drm: fix a use-after-free when GPU acceleration disabled
When GPU acceleration is disabled, drm_vblank_cleanup() will free the
vblank-related data, such as vblank_refcount, vblank_inmodeset, etc.
But we found that drm_vblank_post_modeset() may be called after the
cleanup, which use vblank_refcount and vblank_inmodeset. And this will
cause a kernel panic.
Fix this by return immediately if dev->num_crtcs is zero. This is the
same thing that drm_vblank_pre_modeset() does.
r280299 (cperciva):
When creating VM images, copy the contents of the created filesystem
into a new filesystem before packaging it into a disk image. This
prevents "remnants" of deleted files from showing up in the VM images,
and reduces their compressed size (by about 10% for the cloudware
images) as a result.
r280840 (cperciva):
Clean up filesystem unmounting in vmimage builds:
- Remove vm_umount_base function which is currently unused.
- Add umount_loop function which loops attempting to unmount one
filesystem.
- Replace calls to umount with calls to umount_loop.
- Don't attempt to unmount ${DESTDIR}/dev if it isn't mounted.
The looping is necessary because sometimes umount fails due to
filesystems being busy. The most common cause of such busyness is
periodic(8) jobs running `find / ...`.
r280846 (cperciva):
Improve check for whether ${DESTDIR}/dev is mounted.
r280879 (cperciva):
Add bits for building EC2 disk images.
r280881 (cperciva):
Partially revert r278118 now that the required logic for deciding
whether freebsd-update can be useful has moved into the
firstboot_freebsd_update script.
r280884 (cperciva):
Fix pkg(8) command line in vm_extra_pre_umount().
r280928 (cperciva):
Add code for creating an EC2 AMI.
r281783:
When building VM disk images, vm_copy_base() uses tar(1) to copy
the userland from one md(4)-mounted filesystem to a clean filesystem
to prevent remnants of files that were added and removed from
resulting in an unclean filesystem. When newfs(8) creates the first
filesystem with journaled soft-updates enabled, the /.sujournal file
in the new filesystem cannot be overwritten by the /.sujournal in the
original filesystem.
To avoid this particular error case, do not enable journaled
soft-updates when creating the md(4)-backed filesystems, and instead
use tunefs(8) to enable journaled soft-updates after the new
filesystem is populated in vm_copy_base().
While here, fix a long standing bug where the build environment /boot
files were used by mkimg(1) when creating the VM disk images by using
the files in .OBJDIR.
r281809:
Simplify variable expansion in attempt to fix the vm-image build.
MFC r281495:
Add config option PAE_TABLES for the i386 kernel. It switches pmap to
use PAE format for the page tables, but does not incur other
consequences of the full PAE config. In particular, vm_paddr_t and
bus_addr_t are left 32bit, and max supported memory is still limited
by 4GB.
The option allows to have nx permissions for memory mappings on i386
kernel, while keeping the usual i386 KBI and avoiding the kernel data
sizing problems typical for the PAE config.
Add test case for kern/181741. Right now test fails.
PR: 181741
Sponsored by: Nginx, Inc.
