Commit Graph

2628 Commits

Author SHA1 Message Date
Hans Petter Selasky
0a2f4606bb Add comment describing the use of pagefault_disable() and
pagefault_enable() in the LinuxKPI.

Suggested by:  		rpokala@
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-17 08:02:46 +00:00
Hans Petter Selasky
3803a97f84 Use __LP64__ to detect presence of suword64() to fix linking and
loading of the LinuxKPI on 32-bit platforms.

Reported by:		lwhsu @
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-16 20:39:31 +00:00
Hans Petter Selasky
810ea5b270 The LinuxKPI pagefault disable and enable functions can only be used
pairwise to support the FreeBSD way of pushing and popping the page
fault flags. Ensure this by requiring every occurrence of pagefault
disable function call to have a corresponding pagefault enable call.

Obtained from:		kmacy @
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-16 16:53:22 +00:00
Hans Petter Selasky
0e05589b39 Implement more userspace memory access functions in the LinuxKPI.
Obtained from:		kmacy @
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-16 16:49:27 +00:00
Hans Petter Selasky
b8a8ed7c96 Define some more LinuxKPI task related macros.
Obtained from:		kmacy @
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-16 12:33:34 +00:00
Hans Petter Selasky
8300fb13dd Add helper function similar to ip_dev_find() to the LinuxKPI to lookup
a network device by its IPv6 address in the given VNET.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-16 10:02:45 +00:00
Hans Petter Selasky
404027276b Add basic support for VIMAGE to the LinuxKPI and ibcore.
Support is implemented by mapping Linux's "struct net" into FreeBSD's
"struct vnet". Currently only vnet0 is supported by ibcore.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-16 09:59:35 +00:00
Dmitry Chagin
5bb676adf7 Fix usage of the same 'i' variable in the external and nested loops.
Submitted by: Svyatoslav <razmyslov at viva64.com>
Sponsored by: PVS-Studio

MFC after:	1 week
2017-03-14 18:29:23 +00:00
Hans Petter Selasky
5f50a414ba Set "current" pointer for LinuxKPI interrupts and timer callbacks.
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-14 14:02:47 +00:00
Konstantin Belousov
01feb4c3d4 Use designated initializers for kevent_copyops.
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2017-03-14 09:25:01 +00:00
Hans Petter Selasky
f73286645f Fix implementation of the DECLARE_WORK() macro in the LinuxKPI to fully
initialize the declared work structure and not only the function callback
pointer.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-09 18:37:17 +00:00
Hans Petter Selasky
9760ac0a3e Implement support for mutexes with deadlock avoidance in the LinuxKPI.
When locking a mutex and deadlock is detected the first mutex lock
call that sees the deadlock will return -EDEADLK .

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-09 18:33:40 +00:00
Hans Petter Selasky
4f688d19fb Cleanup the LinuxKPI mutex wrappers.
Add support for using mutexes during KDB and shutdown. This is also
required for doing mode-switching during panic for drm-next.

Add new mutex functions mutex_init_witness() and mutex_destroy()
allowing LinuxKPI mutexes to be tracked by witness.

Declare mutex_is_locked() and mutex_is_owned() like inline functions
to get cleaner warnings. These functions are used inside WARN_ON()
statements which might look a bit odd if these functions get fully
expanded.

Give mutexes better debug names through the mutex_name() macro when
WITNESS_ALL is defined. The mutex_name() macro can prefix parts of the
filename and line number before the mutex name.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-09 17:01:00 +00:00
Hans Petter Selasky
c23b6e238f Don't create any threads before SI_SUB_INIT_IF in the LinuxKPI. Else
kthread_add() will assert it is called too soon. This fixes a startup
issue when COMPAT_LINUXKPI is in enabled the kernel configuration
file.

Reported by:		Michael Butler <imb@protected-networks.net>
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-09 09:17:43 +00:00
Hans Petter Selasky
43ee32f7df Fix compilation warning for powerpc64 by not using const keyword in
return types:

Type qualifiers ignored on function return type [-Wreturn-type]

Reported by:		andreast @
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-08 21:28:53 +00:00
Hans Petter Selasky
14c5024db8 Cleanup the LinuxKPI slab implementation.
Put large functions into linux_slab.c instead of declaring them static
inline.

Add support for more memory allocation wrappers like kmalloc_array()
and __vmalloc().

Make sure either the M_WAITOK or the M_NOWAIT flag is set and mask
away unused memory allocation flags before calling FreeBSD's malloc()
routine.

Move kmalloc_node() definition to slab.h where it belongs.

Implement support for the SLAB_DESTROY_BY_RCU feature when creating a
kmem_cache which basically means kmem_cache memory is freed using
call_rcu().

