__run_hrtimer() is called with the hrtimer_cpu_base.lock held and
interrupts disabled. Before invoking the timer callback the base lock is
dropped, but interrupts stay disabled.
The upcoming support for softirq based hrtimers requires that interrupts
are enabled before the timer callback is invoked.
To avoid code duplication, take hrtimer_cpu_base.lock with
raw_spin_lock_irqsave(flags) at the call site and hand in the flags as
a parameter. So raw_spin_unlock_irqrestore() before the callback invocation
will either keep interrupts disabled in interrupt context or restore to
interrupt enabled state when called from softirq context.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: keescook@chromium.org
Link: http://lkml.kernel.org/r/20171221104205.7269-26-anna-maria@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Among the existing architecture specific versions of
copy_siginfo_to_user32 there are several different implementation
problems. Some architectures fail to handle all of the cases in in
the siginfo union. Some architectures perform a blind copy of the
siginfo union when the si_code is negative. A blind copy suggests the
data is expected to be in 32bit siginfo format, which means that
receiving such a signal via signalfd won't work, or that the data is
in 64bit siginfo and the code is copying nonsense to userspace.
Create a single instance of copy_siginfo_to_user32 that all of the
architectures can share, and teach it to handle all of the cases in
the siginfo union correctly, with the assumption that siginfo is
stored internally to the kernel is 64bit siginfo format.
A special case is made for x86 x32 format. This is needed as presence
of both x32 and ia32 on x86_64 results in two different 32bit signal
formats. By allowing this small special case there winds up being
exactly one code base that needs to be maintained between all of the
architectures. Vastly increasing the testing base and the chances of
finding bugs.
As the x86 copy of copy_siginfo_to_user32 the call of the x86
signal_compat_build_tests were moved into sigaction_compat_abi, so
that they will keep running.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
hrtimer_reprogram() must have access to the hrtimer_clock_base of the new
first expiring timer to access hrtimer_clock_base.offset for adjusting the
expiry time to CLOCK_MONOTONIC. This is required to evaluate whether the
new left most timer in the hrtimer_clock_base is the first expiring timer
of all clock bases in a hrtimer_cpu_base.
The only user of hrtimer_reprogram() is hrtimer_start_range_ns(), which has
a pointer to hrtimer_clock_base() already and hands it in as a parameter. But
hrtimer_start_range_ns() will be split for the upcoming support for softirq
based hrtimers to avoid code duplication and will lose the direct access to
the clock base pointer.
Instead of handing in timer and timer->base as a parameter remove the base
parameter from hrtimer_reprogram() instead and retrieve the clock base internally.
Signed-off-by: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: keescook@chromium.org
Link: http://lkml.kernel.org/r/20171221104205.7269-23-anna-maria@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
hrtimer_reprogram() is conditionally invoked from hrtimer_start_range_ns()
when hrtimer_cpu_base.hres_active is true.
In the !hres_active case there is a special condition for the nohz_active
case:
If the newly enqueued timer expires before the first expiring timer on a
remote CPU then the remote CPU needs to be notified and woken up from a
NOHZ idle sleep to take the new first expiring timer into account.
Previous changes have already established the prerequisites to make the
remote enqueue behaviour the same whether high resolution mode is active or
not:
If the to be enqueued timer expires before the first expiring timer on a
remote CPU, then it cannot be enqueued there.
This was done for the high resolution mode because there is no way to
access the remote CPU timer hardware. The same is true for NOHZ, but was
handled differently by unconditionally enqueuing the timer and waking up
the remote CPU so it can reprogram its timer. Again there is no compelling
reason for this difference.
hrtimer_check_target(), which makes the 'can remote enqueue' decision is
already unconditional, but not yet functional because nothing updates
hrtimer_cpu_base.expires_next in the !hres_active case.
To unify this the following changes are required:
1) Make the store of the new first expiry time unconditonal in
hrtimer_reprogram() and check __hrtimer_hres_active() before proceeding
to the actual hardware access. This check also lets the compiler
eliminate the rest of the function in case of CONFIG_HIGH_RES_TIMERS=n.
