dchagin [Sun, 24 May 2015 16:18:03 +0000 (16:18 +0000)]
Change linux faccessat syscall definition to match actual linux one.
The AT_EACCESS and AT_SYMLINK_NOFOLLOW flags are actually implemented
within the glibc wrapper function for faccessat(). If either of these
flags are specified, then the wrapper function employs fstatat() to
determine access permissions.
dchagin [Sun, 24 May 2015 15:54:58 +0000 (15:54 +0000)]
Refund the proc emuldata struct for future use. For now move flags from
thread emuldata to proc emuldata as it was originally intended.
As we can have both 64 & 32 bit Linuxulator running any eventhandler
can be called twice for us. To prevent this move eventhandlers code
from linux_emul.c to the linux_common.ko module.
dchagin [Sun, 24 May 2015 15:51:18 +0000 (15:51 +0000)]
Introduce a new module linux_common.ko which is intended for the
following primary purposes:
1. Remove the dependency of linsysfs and linprocfs modules from linux.ko,
which will be architecture specific on amd64.
2. Incorporate into linux_common.ko general code for platforms on which
we'll support two Linuxulator modules (for both instruction set - 32 & 64 bit).
3. Move malloc(9) declaration to linux_common.ko, to enable getting memory
usage statistics properly.
Currently linux_common.ko incorporates a code from linux_mib.c and linux_util.c
and linprocfs, linsysfs and linux kernel modules depend on linux_common.ko.
Temporarily remove dtrace garbage from linux_mib.c and linux_util.c
Differential Revision: https://reviews.freebsd.org/D1072
In collaboration with: Vassilis Laganakos.
tijl [Sun, 24 May 2015 15:47:06 +0000 (15:47 +0000)]
When only 2 bytes can be read from a 4 byte UTF-16 character in a base64
encoded chunk of a UTF-7 string, treat that as an incomplete character and
return an error instead of a shift sequence and no error.
Also check that the low 2 bytes have a valid value.
dchagin [Sun, 24 May 2015 15:28:17 +0000 (15:28 +0000)]
Implement vdso - virtual dynamic shared object. Through vdso Linux
exposes functions from kernel with proper DWARF CFI information so that
it becomes easier to unwind through them.
Using vdso is a mandatory for a thread cancelation && cleanup
on a modern glibc.
tijl [Sun, 24 May 2015 15:27:31 +0000 (15:27 +0000)]
Fix decoding of UTF-7 when a base64 encoded chunk appears at the end of
the input buffer.
_citrus_UTF7_mbtoutf16 stored the decoder state at the beginning so it
could restore this state on an incomplete character such that the next
call would restart the decoding. The problem was that "-" (end of base64
mode) at the end of a string was also treated as an incomplete character
but was also removed from the state buffer. So the initial state would be
restored (with base64 mode) and the next call would no longer see the "-"
so it continued in base64 mode.
This state saving/restoring isn't needed here. It's already handled
elsewhere (citrus_iconv_std.c:_citrus_iconv_std_iconv_convert) so just
remove it.
dchagin [Sun, 24 May 2015 15:03:09 +0000 (15:03 +0000)]
To reduce code duplication introduce linux_copyout_rusage() method.
Use it in linux_wait4() system call and move linux_wait4() to the MI path.
While here add a prototype for the static bsd_to_linux_rusage().
dchagin [Sun, 24 May 2015 14:54:12 +0000 (14:54 +0000)]
pthread_join() caller do futex_wait on child_clear_tid. As a results
of multiple simultaneous calls to pthread_join() specifying the same
target thread are undefined wake up the one thread.
dchagin [Sun, 24 May 2015 14:53:16 +0000 (14:53 +0000)]
Switch linuxulator to use the native 1:1 threads.
The reasons:
1. Get rid of the stubs/quirks with process dethreading,
process reparent when the process group leader exits and close
to this problems on wait(), waitpid(), etc.
2. Reuse our kernel code instead of writing excessive thread
managment routines in Linuxulator.
Implementation details:
1. The thread is created via kern_thr_new() in the clone() call with
the CLONE_THREAD parameter. Thus, everything else is a process.