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-08 11:09:27 +00:00
Hans Petter Selasky
0e3bbe9197 Implement eth_zero_addr() in the LinuxKPI.
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-08 09:53:20 +00:00
Hans Petter Selasky
a767c1883d Add support for constant pointer constructs to READ_ONCE() in the
LinuxKPI. When the type of the argument is constant the temporary
variable cannot be assigned after the barrier. Instead assign the
temporary variable by initialization.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-07 20:24:34 +00:00
Dmitry Chagin
b8bec5a415 Linux semop system call return EINVAL in case when the invalid nsops
or semid values specified.

MFC after:	1 month
2017-03-07 17:12:22 +00:00
Dmitry Chagin
c10e04f5a0 Linux kernel does not export to the user space ipc_perm.mode values
other than S_IRWXUGO (0777).

MFC after:	1 month
2017-03-07 17:09:12 +00:00
Dmitry Chagin
ab60bc8488 Reduce code duplication between MD Linux code by moving SYSV IPC 64-bit
related struct definitions out into the MI path.

Invert the native ipc structs to the Linux ipc structs convesion logic.
Since 64-bit variant of ipc structs has more precision convert native ipc
structs to the 64-bit Linux ipc structs and then truncate 64-bit values
into the non 64-bit if needed. Unlike Linux, return EOVERFLOW if the
values do not fit.

Fix SYSV IPC for 64-bit Linuxulator which never sets IPC_64 bit.

MFC after:	1 month
2017-03-07 17:07:16 +00:00
Hans Petter Selasky
249a42207b Implement time_is_after_eq_jiffies() function in the LinuxKPI.
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-07 15:37:51 +00:00
Hans Petter Selasky
661a318c83 Fix implementation of the DECLARE_RWSEM() macro in the LinuxKPI.
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-07 15:34:49 +00:00
Hans Petter Selasky
dc0d19dd4a Make sure jiffies value is cast to an integer in the LinuxKPI before
doing millisecond conversion. Under FreeBSD jiffies are 32-bit.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-07 15:33:38 +00:00
Hans Petter Selasky
4cd34a41c9 Use grouptaskqueue for tasklets in the LinuxKPI.
This avoids creating own per-CPU threads and also ensures the tasklet
execution happens on the same CPU core invoking the tasklet.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-07 13:51:14 +00:00
Hans Petter Selasky
ca2ad6bd77 LinuxKPI workqueue cleanup.
This change makes the workqueue implementation behave more like in
Linux, both functionality wise and structure wise.

All workqueue code has been moved to linux_work.c

Add an atomic based statemachine to the work_struct to ensure proper
operation. Prior to this change struct_work was directly mapped to a
FreeBSD task. When a taskqueue has multiple threads the same task may
end up being executed on more than one worker thread simultaneously.
This might cause problems with code coming from Linux, which expects
serial behaviour, similar to Linux tasklets.

Move all global workqueue function names into the linux_xxx domain to
avoid symbol name clashes in the future.

Implement a few more workqueue related functions and macros.

Create two multithreaded taskqueues for the LinuxKPI during module
load, one for time-consuming callbacks and one for non-time consuming
callbacks.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-07 12:09:14 +00:00
Mahdi Mokhtari
8049c6bfb8 Add UNIMPLEMENTED() placeholder macro for
the syscalls that are not implemented in Linux kernel itself.
Cleanup DUMMY() macros.

Reviewed by:	dchagin, trasz
Approved by:	dchagin
MFC after:	1 month
Differential Revision:	https://reviews.freebsd.org/D9804
2017-03-06 18:11:38 +00:00
Hans Petter Selasky
def277d3ef Implement add_timer_on() function in the LinuxKPI.
Obtained from:		kmacy @
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-06 14:56:57 +00:00
Hans Petter Selasky
19bf8ef562 Implement DECLARE_RWSEM() macro in the LinuxKPI to initialize a
Read-Write semaphore during module init time.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-06 12:22:05 +00:00
Hans Petter Selasky
684bcfec89 Give LinuxKPI Read-Write semaphores better debug names when
WITNESS_ALL is defined. The lock name is based on the filename and
line number where the initialisation happens.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-06 12:20:56 +00:00
Hans Petter Selasky
e0db0ddb39 Remove duplicate prototype in the LinuxKPI to fix compilation warning.
Reported by:		emaste @
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-04 20:06:47 +00:00
Dmitry Chagin
8042c504f5 Style(9).
MFC after:	1 month
2017-03-04 08:59:21 +00:00
Dmitry Chagin
e45e3698ae Remove attribute __packed from some IPC struct definition since
Linuxulator is x86 only.
The only notable differences in algnment for an LP64 64-bit system
when compared to a 32-bit system is an eight or large byte types
alignment.

MFC after:	1 month
2017-03-04 08:57:39 +00:00
Dmitry Chagin
c1c8a12139 Hide Linux socketcall constants under corresponding #ifdef since
they are used only in i386 Linuxulator.