2) Invoke hrtimer_reprogram() unconditionally from
hrtimer_start_range_ns()
3) Remove the remote wakeup special case for the !high_res && nohz_active
case.
Confine the timers_nohz_active static key to timer.c which is the only user
now.
Signed-off-by: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: keescook@chromium.org
Link: http://lkml.kernel.org/r/20171221104205.7269-21-anna-maria@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When the first hrtimer on the current CPU is removed,
hrtimer_force_reprogram() is invoked but only when
CONFIG_HIGH_RES_TIMERS=y and hrtimer_cpu_base.hres_active is set.
hrtimer_force_reprogram() updates hrtimer_cpu_base.expires_next and
reprograms the clock event device. When CONFIG_HIGH_RES_TIMERS=y and
hrtimer_cpu_base.hres_active is set, a pointless hrtimer interrupt can be
prevented.
hrtimer_check_target() makes the 'can remote enqueue' decision. As soon as
hrtimer_check_target() is unconditionally available and
hrtimer_cpu_base.expires_next is updated by hrtimer_reprogram(),
hrtimer_force_reprogram() needs to be available unconditionally as well to
prevent the following scenario with CONFIG_HIGH_RES_TIMERS=n:
- the first hrtimer on this CPU is removed and hrtimer_force_reprogram() is
not executed
- CPU goes idle (next timer is calculated and hrtimers are taken into
account)
- a hrtimer is enqueued remote on the idle CPU: hrtimer_check_target()
compares expiry value and hrtimer_cpu_base.expires_next. The expiry value
is after expires_next, so the hrtimer is enqueued. This timer will fire
late, if it expires before the effective first hrtimer on this CPU and
the comparison was with an outdated expires_next value.
To prevent this scenario, make hrtimer_force_reprogram() unconditional
except the effective reprogramming part, which gets eliminated by the
compiler in the CONFIG_HIGH_RES_TIMERS=n case.
Signed-off-by: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: keescook@chromium.org
Link: http://lkml.kernel.org/r/20171221104205.7269-20-anna-maria@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
hrtimer_force_reprogram() needs to be available unconditionally for softirq
based hrtimers. Move the function and all required struct members out of
the CONFIG_HIGH_RES_TIMERS #ifdef.
There is no functional change because hrtimer_force_reprogram() is only
invoked when hrtimer_cpu_base.hres_active is true and
CONFIG_HIGH_RES_TIMERS=y.
Making it unconditional increases the text size for the
CONFIG_HIGH_RES_TIMERS=n case slightly, but avoids replication of that code
for the upcoming softirq based hrtimers support. Most of the code gets
eliminated in the CONFIG_HIGH_RES_TIMERS=n case by the compiler.
Signed-off-by: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: keescook@chromium.org
Link: http://lkml.kernel.org/r/20171221104205.7269-19-anna-maria@linutronix.de
[ Made it build on !CONFIG_HIGH_RES_TIMERS ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
hrtimer_cpu_base.next_timer stores the pointer to the next expiring timer
in a CPU base.
This pointer cannot be dereferenced and is solely used to check whether a
hrtimer which is removed is the hrtimer which is the first to expire in the
CPU base. If this is the case, then the timer hardware needs to be
reprogrammed to avoid an extra interrupt for nothing.
Again, this is conditional functionality, but there is no compelling reason
to make this conditional. As a preparation, hrtimer_cpu_base.next_timer
needs to be available unconditonally.
Aside of that the upcoming support for softirq based hrtimers requires access
to this pointer unconditionally as well, so our motivation is not entirely
simplicity based.
Make the update of hrtimer_cpu_base.next_timer unconditional and remove the
#ifdef cruft. The impact on CONFIG_HIGH_RES_TIMERS=n && CONFIG_NOHZ=n is
marginal as it's just a store on an already dirtied cacheline.
No functional change.
Signed-off-by: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: keescook@chromium.org
Link: http://lkml.kernel.org/r/20171221104205.7269-17-anna-maria@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
hrtimer_cpu_base.expires_next is used to cache the next event armed in the
timer hardware. The value is used to check whether an hrtimer can be
enqueued remotely. If the new hrtimer is expiring before expires_next, then
remote enqueue is not possible as the remote hrtimer hardware cannot be
accessed for reprogramming to an earlier expiry time.
