kib [Thu, 13 Jun 2013 09:33:22 +0000 (09:33 +0000)]
Fix two issues with the spin loops in the umtx(2) implementation.
- When looping, check for the pending suspension. Otherwise, other
usermode thread which races with the looping one, could try to
prevent the process from stopping or exiting.
- Add missed checks for the faults from casuword*(). The code is
structured in a way which makes the loops exit if the specified
address is invalid, since both fuword() and casuword() return -1 on
the fault. But if the address is mapped readonly, the typical value
read by fuword() is different from -1, while casuword() returns -1.
Absent the checks for casuword() faults, this is interpreted as the
race with other thread and causes non-interruptible spinning in the
kernel.
Reported and tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
glebius [Thu, 13 Jun 2013 06:07:19 +0000 (06:07 +0000)]
Improve locking strategy between keys hash and ID hash.
Before this change state creating sequence was:
1) lock wire key hash
2) link state's wire key
3) unlock wire key hash
4) lock stack key hash
5) link state's stack key
6) unlock stack key hash
7) lock ID hash
8) link into ID hash
9) unlock ID hash
What could happen here is that other thread finds the state via key
hash lookup after 6), locks ID hash and does some processing of the
state. When the thread creating state unblocks, it finds the state
it was inserting already non-virgin.
Now we perform proper interlocking between key hash locks and ID hash
lock:
1) lock wire & stack hashes
2) link state's keys
3) lock ID hash
4) unlock wire & stack hashes
5) link into ID hash
6) unlock ID hash
To achieve that, the following hacking was performed in pf_state_key_attach():
- Key hash mutex is marked with MTX_DUPOK.
- To avoid deadlock on 2 key hash mutexes, we lock them in order determined
by their address value.
- pf_state_key_attach() had a magic to reuse a > FIN_WAIT_2 state. It unlinked
the conflicting state synchronously. In theory this could require locking
a third key hash, which we can't do now.
Now we do not remove the state immediately, instead we leave this task to
the purge thread. To avoid conflicts in a short period before state is
purged, we push to the very end of the TAILQ.
- On success, before dropping key hash locks, pf_state_key_attach() locks
ID hash and returns.
pfg [Thu, 13 Jun 2013 03:23:24 +0000 (03:23 +0000)]
Relax some unnecessary unsigned type changes in ext2fs.
While the changes in r245820 are in line with the ext2 spec,
the code derived from UFS can use negative values so it is
better to relax some types to keep them as they were, and
somewhat more similar to UFS. While here clean some casts.
Some of the original types are still wrong and will require
more work.
jlh [Wed, 12 Jun 2013 21:12:05 +0000 (21:12 +0000)]
Turn libc.so into an ld script rather than a symlink pointing to the
real shared object and libssp_nonshared.a.
This was the last showstopper that prevented from enabling SSP for ports
by default. portmgr@ performed a buildworld which showed no significant
breakage with this patch.
Details:
On i386 for PIC objects, gcc uses the __stack_chk_fail_local hidden
symbol instead of calling __stack_chk_fail directly [1]. This happen
not only with our gcc-4.2.1 but also with the latest gcc-4.8. If you
want the very nasty details, see [2].
OTOH the problem doesn't exist on other architectures. It also doesn't
exist with Clang as the latter will somehow manage to create the
function in the object file at compile time (contrary to only
referencing it through a symbol that will be brought in at link time).
In a perfect world, when an object file is compiled with
-fstack-protector, it will be linked into a binary or a DSO with this
same flag as well, so GCC will add libssp_nonshared.a to the linker
command-line. Unfortunately, we don't control softwares in ports and we
may have such broken DSO. This is the whole point of this patch.
You can reproduce the problem on i386 by compiling a source file into an
object file with "-fstack-protector-all -fPIE" and linking it
into a binary without "-fstack-protector".
This ld script automatically proposes libssp_nonshared.a along with the
real libc DSO to the linker. It is important to understand that the
object file contained in this library will be pulled in the resulting
binary _only if_ the linker notices one of its symbols is needed (i.e.
one of the SSP symbol is missing).
A theorical performance impact could be when compiling, but my testing
showed less than 0.1% of difference.
[1] For 32-bit code gcc saves the PIC register setup by using
__stack_chk_fail_local hidden function instead of calling
__stack_chk_fail directly. See comment line 19460 in:
src/contrib/gcc/config/i386/i386.c
[2] When compiling a source file to an object file, if you use something
which is external to the compilation unit, GCC doesn't know yet if
this symbol will be inside or outside the DSO. So it expects the
worst case and routes the symbol through the GOT, which means
additional space and extra relocation for rtld(1).
Declaring a symbol has hidden tells GCC to use the optimal route (no
GOT), but on the other hand this means the symbol has to be provided
in the same DSO (namely libssp_nonshared.a).
