The rcu_boost_kthread_setaffinity() function removes the outgoing CPU
from the set_cpus_allowed() mask for the corresponding leaf rcu_node
structure's rcub priority-boosting kthread. Except that if the outgoing
CPU will leave that structure without any online CPUs, the mask is set
to the housekeeping CPU mask from housekeeping_cpumask(). Which is fine
unless the outgoing CPU happens to be a housekeeping CPU.
This commit therefore removes the outgoing CPU from the housekeeping mask.
This would of course be problematic if the outgoing CPU was the last
online housekeeping CPU, but in that case you are in a world of hurt
anyway. If someone comes up with a valid use case for a system needing
all the housekeeping CPUs to be offline, further adjustments can be made.
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Kernels built with PREEMPT_RCU=y and RCU_STRICT_GRACE_PERIOD=y trigger
irq-work from rcu_read_unlock(), and the resulting irq-work handler
invokes rcu_preempt_deferred_qs_handle(). The point of this triggering
is to force grace periods to end quickly in order to give tools like KASAN
a better chance of detecting RCU usage bugs such as leaking RCU-protected
pointers out of an RCU read-side critical section.
However, this irq-work triggering is unconditional. This works, but
there is no point in doing this irq-work unless the current grace period
is waiting on the running CPU or task, which is not the common case.
After all, in the common case there are many rcu_read_unlock() calls
per CPU per grace period.
This commit therefore triggers the irq-work only when the current grace
period is waiting on the running CPU or task.
This change was tested as follows on a four-CPU system:
echo rcu_preempt_deferred_qs_handler > /sys/kernel/debug/tracing/set_ftrace_filter
echo 1 > /sys/kernel/debug/tracing/function_profile_enabled
insmod rcutorture.ko
sleep 20
rmmod rcutorture.ko
echo 0 > /sys/kernel/debug/tracing/function_profile_enabled
echo > /sys/kernel/debug/tracing/set_ftrace_filter
This procedure produces results in this per-CPU set of files:
/sys/kernel/debug/tracing/trace_stat/function*
Sample output from one of these files is as follows:
Function Hit Time Avg s^2
-------- --- ---- --- ---
rcu_preempt_deferred_qs_handle 838746 182650.3 us 0.217 us 0.004 us
The baseline sum of the "Hit" values (the number of calls to this
function) was 3,319,015. With this commit, that sum was 1,140,359,
for a 2.9x reduction. The worst-case variance across the CPUs was less
than 25%, so this large effect size is statistically significant.
The raw data is available in the Link: URL.
Link: https://lore.kernel.org/all/20220808022626.12825-1-qiang1.zhang@intel.com/
Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
The dump_cpu_task() function does not print registers on architectures
that do not support NMIs. However, registers can be useful for
debugging. Fortunately, in the case where dump_cpu_task() is invoked
from an interrupt handler and is dumping the current CPU's stack, the
get_irq_regs() function can be used to get the registers.
Therefore, this commit makes dump_cpu_task() check to see if it is being
asked to dump the current CPU's stack from within an interrupt handler,
and, if so, it uses the get_irq_regs() function to obtain the registers.
On systems that do support NMIs, this commit has the further advantage
of avoiding a self-NMI in this case.