The remote enqueue check is currently conditional on
CONFIG_HIGH_RES_TIMERS=y and hrtimer_cpu_base.hres_active. There is no
compelling reason to make this conditional.
Move hrtimer_cpu_base.expires_next out of the CONFIG_HIGH_RES_TIMERS=y
guarded area and remove the conditionals in hrtimer_check_target().
The check is currently a NOOP for the CONFIG_HIGH_RES_TIMERS=n and the
!hrtimer_cpu_base.hres_active case because in these cases nothing updates
hrtimer_cpu_base.expires_next yet. This will be changed with later patches
which further reduce the #ifdef zoo in this code.
Signed-off-by: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: keescook@chromium.org
Link: http://lkml.kernel.org/r/20171221104205.7269-16-anna-maria@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The pointer to the currently running timer is stored in hrtimer_cpu_base
before the base lock is dropped and the callback is invoked.
This results in two levels of indirections and the upcoming support for
softirq based hrtimer requires splitting the "running" storage into soft
and hard IRQ context expiry.
Storing both in the cpu base would require conditionals in all code paths
accessing that information.
It's possible to have a per clock base sequence count and running pointer
without changing the semantics of the related mechanisms because the timer
base pointer cannot be changed while a timer is running the callback.
Unfortunately this makes cpu_clock base larger than 32 bytes on 32-bit
kernels. Instead of having huge gaps due to alignment, remove the alignment
and let the compiler pack CPU base for 32-bit kernels. The resulting cache access
patterns are fortunately not really different from the current
behaviour. On 64-bit kernels the 64-byte alignment stays and the behaviour is
unchanged. This was determined by analyzing the resulting layout and
looking at the number of cache lines involved for the frequently used
clocks.
Signed-off-by: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: keescook@chromium.org
Link: http://lkml.kernel.org/r/20171221104205.7269-12-anna-maria@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The POSIX specification defines that relative CLOCK_REALTIME timers are not
affected by clock modifications. Those timers have to use CLOCK_MONOTONIC
to ensure POSIX compliance.
The introduction of the additional HRTIMER_MODE_PINNED mode broke this
requirement for pinned timers.
There is no user space visible impact because user space timers are not
using pinned mode, but for consistency reasons this needs to be fixed.
Check whether the mode has the HRTIMER_MODE_REL bit set instead of
comparing with HRTIMER_MODE_ABS.
Signed-off-by: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: keescook@chromium.org
Fixes: 597d027573 ("timers: Framework for identifying pinned timers")
Link: http://lkml.kernel.org/r/20171221104205.7269-7-anna-maria@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
schedule_hrtimeout_range_clock() uses an 'int clock' parameter for the
clock ID, instead of the customary predefined "clockid_t" type.
In hrtimer coding style the canonical variable name for the clock ID is
'clock_id', therefore change the name of the parameter here as well
to make it all consistent.
While at it, clean up the description for the 'clock_id' and 'mode'
function parameters. The clock modes and the clock IDs are not
restricted as the comment suggests.
Fix the mode description as well for the callers of schedule_hrtimeout_range_clock().
No functional changes intended.
Signed-off-by: Anna-Maria Gleixner <anna-maria@linutronix.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: keescook@chromium.org
Link: http://lkml.kernel.org/r/20171221104205.7269-5-anna-maria@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The function copy_siginfo_from_user32 is used for two things, in ptrace
since the dawn of siginfo for arbirarily modifying a signal that
user space sees, and in sigqueueinfo to send a signal with arbirary
siginfo data.
Create a single copy of copy_siginfo_from_user32 that all architectures
share, and teach it to handle all of the cases in the siginfo union.
In the generic version of copy_siginfo_from_user32 ensure that all
of the fields in siginfo are initialized so that the siginfo structure
can be safely copied to userspace if necessary.