On i386, GCC actually uses an hidden symbol for SSP in PIC objects
to save PIC register setup, as said in [1].
PR: ports/138228
PR: ports/168010
Reviewed by: kib, kan
emaste [Wed, 12 Jun 2013 20:11:49 +0000 (20:11 +0000)]
Handle DW_TAG_unspecified_type
2011-11-14 Jim Ingham <jingham@apple.com>
* dwarf2read.c (read_type_die): Handle DW_TAG_unspecified_type.
(read_tag_unspecified_type): New function, add a type for the
DW_TAG_unspecified_type die.
mav [Wed, 12 Jun 2013 18:08:11 +0000 (18:08 +0000)]
Replicate r242422 from ata(4) to mvs(4):
Only four specific ATA PIO commands transfer several sectors per DRQ block
(interrupt). All other ATA PIO commands transfer one sector or 512 bytes
at one time. Hardcode these exceptions in mvs(4) with ATA_CAM option.
This fixes timeout of READ LOG EXT command used by `smartctl -x /dev/adaX`.
Also it fixes timeout of DOWNLOAD_MICROCODE on `camcontrol fwdownload`.
adrian [Wed, 12 Jun 2013 14:52:57 +0000 (14:52 +0000)]
Migrate the LNA mixing diversity machinery from the AR9285 HAL to the driver.
The AR9485 chip and AR933x SoC both implement LNA diversity.
There are a few extra things that need to happen before this can be
flipped on for those chips (mostly to do with setting up the different
bias values and LNA1/LNA2 RSSI differences) but the first stage is
putting this code into the driver layer so it can be reused.
This has the added benefit of making it easier to expose configuration
options and diagnostic information via the ioctl API. That's not yet
being done but it sure would be nice to do so.
Tested:
* AR9285, with LNA diversity enabled
* AR9285, with LNA diversity disabled in EEPROM
mav [Wed, 12 Jun 2013 13:36:20 +0000 (13:36 +0000)]
Make CAM return and GEOM DISK pass through new GEOM::lunid attribute.
SPC-4 specification states that serial number may be property of device,
but not a specific logical unit. People reported about FC storages using
serial number in that way, making it unusable for purposes of LUN multipath
detection. SPC-4 states that designators associated with logical unit from
the VPD page 83h "Device Identification" should be used for that purpose.
Report first of them in the new attribute in such preference order: NAA,
EUI-64, T10 and SCSI name string.
While there, make GEOM DISK properly report GEOM::ident in XML output also
using d_getattr() method, if available. This fixes serial numbers reporting
for SCSI disks in `geom disk list` output and confxml.
Discussed with: gibbs, ken
Sponsored by: iXsystems, Inc.
MFC after: 2 weeks
gjb [Wed, 12 Jun 2013 13:15:28 +0000 (13:15 +0000)]
Add new script to run release builds. The script is heavily based upon
the release/generate-release.sh script by nwhitehorn.
This script can use optional configuration file to override defaults,
making running multiple serialzied builds with different specific
configurations or architecture-specific tunings quite easy.
Sample overrideable options are included in the release.conf.sample
file.
mav [Wed, 12 Jun 2013 12:51:43 +0000 (12:51 +0000)]
Use direct custom implementations instead of g_handleattr() for CFI and NAND
d_getattr(). Since these drivers use disk(9) KPI and not directly GEOM, use
of that function means KPI layering violation, causing extra g_io_deliver()
call for the request.
mav [Wed, 12 Jun 2013 09:07:15 +0000 (09:07 +0000)]
Acquire periph reference when handling d_getattr() method call.
While GEOM in general has provider opened while sending BIO_GETATTR,
GEOM DISK does not really need to open disk to read medium-unrelated
attributes for own use.
grog [Wed, 12 Jun 2013 07:52:49 +0000 (07:52 +0000)]
Handle some expression regressions.
Explicitly use GNU cpp for preprocessing.
Remove explicit debugging code.
Change some variable names to be less confusing.
Improve some comments.
Improve indentation.
jhb [Tue, 11 Jun 2013 20:00:49 +0000 (20:00 +0000)]
Borrow the algorithm from kvm_getprocs() to fix procstat_getprocs() to
handle the case where the process tables grows in between the calls to
fetch the size and fetch the table.
jhb [Tue, 11 Jun 2013 19:05:29 +0000 (19:05 +0000)]
Bump the CPU/WCPU column width by one so that it fits values from 100% up
to 999.99% CPU. It still won't be aligned if you have a multithreaded
process using more than 1000% CPU (e.g. idle process on an idle 12-way
system), but 100% is a common case.
