jmallett [Tue, 6 Mar 2012 08:02:10 +0000 (08:02 +0000)]
Store TLS base in the sigframe just as is done in freebsd32_sendsig. Because
the native sigreturn doesn't use set_mcontext like the COMPAT_FREEBSD32 version
does, this wouldn't actually result in overwriting the TLS base. Probably it
makes sense to restructure the native sigreturn to use set_mcontext for
consistency, and to allow sigreturn to change the TLS base.
jmallett [Tue, 6 Mar 2012 07:50:45 +0000 (07:50 +0000)]
Fix two and a half oversights in COMPAT_FREEBSD32 related to contexts and
TLS:
o) The mc_tls field used to store the TLS base when doing context gets and
restores was left a pointer and not converted to a 32-bit integer. This
had the bug of not correctly capturing the TLS value desired by the user,
and the extra nastiness of making the structure the wrong size.
o) The mc_tls field was not being saved by sendsig. As a result, the TLS base
would always be set to NULL when restoring from a signal handler.
Thanks to gonzo for helping track down a bunch of other TLS bugs that came out
of tracking these down.
gonzo [Tue, 6 Mar 2012 03:42:54 +0000 (03:42 +0000)]
- Switch ARM and MIPS to TLS Variant I
- Fix TLS allocation for Variant I: both rtld and libc allocators
assume that tls_static_space includes space for TLS structure.
So increment calculated static size by the size of it.
gonzo [Tue, 6 Mar 2012 03:25:50 +0000 (03:25 +0000)]
Prepare for large TLS redo. Save pointer to the beginning of TLS area,
and offset it only if requested by RDHWR handler. Otherwise things
get overly complicated - we need to track whether address passsed in
request for setting td_md.md_tls is already offseted or not.
jmallett [Tue, 6 Mar 2012 02:23:15 +0000 (02:23 +0000)]
In the trap messages that aid the primitive debugging environment of MIPS,
include the tid as well, so it's easier to tell which thread of a process
with multiple is responsible for a crash.
kib [Mon, 5 Mar 2012 20:43:06 +0000 (20:43 +0000)]
The libmap.conf initialization is performed before TLS is functional.
Since after r232498 the ctype macros require working access to
thread-local variables, rtld crashes when libmap.conf is present.
Use hand-made isspace1() macro which is enough to detect spaces in
libmap.conf.
Reported by: alc, lme, many on current@
Tested by: lme
Reviewed by: dim, kan
MFC after: 1 week
jhb [Mon, 5 Mar 2012 19:53:17 +0000 (19:53 +0000)]
Fix boot2 to handle boot config files that only contain a custom path to
a loader or kernel. Specifically, kname cannot be pointed at cmd[] since
it's value is change to be an empty string after the initial call to
parse, and cmd[]'s value can be changed (thus losing a prior setting for
kname) due to user input at the boot prompt. While here, ensure that that
initial boot config file text is nul-terminated, that ops is initialized
to zero, and that kname is always initialized to a valid string.
ivoras [Mon, 5 Mar 2012 14:19:43 +0000 (14:19 +0000)]
Print out process name and thread id in the debugging message.
This is useful because the message can end up in system logs in
non-debugging operation.
dim [Sun, 4 Mar 2012 21:36:18 +0000 (21:36 +0000)]
Fix a thinko in r232322, where gcc (and its tools) are not built during
the cross-tools stage, if CC=clang and WITH_CLANG_IS_CC is not set.
This causes no 'cc' to be installed in the temporary cross-tools tree,
making lint fall over later in the build, because it ignores ${CC} and
attempts to run 'cc' anyway.
To fix this, only skip building gcc during cross-tools, if WITHOUT_GCC
is set, or if WITH_CLANG_IS_CC is set.
zec [Sun, 4 Mar 2012 18:59:38 +0000 (18:59 +0000)]
Change SYSINIT priorities so that ip_mroute_modevent() is executed
before vnet_mroute_init(), since vnet_mroute_init() depends on mfchashsize
tunable to be set, and that is done in in ip_mroute_modevent().
Apparently I broke that ordering with r208744 almost 2 years ago...
Rather than printing the output from route add for all FIBs just print them
for the default FIB followed by a statement with a list of FIB numbers for
all the other FIBs we install the routes for.
Request by: kib (to make it less noisy)
Tested by: kib
MFC after: 3 days
In nd6_options() ignore the RFC 6106 options completely rather than printing
them if nd6_debug is enabled as unknown. Leave a comment about the RFC4191
option as I am undecided so far.