When copying the embedded sigval union copy the si_int member. That
ensures the 32bit values passes through the kernel unchanged.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Having si_codes in many different files simply encourages duplicate definitions
that can cause problems later. To avoid that merge the blackfin specific si_codes
into uapi/asm-generic/siginfo.h
Update copy_siginfo_to_user to copy with the absence of BUS_MCEERR_AR that blackfin
defines to be something else.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
While the blocked and saved_sigmask fields of task_struct are copied to
userspace (via sigmask_to_save() and setup_rt_frame()), it is always
copied with a static length (i.e. sizeof(sigset_t)).
The only portion of task_struct that is potentially dynamically sized and
may be copied to userspace is in the architecture-specific thread_struct
at the end of task_struct.
cache object allocation:
kernel/fork.c:
alloc_task_struct_node(...):
return kmem_cache_alloc_node(task_struct_cachep, ...);
dup_task_struct(...):
...
tsk = alloc_task_struct_node(node);
copy_process(...):
...
dup_task_struct(...)
_do_fork(...):
...
copy_process(...)
example usage trace:
arch/x86/kernel/fpu/signal.c:
__fpu__restore_sig(...):
...
struct task_struct *tsk = current;
struct fpu *fpu = &tsk->thread.fpu;
...
__copy_from_user(&fpu->state.xsave, ..., state_size);
fpu__restore_sig(...):
...
return __fpu__restore_sig(...);
arch/x86/kernel/signal.c:
restore_sigcontext(...):
...
fpu__restore_sig(...)
This introduces arch_thread_struct_whitelist() to let an architecture
declare specifically where the whitelist should be within thread_struct.
If undefined, the entire thread_struct field is left whitelisted.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Laura Abbott <labbott@redhat.com>
Cc: "Mickaël Salaün" <mic@digikod.net>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Acked-by: Rik van Riel <riel@redhat.com>
In support of usercopy hardening, this patch defines a region in the
thread_stack slab caches in which userspace copy operations are allowed.
Since the entire thread_stack needs to be available to userspace, the
entire slab contents are whitelisted. Note that the slab-based thread
stack is only present on systems with THREAD_SIZE < PAGE_SIZE and
!CONFIG_VMAP_STACK.
cache object allocation:
kernel/fork.c:
alloc_thread_stack_node(...):
return kmem_cache_alloc_node(thread_stack_cache, ...)
dup_task_struct(...):
...
stack = alloc_thread_stack_node(...)
...
tsk->stack = stack;
copy_process(...):
...
dup_task_struct(...)
_do_fork(...):
...
copy_process(...)
This region is known as the slab cache's usercopy region. Slab caches
can now check that each dynamically sized copy operation involving
cache-managed memory falls entirely within the slab's usercopy region.
This patch is modified from Brad Spengler/PaX Team's PAX_USERCOPY
whitelisting code in the last public patch of grsecurity/PaX based on my
understanding of the code. Changes or omissions from the original code are
mine and don't reflect the original grsecurity/PaX code.
Signed-off-by: David Windsor <dave@nullcore.net>
[kees: adjust commit log, split patch, provide usage trace]
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Acked-by: Rik van Riel <riel@redhat.com>
In support of usercopy hardening, this patch defines a region in the
mm_struct slab caches in which userspace copy operations are allowed.
Only the auxv field is copied to userspace.
cache object allocation:
kernel/fork.c:
#define allocate_mm() (kmem_cache_alloc(mm_cachep, GFP_KERNEL))
dup_mm():
...
mm = allocate_mm();
copy_mm(...):
...
dup_mm();
copy_process(...):
...
copy_mm(...)
_do_fork(...):
...
copy_process(...)
example usage trace:
fs/binfmt_elf.c:
create_elf_tables(...):
...
elf_info = (elf_addr_t *)current->mm->saved_auxv;
...
copy_to_user(..., elf_info, ei_index * sizeof(elf_addr_t))
load_elf_binary(...):
...
create_elf_tables(...);
This region is known as the slab cache's usercopy region. Slab caches
can now check that each dynamically sized copy operation involving
cache-managed memory falls entirely within the slab's usercopy region.