r281354:
Fix warnings, fix a typo in a testcase description, bump WARNS to 3
- Remove argc/argv (-Wunused)
- Cast len in comparison to size_t (-Wsign-compare)
Sponsored by: EMC / Isilon Storage Division
r281355:
Fix -Wunused warnings, bump WARNS to 6
The testcase fails today on subtest # 9
The output is still broken if prove -rv is run and the testcase aborts
prematurely (the testcase doesn't really conform to TAP protocol properly,
except when it completes fully)
Sponsored by: EMC / Isilon Storage Division
r281356:
Fix -Wunused warnings, bump WARNS to 6
The output is still broken if prove -rv is run and the testcase aborts
prematurely with fail_assertion (the testcase doesn't really conform to TAP
protocol properly, except when it completes fully)
Sponsored by: EMC / Isilon Storage Division
r281358:
- Parameterize out the number of accept/connect attempts
- Randomize the bind port to allow 2+ consecutive calls in < 10 minutes, and
to also not fail if (for instance) there's a server already listening on port
8080
- Don't leak the listening socket / fds into the child process
- Fix warnings:
-- Remove argc/argv (-Wunused)
-- Mark sig __unused (-Wunused)
-- Mark quit static (-Wmissing-variable-declarations)
Sponsored by: EMC / Isilon Storage Division
r281359:
Remove argc/argv (-Wunused)
Sponsored by: EMC / Isilon Storage Division
r281360:
Fix warnings
- Remove argc/argv (-Wunused)
- Mark some parameters to socket_listen_update __unused (-Wunused)
Sponsored by: EMC / Isilon Storage Division
r281361:
Remove argc/argv (-Wunused)
Sponsored by: EMC / Isilon Storage Division
r281362:
Use _exit, not exit in forked process
Sponsored by: EMC / Isilon Storage Division
r281391:
- Use static buffers for temporary file paths instead of strdup of constant strings
- Don't use /tmp because it's outside ATF's prescribed sandbox
- Use mkstemp instead of mktemp to eliminate warning
Sponsored by: EMC / Isilon Storage Division
r281392:
- Garbage collect argc/argv (-Wunused)
- Bump WARNS to 6
Sponsored by: EMC / Isilon Storage Division
r281393:
Fix warnings and bump WARNS to 6
- Garbage collect argc/argv (-Wunused)
- sleep(3) will always return an unsigned int; don't check for return codes <0
(-Wsign-compare)
Sponsored by: EMC / Isilon Storage Division
r281394:
- Don't use /tmp because it's outside ATF's prescribed sandbox
- Replace a hardcoded PATH_MAX value with sizeof(path)
- Use path like an array, not a pointer, and always try to unlink it in cleanup
Sponsored by: EMC / Isilon Storage Division
r281395:
Fix a -Wuninitialized warning by setting the socket to -1 and bump WARNS to 6
Sponsored by: EMC / Isilon Storage Division
r281397:
Mark signum unused in signal_handler; bump WARNS to 6
Sponsored by: EMC / Isilon Storage Division
r281398:
Garbage collect argc/argv and bump WARNS to 6
Sponsored by: EMC / Isilon Storage Division
r281399:
Fix warnings and bump WARNS to 6
- Staticize variables as needed
- Garbage collect argc/argv
- Fix -Wsign-compare warnings by casting small sizeof to (int)
Sponsored by: EMC / Isilon Storage Division
r281400:
- Garbage collect argc/argv; bump WARNS to 6
- Make the socket path random and move it out of /tmp as that's outside ATF's
prescribed path
Sponsored by: EMC / Isilon Storage Division
r281401:
- Garbage collect argc/argv
- Use random paths instead of one in /tmp
Sponsored by: EMC / Isilon Storage Division
r281402:
Garbage collect argc/argv and bump WARNS to 6
Sponsored by: EMC / Isilon Storage Division
r281403:
Garbage collect argc/argv and bump WARNS to 6
Sponsored by: EMC / Isilon Storage Division
r281404:
Generate temporary files with mkstemp instead of mktemp
Sponsored by: EMC / Isilon Storage Division
r281407:
Fix the knob twiddling to work properly per src.opts.mk
Sponsored by: EMC / Isilon Storage Division
r281408:
- Remove the .t wrapper and put the "magic" of determining the number of
testcases into the .c file
- Require root for now because it fails with SOCK_RAW without root privileges
- Increment the test count properly on socket create failure
Sponsored by: EMC / Isilon Storage Division
r281409:
Fix warnings, bump WARNS to 6, and use a temporary socket instead of one in /tmp
Sponsored by: EMC / Isilon Storage Division
r281410:
Fix more warnings I didn't catch in the first go-around
Cleanup and do minor refactoring before converting testcases to ATF
- Convert errx(-1, ..) to errx(1, ..)
- Move the aio(4) checks to a single function (aio_available); use modfind(2)
instead of depending on SIGSYS (doesn't work when aio(4) support is missing,
not documented in the aio syscall manpages).
- Use aio_available liberally in the testcase functions
- Use mkstemp(3) + unlink(2) instead of mktemp(3)
- Fix some -Wunused warnings
- Bump WARNS to 6
MFC r281544:
snd_hda: add support for the Lenovo X1 20BS model.
This requires a patch to redirect the output to a separate DAC when
the headphones are used. While there, add device strings for Intel
Broadwell HDA controllers and Realtek ALC292 codecs.
v1.1.1 2015-03-26
* Support 4Kn drive.