MFC after:	1 week
2017-03-04 06:54:05 +00:00
Hans Petter Selasky
1f827dab9e Update the LinuxKPI RCU and SRCU wrappers for the concurrency kit, CK.
- Optimise the RCU implementation to not allocate and free
ck_epoch_records during runtime. Instead allocate two sets of
ck_epoch_records per CPU for general purpose use. The first set is
only used for reader locks and the second set is only used for
synchronization and barriers and is protected with a regular mutex to
prevent simultaneous issues.

- Move the task structure away from the rcu_head structure and into
the per-CPU structures. This allows the size of the rcu_head structure
to be reduced down to the size of two pointers.

- Fix a bug where the linux_rcu_barrier() function only waited for one
per-CPU epoch record to be completed instead of all.

- Use a critical section or a mutex to protect ck_epoch_begin() and
ck_epoch_end() depending on RCU or SRCU type. All the ck_epoch_xxx()
functions, except ck_epoch_register(), ck_epoch_unregister() and
ck_epoch_recycle() are not re-entrant and needs a critical section or
a mutex to operate in the LinuxKPI, after inspecting the CK
implementation of the above mentioned functions. The simultaneous
issues arise from per-CPU epoch records being shared between multiple
threads depending on the amount of taskswitching and how many threads
are involved with the RCU and SRCU operations.

- Properly free all epoch records by using safe list traversal at
LinuxKPI module unload. It turns out the ck_epoch_recycle() always
have the records on an internal list and use a flag in the epoch
record to track allocated and free entries. This would lead to use
after free during module unload.

- Remove redundant synchronize_rcu() call from the
linux_compat_uninit() function. Let the linux_rcu_runtime_uninit()
function do the final rcu_barrier() instead.

MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-03-03 16:28:03 +00:00
Konstantin Belousov
ef130ff64c With the removal of IA64, the only arch which uses ia32 compat is amd64.
Sponsored by:	The FreeBSD Foundation
MFC after:	2 weeks
2017-03-01 11:39:29 +00:00
Dmitry Chagin
0633df29be Linux epoll return EEXIST on case when op is EPOLL_CTL_ADD, and the supplied
file descriptor fd is already registered with this epoll instance.

MFC after:	1 month
2017-02-28 19:55:16 +00:00
Dmitry Chagin
8bc4edfd75 Linux epoll return ENOENT error in case when op is EPOLL_CTL_MOD or
EPOLL_CTL_DEL, and fd is not registered with this epoll instance.

MFC after:	1 month
2017-02-28 19:54:22 +00:00
Dmitry Chagin
29b1ecbe01 FreeBSD does not have analgue for epill EPOLLPRI event type.
So, do not set EPOLLPRI event acidently.
Also, do not set EPOLLWRNORM and EPOLLRDNORM events as epoll
do not set this events.

MFC after:	1 month
2017-02-28 19:49:21 +00:00
Gleb Smirnoff
efe3b0de14 Remove SVR4 (System V Release 4) binary compatibility support.
UNIX System V Release 4 is operating system released in 1988. It ceased
to exist in early 2000-s.
2017-02-28 05:14:42 +00:00
Dmitry Chagin
281afc595f Return EINVAL when an invalid file descriptor specified.
MFC after:	1 month
2017-02-27 16:55:09 +00:00
Dmitry Chagin
ed211c40f4 Unify eventfd ioctl method and use it for other similar interfaces.
MFC after:	1 month
2017-02-27 16:53:52 +00:00
Hans Petter Selasky
3cfeca84b4 Implement more bit operation functions in the LinuxKPI.
Some minor whitespace nits while at it.

Obtained from:		kmacy @
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-02-27 14:38:17 +00:00
Hans Petter Selasky
522f4b2c75 Define __sum16 type in the LinuxKPI.
MFC after:		1 week
Sponsored by:		Mellanox Technologies
2017-02-27 13:59:02 +00:00
Dmitry Chagin
2fa6d2fe84 Return EINVAL in case when an invalid size of signal mask specified.
MFC after:	1 month
2017-02-26 20:01:58 +00:00
Dmitry Chagin
80c7315d17 Restore signal mask in epoll_pwait.
MFC after:	1 month
2017-02-26 19:54:17 +00:00
Dmitry Chagin
e0a254f6df Return EINVAL when an invalid file descriptor is specified.
MFC after:	1 month
2017-02-26 19:51:44 +00:00
Dmitry Chagin
dd93b628e9 Implement timerfd family syscalls.
MFC after:	1 month
2017-02-26 09:48:18 +00:00
Dmitry Chagin
01d4a346c2 Nostly style(9) changes, replace unused eventfd_truncate()
by default invfo_truncate() method.

MFC after:	1 month
2017-02-26 09:42:34 +00:00
Dmitry Chagin
0670e972e3 Return EOVERFLOW error in case then the size of tv_sec field of struct timespec
in COMPAT_LINUX32 Linuxulator's not equal to the size of native tv_sec.

MFC after:	1 month
2017-02-26 09:40:42 +00:00