2. The test that the process has a threads is done via P_HADTHREADS
bit p_flag of struct proc.
3. Per thread emulator state data structure is now located in the
struct thread and freed in the thread_dtor() hook.
Mandatory holdig of the p_mtx required when referencing emuldata
from the other threads.
4. PID mangling has changed. Now Linux pid is the native tid
and Linux tgid is the native pid, with the exception of the first
thread in the process where tid and pid are one and the same.
Ugliness:
In case when the Linux thread is the initial thread in the thread
group thread id is equal to the process id. Glibc depends on this
magic (assert in pthread_getattr_np.c). So for system calls that
take thread id as a parameter we should use the special method
to reference struct thread.
dchagin [Sun, 24 May 2015 14:44:06 +0000 (14:44 +0000)]
In preparation for switching linuxulator to the use the native 1:1
threads split sys_sched_getparam(), sys_sched_setparam(),
sys_sched_getscheduler(), sys_sched_setscheduler() to their kern_*
counterparts and add targettd parameter to allow specify the target
thread directly by callee.
dchagin [Sun, 24 May 2015 14:39:26 +0000 (14:39 +0000)]
In preparation for switching linuxulator to the use the native 1:1
threads refactor kern_sched_rr_get_interval() and sys_sched_rr_get_interval().
Add a kern_sched_rr_get_interval() counterpart which takes a targettd
parameter to allow specify target thread directly by callee (new Linuxulator).
Linuxulator temporarily uses first thread in proc.
Move linux_sched_rr_get_interval() to the MI part.
dchagin [Sun, 24 May 2015 14:36:33 +0000 (14:36 +0000)]
In preparation for switching linuxulator to the use the native 1:1
threads split sys_thr_exit() up into sys_thr_exit() and kern_thr_exit().
Move
Where the second will be used in linux_exit() system call later.
dchagin [Sun, 24 May 2015 14:33:19 +0000 (14:33 +0000)]
In preparation for switching linuxulator to the use the native 1:1
threads introduce linux_exit() stub instead of sys_exit() call
(which terminates process).
In the new linuxulator exit() system call terminates the calling
thread (not a whole process).
melifaro [Sun, 24 May 2015 11:24:14 +0000 (11:24 +0000)]
Fix SIOCGI2C structure requirement.
In reality, SIOCGI2C ioctl requires struct ifreq as many other
ioctls. Doing copyin() on (significantly) larger struct ifstat sometimes
triggered EFAULT.
andrew [Sun, 24 May 2015 11:08:06 +0000 (11:08 +0000)]
Rework the PSCI cpu on code to allow it to work before device drivers have
started. This allows this functions to be used with the regular ARM SMP
initialisation sequence.
ganbold [Sun, 24 May 2015 08:45:19 +0000 (08:45 +0000)]
Add kernel config and dts files for an aml8726-m3
based device.
Following u-boot commands allow FreeBSD boot on
Yiyate Android TV Box (aml8726-m3):
tv open 480p
mmc rescan 0
fatload mmc 0 0x80100000 kernel.bin
go 0x80100000
The current FreeBSD driver doesn't program the video
clocks so the u-boot tv command is necessary in order
for the frame buffer to be useful (otherwise it can
be skipped).
The SD card for the Yiyate Android TV Box doesn't need
anything special beyond creating a FAT16 and a UFS
filesystem.
Differential Revision: https://reviews.freebsd.org/D2636
Submitted by: John Wehle
Without this booting the VSATV102 causes the blue "working" led to turn
off when the kernel starts up. With this the led (which is turned on by
the firmware) stays on since that's the default state specified in the FDT.
Expanded the meaning of the led_create_state state parameter in order
to implement support for "keep". The original values were:
== 0 Off
!= 0 On
The new values are:
== -1 don't change / keep current setting
== 0 Off
!= -1 && != 0 On
This should have no effect on acpi_asus_attach which only calls
led_create_state with state set to 1. Updated acpi_ibm_attach
in order to avoid surprises.
andrew [Sat, 23 May 2015 23:08:16 +0000 (23:08 +0000)]
Fix a lock up where we enter swapper() with interrupts disabled.