* Change the SCSI target ID of the disk to be the index of physical
connetion to the HBA.
* Support staggered drive spin up.
* Fix a bug that command would be timeout because of improper
interrupt service routine.
* Error handling to avoid scsi command lost which caused system
hang up.
* Fix a bug that fail to get the devcie's serial number via
FreeNAS WebGUI.
Many thanks to HighPoint for continued support of FreeBSD!
hiren [Fri, 24 Apr 2015 23:26:44 +0000 (23:26 +0000)]
MFC r275358 r275483 r276982 - Removing M_FLOWID by hps@
r275358:
Start process of removing the use of the deprecated "M_FLOWID" flag
from the FreeBSD network code. The flag is still kept around in the
"sys/mbuf.h" header file, but does no longer have any users. Instead
the "m_pkthdr.rsstype" field in the mbuf structure is now used to
decide the meaning of the "m_pkthdr.flowid" field. To modify the
"m_pkthdr.rsstype" field please use the existing "M_HASHTYPE_XXX"
macros as defined in the "sys/mbuf.h" header file.
This patch introduces new behaviour in the transmit direction.
Previously network drivers checked if "M_FLOWID" was set in "m_flags"
before using the "m_pkthdr.flowid" field. This check has now now been
replaced by checking if "M_HASHTYPE_GET(m)" is different from
"M_HASHTYPE_NONE". In the future more hashtypes will be added, for
example hashtypes for hardware dedicated flows.
"M_HASHTYPE_OPAQUE" indicates that the "m_pkthdr.flowid" value is
valid and has no particular type. This change removes the need for an
"if" statement in TCP transmit code checking for the presence of a
valid flowid value. The "if" statement mentioned above is now a direct
variable assignment which is then later checked by the respective
network drivers like before.
r275483:
Remove M_FLOWID from SCTP code.
r276982:
Remove no longer used "M_FLOWID" flag from mbuf.h and update the netisr
manpage.
Pad RX copy alignment calculation to avoid illegal memory accesses
The optimization made in r239940 is valid for struct mbuf's current structure
and size in FreeBSD, but hardcodes assumptions about sizes of struct mbuf,
which are unfortunately broken if additional data is added to the beginning of
struct mbuf
X-MFC note (discussed with rwatson):
This change requires the MPKTHSIZE definition, which is only available after
head@r277203 and will not be MFCed as it breaks mbuf(9) KPI.
A direct commit to stable/10 and merges to other branches to add the necessary
definitions to work with the code as-is will be done to facilitate this MFC
These macros are equivalent to the ones on head, except they are only exposed
when _KERNEL is defined, i.e. to kernel code, whereas the code on head is exposed
to userland as well
This is for improved forwards compatibility with mbuf(9) macros in head@r277203+,
and is required for a clean MFC of r279393
Revert r281543. It causes a panic/hang early in boot for a number of
users, myself included. The original code is likely papering over a
larger bug that needs to be explored, but for now get things back to
a working state.
Set ARCHIVE_EXTRACT_SECURE_SYMLINKS and ARCHIVE_EXTRACT_SECURE_NODOTDOT
as in bsdtar to prevent extraction of archive entries whose pathnames
contain .. or whose target directory would be altered by a symlink.
Also disallow absolute pathnames.
We don't currently provide an option to disable this behaviour (as
bsdtar's -P does). It is unlikely to be a problem in practice for ar(1),
but the -P option is not currently used and available if we want to
consider it for this purpose.
Obtained from: ELF tool chain ar, Ticket #474
Relnotes: Yes
Sponsored by: The FreeBSD Foundation
MFC r281380:
Fix the IPV6_MULTICAST_IF sockopt handling. RFC 3493 says when the
interface index is specified as zero, the system should select the
interface to use for outgoing multicast packets. Even the comment
for the in6p_set_multicast_if() function says about index of zero.
But in fact for zero index the function just returns EADDRNOTAVAIL.
I.e. if you first set some interface and then will try reset it
with zero ifindex, you will get EADDRNOTAVAIL.
Reset im6o_multicast_ifp to NULL when interface index specified as
zero. Also return EINVAL in case when ifnet_byindex() returns NULL.
This will be the same behaviour as when ifindex is bigger than
V_if_index. And return EADDRNOTAVAIL only when interface is not
multicast capable.