Correct typo in the RFC number for the constants based on IANA assignments
for IPv6 Neighbor Discovery Option types for "IPv6 Router Advertisement
Options for DNS Configuration". It is RFC 6106.
kib [Sun, 4 Mar 2012 14:55:37 +0000 (14:55 +0000)]
Instead of incomplete handling of read(2)/write(2) return values that
does not fit into registers, declare that we do not support this case
using CTASSERT(), and remove endianess-unsafe code to split return value
into td_retval.
While there, change the style of the sysctl debug.iosize_max_clamp
definition.
andreast [Sun, 4 Mar 2012 11:55:28 +0000 (11:55 +0000)]
Restore proper dot symbol creation for assembly files in the kernel build case.
Without this patch we were not able to see the assembly function.
Only the function descriptor was visible.
- Distinguish between user-land and kernel when creating the ENTRY() point of
assembly source.
- Make the ENTRY() macro more readable, replace the .align directive with the
gas platform independant .p2align directive.
- Create an END()macro for later use to provide traceback tables on powerpc64.
andreast [Sun, 4 Mar 2012 08:43:33 +0000 (08:43 +0000)]
Add support for PWM controlled fans. I found these fans on my PowerMac9,1.
These fans are not located under the same node as the the RPM controlled ones,
So I had to adapt the current source to parse and fill the properties correctly.
To control the fans we can set the PWM ratio via sysctl between 20 and 100%.
adrian [Sun, 4 Mar 2012 05:52:26 +0000 (05:52 +0000)]
* Introduce new flag for QoS control field;
* Change in mesh_input to validate that QoS is set and Mesh Control field
is present, also both bytes of the QoS are read;
* Moved defragmentation in mesh_input before we try to forward packet as
inferred from amendment spec, because Mesh Control field only present in first
fragment;
* Changed in ieee80211_encap to set QoS subtype and Mesh Control field present,
only first fragment have Mesh Control field present bit equal to 1;
adrian [Sun, 4 Mar 2012 05:49:39 +0000 (05:49 +0000)]
* Added IEEE80211_ACTION_CAT_MESH in ieee80211.h as specified amendment spec;
* Moved old categories as specified by D4.0 to be action fields of MESH category
as specified in amendment spec;
* Modified functions to use MESH category and its action fields:
+ ieee80211_send_action_register
+ ieee80211_send_action
+ ieee80211_recv_action_register
+ieee80211_recv_action;
* Modified ieee80211_hwmp_init and hwmp_send_action so they uses correct
action fields as specified in amendment spec;
* Modified ieee80211_parse_action so that it verifies MESH frames.
* Change Mesh Link Metric to use one information element as amendment spec.
Draft 4.0 defined two different information elements for request and response.
jmallett [Sun, 4 Mar 2012 05:19:55 +0000 (05:19 +0000)]
Fix tls base computation with COMPAT_FREEBSD32 on n64 kernels. The previous
version was missing an else and would always use the n64 TP_OFFSET. Eliminate
some duplication of logic here.
It may be worth getting rid of some of the ifdefs and introducing gratuitous
SV_ILP32 runtime checks on n64 kernels without COMPAT_FREEBSD32 and on o32
kernels, similarly to how PowerPC works.
dim [Sat, 3 Mar 2012 23:49:53 +0000 (23:49 +0000)]
Revert r232473. I have been convinced by Doug Barton and Bjoern Zeeb
that it is better to error out when people attempt to build using the
wrong bsd.*.mk files, than to silently ignore the problem.
This means, that after this commit, if you want to build kernel modules
by hand (or via a port) from a head source tree, you *must* make sure
the files in /usr/share/mk are in sync with that tree. If that isn't
possible, for example when you are running on an older FreeBSD branch,
you can:
- Run "make buildenv" from your head source tree, to have the correct
environment setup. (It's advisable to have run "make buildworld", or
at a minimum "make toolchain" first.)
- Alternatively, set MAKESYSPATH to the share/mk directory under your
head source tree. If your build tools are too old, other problems may
still occur.
- Alternatively, use "make -m" and specify the share/mk directory under
your head source tree. Again, build tools that are too old may still
result in trouble.
jmallett [Sat, 3 Mar 2012 21:39:12 +0000 (21:39 +0000)]
On MIPS, _ALIGN always aligns to 8 bytes, even for 32-bit binaries. This might
not be ideal, but is the ABI we've shipped so far. Fix macros which reflect
the results of _ALIGN on 32-bit MIPS to use the right alignment.