This is an example of rcu self-detected stall on arm64, which does not
support NMIs:
[ 27.501721] rcu: INFO: rcu_preempt self-detected stall on CPU
[ 27.502238] rcu: 0-....: (1250 ticks this GP) idle=4f7/1/0x4000000000000000 softirq=2594/2594 fqs=619
[ 27.502632] (t=1251 jiffies g=2989 q=29 ncpus=4)
[ 27.503845] CPU: 0 PID: 306 Comm: test0 Not tainted 5.19.0-rc7-00009-g1c1a6c29ff99-dirty #46
[ 27.504732] Hardware name: linux,dummy-virt (DT)
[ 27.504947] pstate: 20000005 (nzCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[ 27.504998] pc : arch_counter_read+0x18/0x24
[ 27.505301] lr : arch_counter_read+0x18/0x24
[ 27.505328] sp : ffff80000b29bdf0
[ 27.505345] x29: ffff80000b29bdf0 x28: 0000000000000000 x27: 0000000000000000
[ 27.505475] x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000
[ 27.505553] x23: 0000000000001f40 x22: ffff800009849c48 x21: 000000065f871ae0
[ 27.505627] x20: 00000000000025ec x19: ffff80000a6eb300 x18: ffffffffffffffff
[ 27.505654] x17: 0000000000000001 x16: 0000000000000000 x15: ffff80000a6d0296
[ 27.505681] x14: ffffffffffffffff x13: ffff80000a29bc18 x12: 0000000000000426
[ 27.505709] x11: 0000000000000162 x10: ffff80000a2f3c18 x9 : ffff80000a29bc18
[ 27.505736] x8 : 00000000ffffefff x7 : ffff80000a2f3c18 x6 : 00000000759bd013
[ 27.505761] x5 : 01ffffffffffffff x4 : 0002dc6c00000000 x3 : 0000000000000017
[ 27.505787] x2 : 00000000000025ec x1 : ffff80000b29bdf0 x0 : 0000000075a30653
[ 27.505937] Call trace:
[ 27.506002] arch_counter_read+0x18/0x24
[ 27.506171] ktime_get+0x48/0xa0
[ 27.506207] test_task+0x70/0xf0
[ 27.506227] kthread+0x10c/0x110
[ 27.506243] ret_from_fork+0x10/0x20
This is a marked improvement over the old output:
[ 27.944550] rcu: INFO: rcu_preempt self-detected stall on CPU
[ 27.944980] rcu: 0-....: (1249 ticks this GP) idle=cbb/1/0x4000000000000000 softirq=2610/2610 fqs=614
[ 27.945407] (t=1251 jiffies g=2681 q=28 ncpus=4)
[ 27.945731] Task dump for CPU 0:
[ 27.945844] task:test0 state:R running task stack: 0 pid: 306 ppid: 2 flags:0x0000000a
[ 27.946073] Call trace:
[ 27.946151] dump_backtrace.part.0+0xc8/0xd4
[ 27.946378] show_stack+0x18/0x70
[ 27.946405] sched_show_task+0x150/0x180
[ 27.946427] dump_cpu_task+0x44/0x54
[ 27.947193] rcu_dump_cpu_stacks+0xec/0x130
[ 27.947212] rcu_sched_clock_irq+0xb18/0xef0
[ 27.947231] update_process_times+0x68/0xac
[ 27.947248] tick_sched_handle+0x34/0x60
[ 27.947266] tick_sched_timer+0x4c/0xa4
[ 27.947281] __hrtimer_run_queues+0x178/0x360
[ 27.947295] hrtimer_interrupt+0xe8/0x244
[ 27.947309] arch_timer_handler_virt+0x38/0x4c
[ 27.947326] handle_percpu_devid_irq+0x88/0x230
[ 27.947342] generic_handle_domain_irq+0x2c/0x44
[ 27.947357] gic_handle_irq+0x44/0xc4
[ 27.947376] call_on_irq_stack+0x2c/0x54
[ 27.947415] do_interrupt_handler+0x80/0x94
[ 27.947431] el1_interrupt+0x34/0x70
[ 27.947447] el1h_64_irq_handler+0x18/0x24
[ 27.947462] el1h_64_irq+0x64/0x68 <--- the above backtrace is worthless
[ 27.947474] arch_counter_read+0x18/0x24
[ 27.947487] ktime_get+0x48/0xa0
[ 27.947501] test_task+0x70/0xf0
[ 27.947520] kthread+0x10c/0x110
[ 27.947538] ret_from_fork+0x10/0x20
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Ben Segall <bsegall@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Valentin Schneider <vschneid@redhat.com>
The trigger_all_cpu_backtrace() function attempts to send an NMI to the
target CPU, which usually provides much better stack traces than the
dump_cpu_task() function's approach of dumping that stack from some other
CPU. So much so that most calls to dump_cpu_task() only happen after
a call to trigger_all_cpu_backtrace() has failed. And the exception to
this rule really should attempt to use trigger_all_cpu_backtrace() first.