MFC 281159:
Move the message complaining about failed system resource allocations
under bootverbose. Every example I've seen to date has been due to
an ACPI system resource device reserving a range that overlaps with
system memory (which ram0 attempts to reserve) or a local or I/O APIC
(which apic0 attempts to reserve). These are always harmless but look
scary to users.
MFC r281309:
Fix the check for maximum mbuf's size needed to send ND6 NA and NS.
It is acceptable that the size can be equal to MCLBYTES. In the later
KAME's code this check has been moved under DIAGNOSTIC ifdef, because
the size of NA and NS is much smaller than MCLBYTES. So, it is safe to
replace the check with KASSERT.
MFC revisions 277693,278335,280382-280385,280923-280926,280931,
280933-280939,280974-280976,281002,281009,281081,281176-281180,
281271,281275,281616 (described in-breif below):
r277693: Font fix (des)
r278335: Revert that
r280382: Whitespace, comments, and copyright update
r280383: Prevent inadvertent bootlock condition
r280384: Increase max passowrd length from 16 to 255 chars
r280385: Add missing variable hints to loader.conf(5) defaults
r280923: Whitespace
r280924: Comments
r280925: Optimize bootmsg to use fg/bg/me from screen.4th
r280926: Whitespace and cleanup
r280931: Comments
r280933: Move beastie to logo-*.4th; brands to brand-*.4th
r280934: Add remainder of supported ANSI escape sequences
r280935: Securely overwrite (zero) user input after password checks
r280936: Use equals for ASCII double frames
r280937: Solve dreaded "dictionary full" issue
r280938: Add "GELI Passphrase:" prompt to boot loader
r280939: Revert that (premature commit)
r280974: Use fg/b/me from screen.4th instead of literals
r280975: Eliminate literal escape sequences from *.4th
r280976: Use ^[[m mode-ending versus ^[[37m
r281002: Install newly added brand-*.4th and logo-*.4th files (jkim)
r281009: Revert .PATH changes to fix mips build (jkim)
r281081: Make sure forth manpages are only installed once (bapt)
r281176: Back to previous mode-endings based on feedback
r281177: Back to previous mode-endings based on feedback
r281178: Back to previous mode-endings based on feedback
r281179: Back to previous mode-endings based on feedback
r281180: Eliminate literal escape sequences from *.rc
r281271: Fix a bootlock condition if loader_version is set
NB: Commit message of r281271 has a typo, s/_logo/_version/
r281275: Re-do proper mode-endings
r281616: Add "GELI Passphrase:" prompt to boot loader
Relnotes: Added "GELI Passphrase:" prompt to boot loader
MFC r281670, r281703:
Drop experimental ext2fs dir_index support.
The htree directory index is a highly desirable feature for research
purposes and was meant to improve performance in our ext2/3 driver.
Unfortunately our implementation has two problems:
- It never really delivered any performance improvement.
- It appears to corrupt the filesystem in undetermined circumstances.
Strictly speaking dir_index is not required for read/write support in
ext2/3 and our limited ext4 support still works fine without it.
Regain stability in the ext2 driver by removing it. We may need it back
(fixed) if we want to support encrypted ext4 support but thanks to the
wonders of version control we can always revert this change and bring it
back.
Per Austin group issue #884, sh should not import IFS from the environment
but always set it to $' \t\n'. For wordexp(), however, it is documented and
useful for it to use IFS from the environment.
Since sh currently imports IFS from the environment, this change has no
functional effect.
MFC r281082: fts: Don't return FTS_SLNONE if it's not a symlink (if race).
When following symlinks, fts returned FTS_SLNONE when fstatat(flag=0)
failed, but a subsequent fstatat(flag=AT_SYMLINK_NOFOLLOW) succeeded. This
incorrectly triggered if a filename existed to be read from the directory,
was deleted before the fstatat(flag=0) and created again after the
fstatat(flag=0).
Fix this by only returning FTS_SLNONE if the result from
fstatat(flag=AT_SYMLINK_NOFOLLOW) is actually a symlink. If it is not a
symlink, treat it as if fstatat(flag=0) succeeded.
MFC r260571: fts: Stat things relative to the directory fd, if possible.