This fixes sendmsg under COMPAT_FREEBSD32 on n64 MIPS kernels.
dim [Sat, 3 Mar 2012 18:58:15 +0000 (18:58 +0000)]
After r232322, it turned out many people (and some ports) are building
kernel modules using their old installed /usr/share/mk/bsd.*.mk files,
instead of the updated ones in their source tree. This leads to errors
like:
Obviously, these errors will go away after a "make installworld", or
alternatively, by using "make buildenv" before attempting to manually
build modules.
However, since it is apparently an expected use case to build using old
.mk files, change the way we test for clang, so it also works when the
MK_CLANG_IS_CC macro doesn't exist.
Note the conditional expressions are becoming rather unreadable now, but
I will attempt to fix that on a followup commit.
jhb [Sat, 3 Mar 2012 18:08:57 +0000 (18:08 +0000)]
Expand the set of APIs available for locating PCI capabilities:
- pci_find_extcap() is repurposed to be used for fetching PCI-express
extended capabilities (PCIZ_* constants in <dev/pci/pcireg.h>).
- pci_find_htcap() can be used to locate a specific HyperTransport
capability (PCIM_HTCAP_* constants in <dev/pci/pcireg.h>).
- Cache the starting location of the PCI-express capability for PCI-express
devices in PCI device ivars.
rmacklem [Sat, 3 Mar 2012 16:13:20 +0000 (16:13 +0000)]
The name caching changes of r230394 exposed an intermittent bug
in the new NFS server for NFSv4, where it would report ENOENT
when the file actually existed on the server. This turned out
to be caused by not initializing ni_topdir before calling lookup()
and there was a rare case where the value on the stack location
assigned to ni_topdir happened to be a pointer to a ".." entry,
such that "dp == ndp->ni_topdir" succeeded in lookup().
This patch initializes ni_topdir to fix the problem.
jhb [Sat, 3 Mar 2012 14:25:36 +0000 (14:25 +0000)]
Update the pci_get_vpd_readonly() wrapper to use 'vptr' instead of
'identptr' for its last parameter to match the default implementation
as well as the method definition in pci_if.m.
jhb [Sat, 3 Mar 2012 14:23:54 +0000 (14:23 +0000)]
Expand and reorganize the pci(9) manpage a bit:
- Document the following routines: pci_alloc_msi(), pci_alloc_msix(),
pci_find_cap(), pci_get_max_read_req(), pci_get_vpd_ident(),
pci_get_vpd_readonly(), pci_msi_count(), pci_msix_count(),
pci_pending_msix(), pci_release_msi(), pci_remap_msix(), and
pci_set_max_read_req().
- Group the functions into five sub-sections: raw configuration access,
locating devices, device information, device configuration, and
message signaled interrupts.
- Discourage use of pci_disable_io() and pci_enable_io() in device drivers.
The PCI bus driver handles this automatically as resources are activated.
mav [Sat, 3 Mar 2012 11:50:48 +0000 (11:50 +0000)]
Fix bug of r232207, when cpu_search() could prefer CPU group with best
load, but with no CPU matching given limitations. It caused kernel panics
in some cases when thread was bound to specific CPUs with cpuset(1).
jmallett [Sat, 3 Mar 2012 08:19:18 +0000 (08:19 +0000)]
o) Add COMPAT_FREEBSD32 support for MIPS kernels using the n64 ABI with userlands
using the o32 ABI. This mostly follows nwhitehorn's lead in implementing
COMPAT_FREEBSD32 on powerpc64.
o) Add a new type to the freebsd32 compat layer, time32_t, which is time_t in the
32-bit ABI being used. Since the MIPS port is relatively-new, even the 32-bit
ABIs use a 64-bit time_t.
o) Because time{spec,val}32 has the same size and layout as time{spec,val} on MIPS
with 32-bit compatibility, then, disable some code which assumes otherwise
wrongly when built for MIPS. A more general macro to check in this case would
seem like a good idea eventually. If someone adds support for using n32
userland with n64 kernels on MIPS, then they will have to add a variety of
flags related to each piece of the ABI that can vary. That's probably the
right time to generalize further.