Submitted by: Jeremy Chadwick (partial)
MFC after: 1 week
flo [Tue, 11 Jun 2013 17:46:32 +0000 (17:46 +0000)]
Move the check whether the clear flag is set. This has 2 advantages
- When operating on a core file (-M) and -c is specified we don't clear
the message buffer of the running system.
- If we don't have permission to clear the buffer print the error message
only. That's what Linux does in this case, where this feature was ported
from, and it ensures that the error message doesn't get lost in the noise.
jhb [Tue, 11 Jun 2013 15:37:07 +0000 (15:37 +0000)]
Store a reference to the vnode associated with a file descriptor in the
linux_file structure and use it instead of directly accessing td_fpop
when destroying the linux_file structure. The td_fpop pointer is not
valid when a cdevpriv destructor is run, and the type-specific close
method has already been called, so f_vnode may not be valid (and the
vnode might have been recycled without our own reference).
mav [Tue, 11 Jun 2013 10:06:07 +0000 (10:06 +0000)]
Don't update provider properties and don't set DISKFLAG_OPEN if d_open()
disk method call returned error. GEOM considers devices in such case as
still closed, and won't call symmetric d_close() for them.
davidcs [Mon, 10 Jun 2013 17:12:22 +0000 (17:12 +0000)]
There is a one-to-one correspondence between the MSI-X vector # and the Status Descriptor Ring Index. Hence there is no need to check the Interrupt Source Register.
marcel [Mon, 10 Jun 2013 15:22:27 +0000 (15:22 +0000)]
Revert r251590. It unexpectedly broke the build and there were some
questions on locking. As part of commit-bit grooming, I'd like Steve
to handle this, but can't leave things broken in the mean time.
emaste [Mon, 10 Jun 2013 13:55:03 +0000 (13:55 +0000)]
Handle options that override other options
This is a workaround for WITH_LDNS_UTILS forcing BIND_UTILS off. It can
be reverted when we no longer have these conflicting options, or made more
general if we grow more cases like this.
yongari [Mon, 10 Jun 2013 07:31:49 +0000 (07:31 +0000)]
Avoid unnecessary controller reinitialization by checking driver
running state. fxp(4) requires controller reinitialization for the
following cases.
o RX lockup condition on i82557
o promiscuous mode change
o multicast filter change
o WOL configuration
o TSO/VLAN hardware tagging/checksum offloading configuration
o MAC reprogramming after speed/duplex/flow-control resolution
o Any events that result in MAC reprogramming(link UP/DOWN,
remote link partner's restart of auto-negotiation etc)
o Microcode loading/unloading
Apart from above cases which come from hardware limitation, upper
stack also blindly reinitializes controller whenever an IP address
is assigned. After r194573, fxp(4) no longer needs to reinitialize
the controller to program multicast filter after upping the
interface. So keeping track of driver running state should remove
all unnecessary controller reinitializations.
This change will also address endless controller reinitialization
triggered by dhclient(8).
Tested by: hrs, Alban Hertroys <haramrae@gmail.com>
alc [Mon, 10 Jun 2013 01:48:21 +0000 (01:48 +0000)]
Revise the interface between vm_object_madvise() and vm_page_dontneed() so
that pointless calls to pmap_is_modified() can be easily avoided when
performing madvise(..., MADV_FREE).
marcel [Sun, 9 Jun 2013 23:51:26 +0000 (23:51 +0000)]
Add vfs_mounted and vfs_unmounted events so that components can be informed
about mount and unmount events. This is used by Juniper to implement a more
optimal implementation of NetBSD's veriexec.
marius [Sun, 9 Jun 2013 23:50:30 +0000 (23:50 +0000)]
- The method introduced as part of r234898 for not altering the boot path
when booting from ZFS turned out to also cause the boot path not being
adjusted if booting from CD-ROM with firmware versions that do not employ
the "cdrom" alias in that case. So shuffle the code around instead in order
to achieve the original intent. Ideally, we shouldn't fiddle with the boot
path when booting from UFS on a disk either; unfortunately, there doesn't
seem to be an universal way of telling disks and CD-ROMs apart, though. [1]
- Use NULL instead of 0 for pointers.
marcel [Sun, 9 Jun 2013 23:34:26 +0000 (23:34 +0000)]
Change the set and unset ctlreqs by making the index argument optional.
This allows setting attributes on tables. One simply does not provide
an index in that case. Otherwise the entry corresponding the index has
the attribute set or unset.
Use this change to fix a relatively longstanding bug in our GPT scheme
that's the result of rev 198097 (relatively harmless) followed by rev
237057 (damaging). The damaging part being that our GPT scheme always
has the active flag set on the PMBR slice. This is in violation with
EFI. Existing EFI implementions for both x86 and ia64 reject the GPT.
As such, GPT disks created by us aren't usable under EFI because of
that.