This patch is modified from Brad Spengler/PaX Team's PAX_USERCOPY
whitelisting code in the last public patch of grsecurity/PaX based on my
understanding of the code. Changes or omissions from the original code are
mine and don't reflect the original grsecurity/PaX code.
Signed-off-by: David Windsor <dave@nullcore.net>
[kees: adjust commit log, split patch, provide usage trace]
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Acked-by: Rik van Riel <riel@redhat.com>
ftrace_module_init happen after dynamic_debug_setup, it is desired that
cleanup should be called after this label however in current implementation
it is called in free module label,ie:even though ftrace in not initialized,
from so many fail case ftrace_release_mod() will be called and unnecessary
traverse the whole list.
In below patch we moved ftrace_release_mod() from free_module label to
ddebug_cleanup label. that is the best possible location, other solution
is to make new label to ftrace_release_mod() but since ftrace_module_init()
is not return with minimum changes it should be in ddebug_cleanup label.
Signed-off-by: Namit Gupta <gupta.namit@samsung.com>
Reviewed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Jessica Yu <jeyu@kernel.org>
I regularly get 50 MB - 60 MB files during kernel randconfig builds.
These large files mostly contain (many repeats of; e.g., 124,594):
In file included from ../include/linux/string.h:6:0,
from ../include/linux/uuid.h:20,
from ../include/linux/mod_devicetable.h:13,
from ../scripts/mod/devicetable-offsets.c:3:
../include/linux/compiler.h:64:4: warning: '______f' is static but declared in inline function 'strcpy' which is not static [enabled by default]
______f = { \
^
../include/linux/compiler.h:56:23: note: in expansion of macro '__trace_if'
^
../include/linux/string.h:425:2: note: in expansion of macro 'if'
if (p_size == (size_t)-1 && q_size == (size_t)-1)
^
This only happens when CONFIG_FORTIFY_SOURCE=y and
CONFIG_PROFILE_ALL_BRANCHES=y, so prevent PROFILE_ALL_BRANCHES if
FORTIFY_SOURCE=y.
Link: http://lkml.kernel.org/r/9199446b-a141-c0c3-9678-a3f9107f2750@infradead.org
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Commit 1a149d7d3f ("ring-buffer: Rewrite trace_recursive_(un)lock() to be
simpler") replaced the context level recursion checks with a simple counter.
This would prevent the ring buffer code from recursively calling itself more
than the max number of contexts that exist (Normal, softirq, irq, nmi). But
this change caused a lockup in a specific case, which was during suspend and
resume using a global clock. Adding a stack dump to see where this occurred,
the issue was in the trace global clock itself:
trace_buffer_lock_reserve+0x1c/0x50
__trace_graph_entry+0x2d/0x90
trace_graph_entry+0xe8/0x200
prepare_ftrace_return+0x69/0xc0
ftrace_graph_caller+0x78/0xa8
queued_spin_lock_slowpath+0x5/0x1d0
trace_clock_global+0xb0/0xc0
ring_buffer_lock_reserve+0xf9/0x390
The function graph tracer traced queued_spin_lock_slowpath that was called
by trace_clock_global. This pointed out that the trace_clock_global() is not
reentrant, as it takes a spin lock. It depended on the ring buffer recursive
lock from letting that happen.
By removing the context detection and adding just a max number of allowable
recursions, it allowed the trace_clock_global() to be entered again and try
to retake the spinlock it already held, causing a deadlock.
Fixes: 1a149d7d3f ("ring-buffer: Rewrite trace_recursive_(un)lock() to be simpler")
Reported-by: David Weinehall <david.weinehall@gmail.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Instead of the cfs workitem library, use workqueues.
As lnet wants to provide a cpu mask of allowed cpus, it
needs to be a WQ_UNBOUND work queue so that tasks can
run on cpus other than where they were submitted.
This patch also exported apply_workqueue_attrs() which is
a documented part of the workqueue API, that isn't currently
exported. lustre needs it to allow workqueue thread to be limited
to a subset of CPUs.