As a result, the kernel needs to process shorter pathnames if fts is not
changing directories (if fts follows symlinks (-L option to utilities), fts
cannot open "." or FTS_NOCHDIR was specified).
Side effect: If pathnames exceed PATH_MAX, [ENAMETOOLONG] is not hit at the
stat stage but later (opendir or application fts_accpath) or not at all.
Enable two finger scrolling by default and disable the edge scrolling if
the touchpad has no physical zone for it. Disable directional scrolling
by default to avoid using extended buttons as scroll buttons.
Add support for ClickPad. On Lenovo laptops, this is the button
reported when one presses the touchpad.
While there, fix a problem where the extended buttons were not reporting
the button release event correctly: we need to save the state of the
buttons and report it to sysmouse until we receive a packet from the
touchpad indicating the button has been released. This makes it
possible to use an extended button to resize a window. On Lenovo
laptops, the major buttons are actually reported as extended buttons.
MFC 280222:
Clear an mbuf's external storage flags in m_extaddref(). They are cleared
in other places that set the external storage type (ext_type) such as
m_cljset(), m_extadd(), mb_ctor_clust(), and vn_sendfile().
bthidd: Consider usage ranges when dealing with array inputs.
So far, we were always using HID_USAGE() to determine the Usage ID of a
certain HID report input item. This does not work as intended if a field
is an array and the allowed usages are specified with a usage range, as
HID_USAGE() will return 0. We need to use the field value as an index in
the usage range list in this case instead.
This makes the volume keys in a Microsoft Bluetooth Mobile Keyboard
5000 be properly recognized. The relevant part of the HID report looks
like this:
MFC 278325,280866:
Revert the IPI startup sequence to match what is described in the
Intel Multiprocessor Specification v1.4. The Intel SDM claims that
278325:
Revert the IPI startup sequence to match what is described in the
Intel Multiprocessor Specification v1.4. The Intel SDM claims that
the INIT IPIs here are invalid, but other systems follow the MP
spec instead.
While here, fix the IPI wait routine to accept a timeout in microseconds
instead of a raw spin count, and don't spin forever during AP startup.
Instead, panic if a STARTUP IPI is not delivered after 20 us.
280866:
Wait 100 microseconds for a local APIC to dispatch each startup-related IPI
rather than 20. The MP 1.4 specification states in Appendix B.2:
"A period of 20 microseconds should be sufficient for IPI dispatch to
complete under normal operating conditions".
(Note that this appears to be separate from the 10 millisecond (INIT) and
200 microsecond (STARTUP) waits after the IPIs are dispatched.) The
Intel SDM is silent on this issue as far as I can tell.
At least some hardware requires 60 microseconds as noted in the PR, so
bump this to 100 to be on the safe side.
MFC r281272:
Explain that vm_page_array is mapped to describe the memory, not the
memory itself. Provide the formula to calculate the number of
required page tables. Correct the size of the struct vm_page for
non-PAE case.
MFC 278474,278476,280279:
Expand the bitcount*() API and use it to implement CPU_COUNT for cpusets.
278474:
Use __builtin_popcnt() to implement a BIT_COUNT() operation for bitsets and
use this to implement CPU_COUNT() to count the number of CPUs in a cpuset.
278476:
Use __builtin_popcountl() instead of __builtin_popcount().
280279:
Expand the bitcount* API to support 64-bit integers, plain ints and longs
and create a "hidden" API that can be used in other system headers without
adding namespace pollution.
- If the POPCNT instruction is enabled at compile time, use
__builtin_popcount*() to implement __bitcount*(), otherwise fall back
to software implementations.
- Use the existing bitcount16() and bitcount32() from <sys/systm.h> to
implement the non-POPCNT __bitcount16() and __bitcount32() in
<sys/types.h>.
- For the non-POPCNT __bitcount64(), use a similar SWAR method on 64-bit
systems. For 32-bit systems, use two __bitcount32() operations on the
two halves.
- Use __bitcount32() to provide a __bitcount() that operates on plain ints.
- Use either __bitcount32() or __bitcount64() to provide a
__bitcountl() that operates on longs.
- Add public bitcount*() wrappers for __bitcount*() for use in the kernel
in <sys/libkern.h>.
- Use __bitcountl() instead of __builtin_popcountl() in BIT_COUNT().