r343904 pluknet Feb. 8, 2019, 11:49 a.m.
r343899 kib Feb. 8, 2019, 8:20 a.m.
The vp vnode is unlocked during the execution of the VOP method and
can be reclaimed, zeroing vp->v_data.  Caching allows to use the
correct mount point.

Reported and tested by:	pho
PR: 235549
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
ViewVC
r343898 kib Feb. 8, 2019, 8:17 a.m.
Reported and tested by:	pho
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
ViewVC
r343897 kib Feb. 8, 2019, 8:15 a.m.
Tested by:	pho
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
ViewVC
r343896 kib Feb. 8, 2019, 7:36 a.m.
This is also OpenBSD rev. 1.117, as pointed out by
Ryan Moeller <ryan@ixsystems.com>.

Submitted by:	Nash Kaminski <nashkaminski@gmail.com>
MFC after:	1 week
ViewVC
r343895 glebius Feb. 8, 2019, 6:19 a.m.
PR:		235500
MFC after:	1 week
ViewVC
r343891 kib Feb. 8, 2019, 4:18 a.m.
When renameat(2) is used with:
- absolute path for to;
- tofd not set to AT_FDCWD;
- the target exists
kern_renameat() requires CAP_UNLINK capability on tofd, but
corresponding namei ni_filecap is not initialized at all because the
lookup is absolute.  As result, the check was done against empty filecap
and syscall fails erronously.

Fix it by creating a return flags namei member and reporting if the
lookup was absolute, then do not touch to.ni_filecaps at all.

PR:	222258
Reviewed by:	jilles, ngie
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
X-MFC-note:	KBI breakage
Differential revision:	https://reviews.freebsd.org/D19096
ViewVC
r343890 kib Feb. 8, 2019, 4:06 a.m.
Code after exec_fail_dealloc label expects that the image vnode is
locked if present.  When copyout() of the strings or auxv vectors fails,
goto to the error handling did not relocked the vnode as required.

The copyout() can be made failing e.g. by creating an ELF image with
PT_GNU_STACK segment disabling the write.

Reported by:	Jonathan Stuart <n0t.jcs@gmail.com> (found by fuzzing)
Sponsored by:	The FreeBSD Foundation
MFC after:	3 days
ViewVC
r343889 np Feb. 8, 2019, 3:31 a.m.
r343884 brooks Feb. 7, 2019, 11:17 p.m.
Back in 1993, the fgetln (then fgetline) interface was changed to not
return a C string.  The change was accomplished by ifdefing out the code
that did the termination.  Changing the interface would violate our API
stability rules so remove the old implementation.

Sponsored by:	DARPA, AFRL
ViewVC
r343882 sef Feb. 7, 2019, 10:10 p.m.
PR:             233849
Reported by:    Andre Albsmeier
MFC after:      1 month
Sponsored by:   iXsystems Inc
Differential Revision:  https://reviews.freebsd.org/D18785
ViewVC
r343881 sef Feb. 7, 2019, 9:51 p.m.
Refactor the function calls and tests so that, on UFS, the proper fields
are filled out.

PR:		233849
Reported by:	Andre Albsmeier
Reviewed by:	mav, delphij
MFC after:	1 month
Sponsored by:	iXsystems Inc
Differential Revision:	https://reviews.freebsd.org/D18785
ViewVC
r343880 bdrewery Feb. 7, 2019, 9:41 p.m.
MFC after:	2 weeks
X-MFC-with:	r343877
Pointyhat to:	bdrewery
ViewVC
r343877 bdrewery Feb. 7, 2019, 9:05 p.m.
Properly pass down is_all to table_show_list().  This restores the behavior
before r272840 so that only 'ipfw table all list' shows the headers.

MFC after:	2 weeks
Relnotes:	yes
ViewVC
r343876 andrew Feb. 7, 2019, 8:58 p.m.
When moving from an invalid to a valid entry we don't need to invalidate
the tlb, however we do need to ensure the store is ordered before later
memory accesses. This is because this later access may be to a virtual
address within the newly mapped region.

Add the needed barriers to places where we don't later invalidate the
tlb. When we do invalidate the tlb there will be a barrier to correctly
order this.

This fixes a panic on boot on ThunderX2 when INVARIANTS is turned off:
panic: vm_fault_hold: fault on nofault entry, addr: 0xffff000040c11000

Reported by:	jchandra
Tested by:	jchandra
Sponsored by:	DARPA, AFRL
Differential Revision:	https://reviews.freebsd.org/D19097
ViewVC