Acked-by: Tejun Heo <tj@kernel.org> (for export of apply_workqueue_attrs)
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
BPF map offload follow similar path to program offload. At creation
time users may specify ifindex of the device on which they want to
create the map. Map will be validated by the kernel's
.map_alloc_check callback and device driver will be called for the
actual allocation. Map will have an empty set of operations
associated with it (save for alloc and free callbacks). The real
device callbacks are kept in map->offload->dev_ops because they
have slightly different signatures. Map operations are called in
process context so the driver may communicate with HW freely,
msleep(), wait() etc.
Map alloc and free callbacks are muxed via existing .ndo_bpf, and
are always called with rtnl lock held. Maps and programs are
guaranteed to be destroyed before .ndo_uninit (i.e. before
unregister_netdev() returns). Map callbacks are invoked with
bpf_devs_lock *read* locked, drivers must take care of exclusive
locking if necessary.
All offload-specific branches are marked with unlikely() (through
bpf_map_is_dev_bound()), given that branch penalty will be
negligible compared to IO anyway, and we don't want to penalize
SW path unnecessarily.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Add a helper to check if netdev could be found and whether it
has .ndo_bpf callback. There is no need to check the callback
every time it's invoked, ndos can't reasonably be swapped for
a set without .ndp_bpf while program is loaded.
bpf_dev_offload_check() will also be used by map offload.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Number of attribute checks are currently performed after hashtab
is already allocated. Move them to be able to split them out to
the check function later on. Checks have to now be performed on
the attr union directly instead of the members of bpf_map, since
bpf_map will be allocated later. No functional changes.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
.map_alloc callbacks contain a number of checks validating user-
-provided map attributes against constraints of a particular map
type. For offloaded maps we will need to check map attributes
without actually allocating any memory on the host. Add a new
callback for validating attributes before any memory is allocated.
This callback can be selectively implemented by map types for
sharing code with offloads, or simply to separate the logical
steps of validation and allocation.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Because the return value of cpu_timer_sample_group() is not checked,
compilers and static checkers can legitimately warn about a potential use
of the uninitialized variable 'now'. This is not a runtime issue as all call
sites hand in valid clock ids.
Also cpu_timer_sample_group() is invoked unconditionally even when the
result is not used because *oldval is NULL.
Make the invocation conditional and check the return value.
[ tglx: Massage changelog ]
Signed-off-by: Max R. P. Grossmann <m@max.pm>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: john.stultz@linaro.org
Link: https://lkml.kernel.org/r/20180108190157.10048-1-m@max.pm
Julia reported futex state corruption in the following scenario:
waiter waker stealer (prio > waiter)
futex(WAIT_REQUEUE_PI, uaddr, uaddr2,
timeout=[N ms])
futex_wait_requeue_pi()
futex_wait_queue_me()
freezable_schedule()
<scheduled out>
futex(LOCK_PI, uaddr2)
futex(CMP_REQUEUE_PI, uaddr,
uaddr2, 1, 0)
/* requeues waiter to uaddr2 */
futex(UNLOCK_PI, uaddr2)
wake_futex_pi()
cmp_futex_value_locked(uaddr2, waiter)
wake_up_q()
<woken by waker>
<hrtimer_wakeup() fires,
clears sleeper->task>
futex(LOCK_PI, uaddr2)
__rt_mutex_start_proxy_lock()
try_to_take_rt_mutex() /* steals lock */
rt_mutex_set_owner(lock, stealer)
<preempted>
<scheduled in>
rt_mutex_wait_proxy_lock()
__rt_mutex_slowlock()
try_to_take_rt_mutex() /* fails, lock held by stealer */
if (timeout && !timeout->task)
return -ETIMEDOUT;
fixup_owner()
/* lock wasn't acquired, so,
fixup_pi_state_owner skipped */
return -ETIMEDOUT;
/* At this point, we've returned -ETIMEDOUT to userspace, but the
* futex word shows waiter to be the owner, and the pi_mutex has
* stealer as the owner */
futex_lock(LOCK_PI, uaddr2)
-> bails with EDEADLK, futex word says we're owner.