After this change, GPT disks never have the active flag set on the PMBR
slice. In order to make the GPT disk bootable under some x86 BIOSes,
the reason of rev 198097, one must now set the active attribute on the
gpt table. The kernel will apply this to the PMBR slice For (S)ATA:
gpart set -a active ada0
To fix an existing GPT disk that has the active flag set in the PMBR,
and that does not need the flag, use (again for (S)ATA):
gpart unset -a active ada0
The EBR, MBR & PC98 schemes, which also impement at least 1 attribute,
now check to make sure the entry passed is valid. They do not have
attributes that apply to the table.
At boot time, all of the static routes are installed as before.
The differences are:
- "/etc/rc.d/netif start/stop <if>" now configures static routes
with :<if> if any.
- "/etc/rc.d/routing start/stop <af> <if>" works as well. <af> cannot be
omitted when <if> is specified, but a keyword "any" or "all" can be used
for <af> and <if>.
hrs [Sun, 9 Jun 2013 16:33:32 +0000 (16:33 +0000)]
- Add file system type number (vfc_typenum) in the list. This is useful
for debugging when changing vfs.typenumhash configuration.
- Refactor fmt_flags().
pfg [Sun, 9 Jun 2013 01:38:51 +0000 (01:38 +0000)]
e2fs_bpg and e2fs_isize are always unsigned.
The superblock in ext2fs defines all the fields as unsigned but for
some reason the in-memory superblock was carrying e2fs_bpg and
e2fs_isize as signed.
We should preserve the specified types for consistency.
pfg [Sun, 9 Jun 2013 01:19:22 +0000 (01:19 +0000)]
libstand: Reset the seek pointer in ext2fs as done in UFS.
Based on r134760:
Reset the seek pointer to 0 when a file is successfully opened,
since otherwise the initial seek offset will contain the directory
offset of the filesystem block that contained its directory entry.
This bug was mostly harmless because typically the directory is
less than one filesystem block in size so the offset would be zero.
It did however generally break loading a kernel from the (large)
kernel compile directory.
Also reset the seek pointer when a new inode is opened in read_inode(),
though this is not actually necessary now because all callers set
it afterwards.
PR: 177328
Submitted by: Eric van Gyzen
Reviewed by: iedowse
MFC after: 5 days
Merge the 1 and 2 byte versions of the atomic functions into one.
After pushing in my fix for the 2 byte functions, I realized that the
functions for 1 and 2 byte operations had become identical. Reduce the
code size by merging the functions for 1 and 2 byte operations together.
While there, slightly improve variable naming and comments.
This small utility performs a sequence of atomic operations with random
parameters on an atomic variable. For every type, we also create 16
variables, to ensure that we test the correctness at different
alignments.
dteske [Sat, 8 Jun 2013 21:45:32 +0000 (21:45 +0000)]
Fix a bug introduced with r251190, in which a small hunk was forgotten in
the performance conversion process.
The effect of this was the following error when selecting the menu
"Startup"->"View/Edit Startup Configuration"->"Add New"->"Add From List":
[: -eq: unexpected operator
By running `bsdconfig -d' as root to enable debugging, this turns into:
DEBUG: f_getvar: var=[text] value=[
Error: Expected a number for token 4 of --menu.
Use --help to list options.] r=0
[: -eq: unexpected operator
Indicating that the fourth token for --menu which should be $height was
instead a string (the first item of $menu_list) because it was using the
old size-calculation method and $size was now null (needed to use instead
the new size variables of $height $width and $rows).
dteske [Sat, 8 Jun 2013 21:33:09 +0000 (21:33 +0000)]
Fix a bug introduced with r249751, in which a small hunk was forgotten in
the performance conversion process.
The effect of this was, when your /etc/defaults/rc.conf file changed
(based on md5(1)) and re-generating the file startup_rcconf_map.cache in
/var/run/bsdconfig/ you would get a screen-dump of its contents before the
menu would appear.
dteske [Sat, 8 Jun 2013 20:27:16 +0000 (20:27 +0000)]
Fix a one-character typo. The effects of which were "Console"->"Ttys" menu
would not launch (and gave no error). This was easily diagnosed by running
`bsdconfig -d' as root and seeing the following error right after selecting
the "Ttys" sub-menu:
DEBUG: f_getvar: var=[text] value=[
Error: Expected at least 6 tokens for --menu, have 4.
Use --help to list options.] r=0
adrian [Sat, 8 Jun 2013 20:21:17 +0000 (20:21 +0000)]
Add 8devices CARAMBOLA2 support.
This is based on the AR933x (Hornet) SoC from Qualcomm Atheros.
It's a much nicer board to do development on - 64MB RAM, 16MB flash.
The development board breaks out the GPIO pins, ethernet, serial (via
a USB<->RS232 chip), USB host and of course a small wifi antenna.