o) Add MIPS to the list of architectures which use PAD64_REQUIRED in the
freebsd32 compat code. Probably this should be generalized at some point.
hselasky [Sat, 3 Mar 2012 08:11:04 +0000 (08:11 +0000)]
Make sure that the USB system suspend event is executed synchronously
and not asynchronously. This fixes problems related to USB system
suspend and resume. It is assumed that we are always allowed to sleep
from the device_suspend() method.
rmacklem [Sat, 3 Mar 2012 01:06:54 +0000 (01:06 +0000)]
Post r230394, the Lookup RPC counts for both NFS clients increased
significantly. Upon investigation this was caused by name cache
misses for lookups of "..". For name cache entries for non-".."
directories, the cache entry serves double duty. It maps both the
named directory plus ".." for the parent of the directory. As such,
two ctime values (one for each of the directory and its parent) need
to be saved in the name cache entry.
This patch adds an entry for ctime of the parent directory to the
name cache. It also adds an additional uma zone for large entries
with this time value, in order to minimize memory wastage.
As well, it fixes a couple of cases where the mtime of the parent
directory was being saved instead of ctime for positive name cache
entries. With this patch, Lookup RPC counts return to values similar
to pre-r230394 kernels.
ken [Fri, 2 Mar 2012 22:00:40 +0000 (22:00 +0000)]
Fix a problem that was causing the mpt(4) driver to attach to MegaRAID
cards that should be handled by the mfi(4) driver.
The root of the problem is that the mpt(4) driver was masking off the
bottom bit of the PCI device ID when deciding which cards to attach to.
It appears that a number of the mpt(4) Fibre Channel cards had a LAN
variant whose PCI device ID was just one bit off from the FC card's device
ID. The FC cards were even and the LAN cards were odd.
The problem was that this pattern wasn't carried over on the SAS and
parallel SCSI mpt(4) cards. Luckily the SAS and parallel SCSI PCI device
IDs were either even numbers, or they would get masked to a supported
adjacent PCI device ID, and everything worked well.
Now LSI is using some of the odd-numbered PCI device IDs between the 3Gb
SAS device IDs for their new MegaRAID cards. This is causing the mpt(4)
driver to attach to the RAID cards instead of the mfi(4) driver.
The solution is to stop masking off the bottom bit of the device ID, and
explicitly list the PCI device IDs of all supported cards.
This change should be a no-op for mpt(4) hardware. The only intended
functional change is that for the 929X, the is_fc variable gets set. It
wasn't being set previously, but needs to be because the 929X is a Fibre
Channel card.
Reported by: Kashyap Desai <Kashyap.Desai@lsi.com>
MFC After: 3 days
jmallett [Fri, 2 Mar 2012 21:46:31 +0000 (21:46 +0000)]
When creating a handle for a subregion, be sure to actually math out the new
handle address, where we're using handles as raw addresses.
This fixes devices with subregions on Octeon PCI specifically, and likely also on
MIPS more generally, where there isn't another bus_space in use that was doing the
math already.
jhb [Fri, 2 Mar 2012 20:38:04 +0000 (20:38 +0000)]
- Add a bus_dma tag to each PCI bus that is a child of a Host-PCI bridge.
The tag enforces a single restriction that all DMA transactions must not
cross a 4GB boundary. Note that while this restriction technically only
applies to PCI-express, this change applies it to all PCI devices as it
is simpler to implement that way and errs on the side of caution.
- Add a softc structure for PCI bus devices to hold the bus_dma tag and
a new pci_attach_common() routine that performs actions common to the
attach phase of all PCI bus drivers. Right now this only consists of
a bootverbose printf and the allocate of a bus_dma tag if necessary.
- Adjust all PCI bus drivers to allocate a PCI bus softc and to call
pci_attach_common() from their attach routines.
jmallett [Fri, 2 Mar 2012 20:34:15 +0000 (20:34 +0000)]
Unbreak SMP on stock Octeon systems -- copy the core_mask from bootinfo into
sysinfo. This should have been done as part of replacing bootinfo with sysinfo.
jhb [Fri, 2 Mar 2012 18:55:19 +0000 (18:55 +0000)]
Similar to the fixes in 226967 and 226987, purge any name cache entries
associated with the previous vnode (if any) associated with the target of
a rename(). Otherwise, a lookup of the target pathname concurrent with a
rename() could re-add a name cache entry after the namei(RENAME) lookup
in kern_renameat() had purged the target pathname.