Therefore, move the trigger_all_cpu_backtrace() invocation into
dump_cpu_task().
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Ben Segall <bsegall@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Valentin Schneider <vschneid@redhat.com>
Given that rcu_all_qs() is in non-preemptible kernels, why on earth should
it invoke preempt_disable()? This commit adds the reason, which is to
work nicely with debugging enabled in CONFIG_PREEMPT_COUNT=y kernels.
Reported-by: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Reported-by: Boqun Feng <boqun.feng@gmail.com>
Reported-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Currently, only Tree RCU leaks callbacks setting when it detects a
duplicate call_rcu(). This commit causes Tiny RCU to also leak
callbacks in this situation.
Because this is Tiny RCU, kernel size is important:
1. CONFIG_TINY_RCU=y and CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
(Production kernel)
Original:
text data bss dec hex filename
26290663 20159823 15212544 61663030 3ace736 vmlinux
With this commit:
text data bss dec hex filename
26290663 20159823 15212544 61663030 3ace736 vmlinux
2. CONFIG_TINY_RCU=y and CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
(Debugging kernel)
Original:
text data bss dec hex filename
26291319 20160143 15212544 61664006 3aceb06 vmlinux
With this commit:
text data bss dec hex filename
26291319 20160431 15212544 61664294 3acec26 vmlinux
These results show that the kernel size is unchanged for production
kernels, as desired.
Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Kernels built with CONFIG_PREEMPTION=n and CONFIG_PREEMPT_COUNT=y maintain
preempt_count() state. Because such kernels map __rcu_read_lock()
and __rcu_read_unlock() to preempt_disable() and preempt_enable(),
respectively, this allows the expedited grace period's !CONFIG_PREEMPT_RCU
version of the rcu_exp_handler() IPI handler function to use
preempt_count() to detect quiescent states.
This preempt_count() usage might seem to risk failures due to
use of implicit RCU readers in portions of the kernel under #ifndef
CONFIG_PREEMPTION, except that rcu_core() already disallows such implicit
RCU readers. The moral of this story is that you must use explicit
read-side markings such as rcu_read_lock() or preempt_disable() even if
the code knows that this kernel does not support preemption.
This commit therefore adds a preempt_count()-based check for a quiescent
state in the !CONFIG_PREEMPT_RCU version of the rcu_exp_handler()
function for kernels built with CONFIG_PREEMPT_COUNT=y, reporting an
immediate quiescent state when the interrupted code had both preemption
and softirqs enabled.
This change results in about a 2% reduction in expedited grace-period
latency in kernels built with both CONFIG_PREEMPT_RCU=n and
CONFIG_PREEMPT_COUNT=y.
Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Link: https://lore.kernel.org/all/20220622103549.2840087-1-qiang1.zhang@intel.com/
In non-premptible kernels, tasks never do context switches within
RCU read-side critical sections. Therefore, in such kernels, each
leaf rcu_node structure's ->blkd_tasks list will always be empty.
The comment on the non-preemptible version of rcu_preempt_deferred_qs()
confuses this point, so this commit therefore fixes it.
Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Kernels built with CONFIG_PREEMPT=n and CONFIG_RCU_STRICT_GRACE_PERIOD=y
report the quiescent state directly from the outermost rcu_read_unlock().
However, the current CPU's rcu_data structure's ->cpu_no_qs.b.norm
might still be set, in which case rcu_report_qs_rdp() will exit early,
thus failing to report quiescent state.
This commit therefore causes rcu_read_unlock_strict() to clear
CPU's rcu_data structure's ->cpu_no_qs.b.norm field before invoking
rcu_report_qs_rdp().
Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
We can still see that a majority of the time is spent hashing task pointers:
...
16.98% [kernel] [k] rhashtable_jhash2
...
Doing the bookkeeping in toggle_bp_slots() is currently O(#cpus),
calling task_bp_pinned() for each CPU, even if task_bp_pinned() is
CPU-independent. The reason for this is to update the per-CPU
'tsk_pinned' histogram.