In smp_rendezvous_cpus we expect to wait for all cpus to enter
smp_rendezvous_action. If we call this holding a proc lock swapper may
attempt to also lock it, however as interrupts are disabled the cpu never
handles the ipi. Because smp_rendezvous_action waits for all signaled
cpus before contining it may get caught waiting for the cpu running swapper
as the proc mutex will be unlocked after smp_rendezvous_cpus finishes.
The fix is to enable interrupts in the configure stage as we should be
doing.
andrew [Sat, 23 May 2015 22:28:59 +0000 (22:28 +0000)]
Use the wait-for-event instruction to put the core we have just enabled
to sleep while it waits to start scheduling. The boot core can then use
the send-event instruction to wake the cores when they should enter the
scheduler.
rmacklem [Sat, 23 May 2015 21:58:41 +0000 (21:58 +0000)]
The NFS client generated directory block(s) with d_fileno == 0
so that it would not return less data than requested.
Since returning less directory data than requested is not a problem
for FreeBSD and even UFS no longer returns directory structures
with d_fileno == 0, this patch stops the client from doing this.
Although entries with d_fileno == 0 should not be a problem,
the man pages no longer document that these entries should be
ignored, so there was a concern that these entries might be an
issue in the future.
kib [Sat, 23 May 2015 19:09:04 +0000 (19:09 +0000)]
If thread requested to not stop on non-boundary, then not only
stopping signals should obey, but also all forms of single-threading.
Otherwise, thread might sleep interruptible while owning some
resources, and single-threading thread could try to access them.
An example is owning vnode lock while dumping core.
alc [Sat, 23 May 2015 16:54:46 +0000 (16:54 +0000)]
Simplify the allocation of page table pages for mips64. Specifically,
when allocation fails, it suffices to perform VM_WAIT. The direct map
covers the entirely of physical memory, so unlike 32-bit mips any
physical page can be used as a page table page.
ae [Sat, 23 May 2015 09:49:40 +0000 (09:49 +0000)]
Properly update TX statistics for wlan(4).
ieee80211_pwrsave() can fail due to queue overflow, check its return code
and increment oerrors counter when it fails. Also handle more error cases
and update oerrors counter when we don't send mbuf due to some errors.
Return ENETDOWN when parent interface isn't ready. Update obytes and omcasts
counters in corresponding places.
imp [Fri, 22 May 2015 20:50:35 +0000 (20:50 +0000)]
Export the eflags field from the elf header. This allows better
discrimination between different subarch binaries, at least for mips
and arm. Arm is implemented, mips is still tbd, so not currently
exported. aarch64 does not export this because aarch64 binaries use
different tags and flags than arm.
neel [Fri, 22 May 2015 17:34:22 +0000 (17:34 +0000)]
Don't rely on the 'VM-exit instruction length' field in the VMCS to always
have an accurate length on an EPT violation. This is not needed by the
instruction decoding code because it also has to work with AMD/SVM that
does not provide a valid instruction length on a Nested Page Fault.
In collaboration with: Leon Dang (ldang@nahannisys.com)
Discussed with: grehan
MFC after: 1 week
jkim [Fri, 22 May 2015 17:05:21 +0000 (17:05 +0000)]
CALLOUT_MPSAFE has lost its meaning since r141428, i.e., for more than ten
years for head. However, it is continuously misused as the mpsafe argument
for callout_init(9). Deprecate the flag and clean up callout_init() calls
to make them more consistent.
jhb [Fri, 22 May 2015 11:09:41 +0000 (11:09 +0000)]
Expand ktr_mask to be a 64-bit unsigned integer.
The mask does not really need to be updated with atomic operations and
the downside of losing races during transitions is not great (it is
not marked volatile, so those races are pretty wide open as it is).
jhb [Fri, 22 May 2015 11:04:54 +0000 (11:04 +0000)]
Only reparent a traced process to its old parent if the tracing process is
not the old parent. Otherwise, proc_reap() will leave the zombie in place
resulting in the process' status being returned twice to its parent.
Add test cases for PT_TRACE_ME and PT_ATTACH which are fixed by
this change.