficients from 28bit to 30bit.	2009-07-05 18:15:06 +00:00
ufs	Don't build ufs_gjournal.c at all if UFS_GJOURNAL option is not given	2009-09-22 16:22:05 +00:00
vm	Align and pad the page queue and free page queue locks so that the linker	2009-10-04 18:53:10 +00:00
xdr	Add a check for a NULL mbuf ptr at the beginning of xdrmbuf_inline()	2009-08-12 16:27:51 +00:00
xen	Temporarily revert the new-bus locking for 8.0 release. It will be	2009-08-20 19:17:53 +00:00
Makefile