And suggested that what commit:
73d786bd04 ("futex: Rework inconsistent rt_mutex/futex_q state")
removes from fixup_owner() looks to be just what is needed. And indeed
it is -- I completely missed that requeue_pi could also result in this
case. So we need to restore that, except that subsequent patches, like
commit:
16ffa12d74 ("futex: Pull rt_mutex_futex_unlock() out from under hb->lock")
changed all the locking rules. Even without that, the sequence:
- if (rt_mutex_futex_trylock(&q->pi_state->pi_mutex)) {
- locked = 1;
- goto out;
- }
- raw_spin_lock_irq(&q->pi_state->pi_mutex.wait_lock);
- owner = rt_mutex_owner(&q->pi_state->pi_mutex);
- if (!owner)
- owner = rt_mutex_next_owner(&q->pi_state->pi_mutex);
- raw_spin_unlock_irq(&q->pi_state->pi_mutex.wait_lock);
- ret = fixup_pi_state_owner(uaddr, q, owner);
already suggests there were races; otherwise we'd never have to look
at next_owner.
So instead of doing 3 consecutive wait_lock sections with who knows
what races, we do it all in a single section. Additionally, the usage
of pi_state->owner in fixup_owner() was only safe because only the
rt_mutex owner would modify it, which this additional case wrecks.
Luckily the values can only change away and not to the value we're
testing, this means we can do a speculative test and double check once
we have the wait_lock.
Fixes: 73d786bd04 ("futex: Rework inconsistent rt_mutex/futex_q state")
Reported-by: Julia Cartwright <julia@ni.com>
Reported-by: Gratian Crisan <gratian.crisan@ni.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Julia Cartwright <julia@ni.com>
Tested-by: Gratian Crisan <gratian.crisan@ni.com>
Cc: Darren Hart <dvhart@infradead.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20171208124939.7livp7no2ov65rrc@hirez.programming.kicks-ass.net
Divides by zero are not nice, lets avoid them if possible.
Also do_div() seems not needed when dealing with 32bit operands,
but this seems a minor detail.
Fixes: bd4cf0ed33 ("net: filter: rework/optimize internal BPF interpreter's instruction set")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Merge misc fixlets from Andrew Morton:
"4 fixes"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
tools/objtool/Makefile: don't assume sync-check.sh is executable
kdump: write correct address of mem_section into vmcoreinfo
kmemleak: allow to coexist with fault injection
MAINTAINERS, nilfs2: change project home URLs
Support in-kernel fault-injection framework via debugfs.
This allows you to inject a conditional error to specified
function using debugfs interfaces.
Here is the result of test script described in
Documentation/fault-injection/fault-injection.txt
===========
# ./test_fail_function.sh
1+0 records in
1+0 records out
1048576 bytes (1.0 MB, 1.0 MiB) copied, 0.0227404 s, 46.1 MB/s
btrfs-progs v4.4
See http://btrfs.wiki.kernel.org for more information.
Label: (null)
UUID: bfa96010-12e9-4360-aed0-42eec7af5798
Node size: 16384
Sector size: 4096
Filesystem size: 1001.00MiB
Block group profiles:
Data: single 8.00MiB
Metadata: DUP 58.00MiB
System: DUP 12.00MiB
SSD detected: no
Incompat features: extref, skinny-metadata
Number of devices: 1
Devices:
ID SIZE PATH
1 1001.00MiB /dev/loop2
mount: mount /dev/loop2 on /opt/tmpmnt failed: Cannot allocate memory
SUCCESS!
===========
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Reviewed-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Since error-injection framework is not limited to be used
by kprobes, nor bpf. Other kernel subsystems can use it
freely for checking safeness of error-injection, e.g.
livepatch, ftrace etc.
So this separate error-injection framework from kprobes.
Some differences has been made:
- "kprobe" word is removed from any APIs/structures.
- BPF_ALLOW_ERROR_INJECTION() is renamed to
ALLOW_ERROR_INJECTION() since it is not limited for BPF too.
- CONFIG_FUNCTION_ERROR_INJECTION is the config item of this
feature. It is automatically enabled if the arch supports
error injection feature for kprobe or ftrace etc.
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Reviewed-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