To optimize the CPU-independent case to O(1), keep a separate
CPU-independent 'tsk_pinned_all' histogram.
The major source of complexity are transitions between "all
CPU-independent task breakpoints" and "mixed CPU-independent and
CPU-dependent task breakpoints". The code comments list all cases that
require handling.
After this optimization:
| $> perf bench -r 100 breakpoint thread -b 4 -p 128 -t 512
| # Running 'breakpoint/thread' benchmark:
| # Created/joined 100 threads with 4 breakpoints and 128 parallelism
| Total time: 1.758 [sec]
|
| 34.336621 usecs/op
| 4395.087500 usecs/op/cpu
38.08% [kernel] [k] queued_spin_lock_slowpath
10.81% [kernel] [k] smp_cfm_core_cond
3.01% [kernel] [k] update_sg_lb_stats
2.58% [kernel] [k] osq_lock
2.57% [kernel] [k] llist_reverse_order
1.45% [kernel] [k] find_next_bit
1.21% [kernel] [k] flush_tlb_func_common
1.01% [kernel] [k] arch_install_hw_breakpoint
Showing that the time spent hashing keys has become insignificant.
With the given benchmark parameters, that's an improvement of 12%
compared with the old O(#cpus) version.
And finally, using the less aggressive parameters from the preceding
changes, we now observe:
| $> perf bench -r 30 breakpoint thread -b 4 -p 64 -t 64
| # Running 'breakpoint/thread' benchmark:
| # Created/joined 30 threads with 4 breakpoints and 64 parallelism
| Total time: 0.067 [sec]
|
| 35.292187 usecs/op
| 2258.700000 usecs/op/cpu
Which is an improvement of 12% compared to without the histogram
optimizations (baseline is 40 usecs/op). This is now on par with the
theoretical ideal (constraints disabled), and only 12% slower than no
breakpoints at all.
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20220829124719.675715-15-elver@google.com
Running the perf benchmark with (note: more aggressive parameters vs.
preceding changes, but same 256 CPUs host):
| $> perf bench -r 100 breakpoint thread -b 4 -p 128 -t 512
| # Running 'breakpoint/thread' benchmark:
| # Created/joined 100 threads with 4 breakpoints and 128 parallelism
| Total time: 1.989 [sec]
|
| 38.854160 usecs/op
| 4973.332500 usecs/op/cpu
20.43% [kernel] [k] queued_spin_lock_slowpath
18.75% [kernel] [k] osq_lock
16.98% [kernel] [k] rhashtable_jhash2
8.34% [kernel] [k] task_bp_pinned
4.23% [kernel] [k] smp_cfm_core_cond
3.65% [kernel] [k] bcmp
2.83% [kernel] [k] toggle_bp_slot
1.87% [kernel] [k] find_next_bit
1.49% [kernel] [k] __reserve_bp_slot
We can see that a majority of the time is now spent hashing task
pointers to index into task_bps_ht in task_bp_pinned().
Obtaining the max_bp_pinned_slots() for CPU-independent task targets
currently is O(#cpus), and calls task_bp_pinned() for each CPU, even if
the result of task_bp_pinned() is CPU-independent.
The loop in max_bp_pinned_slots() wants to compute the maximum slots
across all CPUs. If task_bp_pinned() is CPU-independent, we can do so by
obtaining the max slots across all CPUs and adding task_bp_pinned().
To do so in O(1), use a bp_slots_histogram for CPU-pinned slots.
After this optimization:
| $> perf bench -r 100 breakpoint thread -b 4 -p 128 -t 512
| # Running 'breakpoint/thread' benchmark:
| # Created/joined 100 threads with 4 breakpoints and 128 parallelism
| Total time: 1.930 [sec]
|
| 37.697832 usecs/op
| 4825.322500 usecs/op/cpu
19.13% [kernel] [k] queued_spin_lock_slowpath
18.21% [kernel] [k] rhashtable_jhash2
15.46% [kernel] [k] osq_lock
6.27% [kernel] [k] toggle_bp_slot
5.91% [kernel] [k] task_bp_pinned
5.05% [kernel] [k] smp_cfm_core_cond
1.78% [kernel] [k] update_sg_lb_stats
1.36% [kernel] [k] llist_reverse_order
1.34% [kernel] [k] find_next_bit
1.19% [kernel] [k] bcmp
Suggesting that time spent in task_bp_pinned() has been reduced.
However, we're still hashing too much, which will be addressed in the
subsequent change.
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20220829124719.675715-14-elver@google.com
Factor out the existing `atomic_t count[N]` into its own struct called
'bp_slots_histogram', to generalize and make its intent clearer in
preparation of reusing elsewhere. The basic idea of bucketing "total
uses of N slots" resembles a histogram, so calling it such seems most
intuitive.
No functional change.
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20220829124719.675715-13-elver@google.com
While optimizing task_bp_pinned()'s runtime complexity to O(1) on
average helps reduce time spent in the critical section, we still suffer
due to serializing everything via 'nr_bp_mutex'. Indeed, a profile shows
that now contention is the biggest issue:
95.93% [kernel] [k] osq_lock
0.70% [kernel] [k] mutex_spin_on_owner
0.22% [kernel] [k] smp_cfm_core_cond
0.18% [kernel] [k] task_bp_pinned
0.18% [kernel] [k] rhashtable_jhash2
0.15% [kernel] [k] queued_spin_lock_slowpath
when running the breakpoint benchmark with (system with 256 CPUs):
| $> perf bench -r 30 breakpoint thread -b 4 -p 64 -t 64
| # Running 'breakpoint/thread' benchmark:
| # Created/joined 30 threads with 4 breakpoints and 64 parallelism
| Total time: 0.207 [sec]
|
| 108.267188 usecs/op
| 6929.100000 usecs/op/cpu
The main concern for synchronizing the breakpoint constraints data is
that a consistent snapshot of the per-CPU and per-task data is observed.
The access pattern is as follows:
1. If the target is a task: the task's pinned breakpoints are counted,
checked for space, and then appended to; only bp_cpuinfo::cpu_pinned
is used to check for conflicts with CPU-only breakpoints;
bp_cpuinfo::tsk_pinned are incremented/decremented, but otherwise
unused.
2. If the target is a CPU: bp_cpuinfo::cpu_pinned are counted, along
with bp_cpuinfo::tsk_pinned; after a successful check, cpu_pinned is
incremented. No per-task breakpoints are checked.
Since rhltable safely synchronizes insertions/deletions, we can allow
concurrency as follows:
1. If the target is a task: independent tasks may update and check the
constraints concurrently, but same-task target calls need to be
serialized; since bp_cpuinfo::tsk_pinned is only updated, but not
checked, these modifications can happen concurrently by switching
tsk_pinned to atomic_t.
2. If the target is a CPU: access to the per-CPU constraints needs to
be serialized with other CPU-target and task-target callers (to
stabilize the bp_cpuinfo::tsk_pinned snapshot).
We can allow the above concurrency by introducing a per-CPU constraints
data reader-writer lock (bp_cpuinfo_sem), and per-task mutexes (reuses
task_struct::perf_event_mutex):
1. If the target is a task: acquires perf_event_mutex, and acquires
bp_cpuinfo_sem as a reader. The choice of percpu-rwsem minimizes
contention in the presence of many read-lock but few write-lock
acquisitions: we assume many orders of magnitude more task target
breakpoints creations/destructions than CPU target breakpoints.
2. If the target is a CPU: acquires bp_cpuinfo_sem as a writer.
With these changes, contention with thousands of tasks is reduced to the
point where waiting on locking no longer dominates the profile:
| $> perf bench -r 30 breakpoint thread -b 4 -p 64 -t 64
| # Running 'breakpoint/thread' benchmark:
| # Created/joined 30 threads with 4 breakpoints and 64 parallelism
| Total time: 0.077 [sec]
|
| 40.201563 usecs/op
| 2572.900000 usecs/op/cpu
21.54% [kernel] [k] task_bp_pinned
20.18% [kernel] [k] rhashtable_jhash2
6.81% [kernel] [k] toggle_bp_slot
5.47% [kernel] [k] queued_spin_lock_slowpath
3.75% [kernel] [k] smp_cfm_core_cond
3.48% [kernel] [k] bcmp
On this particular setup that's a speedup of 2.7x.
We're also getting closer to the theoretical ideal performance through
optimizations in hw_breakpoint.c -- constraints accounting disabled:
| perf bench -r 30 breakpoint thread -b 4 -p 64 -t 64
| # Running 'breakpoint/thread' benchmark:
| # Created/joined 30 threads with 4 breakpoints and 64 parallelism
| Total time: 0.067 [sec]
|
| 35.286458 usecs/op
| 2258.333333 usecs/op/cpu
Which means the current implementation is ~12% slower than the
theoretical ideal.
For reference, performance without any breakpoints:
| $> bench -r 30 breakpoint thread -b 0 -p 64 -t 64
| # Running 'breakpoint/thread' benchmark:
| # Created/joined 30 threads with 0 breakpoints and 64 parallelism
| Total time: 0.060 [sec]
|
| 31.365625 usecs/op
| 2007.400000 usecs/op/cpu
On a system with 256 CPUs, the theoretical ideal is only ~12% slower
than no breakpoints at all; the current implementation is ~28% slower.
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20220829124719.675715-12-elver@google.com
Flexible breakpoints have never been implemented, with
bp_cpuinfo::flexible always being 0. Unfortunately, they still occupy 4
bytes in each bp_cpuinfo and bp_busy_slots, as well as computing the max
flexible count in fetch_bp_busy_slots().
This again causes suboptimal code generation, when we always know that
`!!slots.flexible` will be 0.
Just get rid of the flexible "placeholder" and remove all real code
related to it. Make a note in the comment related to the constraints
algorithm but don't remove them from the algorithm, so that if in future
flexible breakpoints need supporting, it should be trivial to revive
them (along with reverting this change).
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20220829124719.675715-9-elver@google.com
Due to being a __weak function, hw_breakpoint_weight() will cause the
compiler to always emit a call to it. This generates unnecessarily bad
code (register spills etc.) for no good reason; in fact it appears in
profiles of `perf bench -r 100 breakpoint thread -b 4 -p 128 -t 512`:
...
0.70% [kernel] [k] hw_breakpoint_weight
...
While a small percentage, no architecture defines its own
hw_breakpoint_weight() nor are there users outside hw_breakpoint.c,
which makes the fact it is currently __weak a poor choice.
Change hw_breakpoint_weight()'s definition to follow a similar protocol
to hw_breakpoint_slots(), such that if <asm/hw_breakpoint.h> defines
hw_breakpoint_weight(), we'll use it instead.
The result is that it is inlined and no longer shows up in profiles.
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20220829124719.675715-8-elver@google.com
Optimize internal hw_breakpoint state if the architecture's number of
breakpoint slots is constant. This avoids several kmalloc() calls and
potentially unnecessary failures if the allocations fail, as well as
subtly improves code generation and cache locality.
The protocol is that if an architecture defines hw_breakpoint_slots via
the preprocessor, it must be constant and the same for all types.
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20220829124719.675715-7-elver@google.com
On a machine with 256 CPUs, running the recently added perf breakpoint
benchmark results in:
| $> perf bench -r 30 breakpoint thread -b 4 -p 64 -t 64
| # Running 'breakpoint/thread' benchmark:
| # Created/joined 30 threads with 4 breakpoints and 64 parallelism
| Total time: 236.418 [sec]
|
| 123134.794271 usecs/op
| 7880626.833333 usecs/op/cpu
The benchmark tests inherited breakpoint perf events across many
threads.
Looking at a perf profile, we can see that the majority of the time is
spent in various hw_breakpoint.c functions, which execute within the
'nr_bp_mutex' critical sections which then results in contention on that
mutex as well:
37.27% [kernel] [k] osq_lock
34.92% [kernel] [k] mutex_spin_on_owner
12.15% [kernel] [k] toggle_bp_slot
11.90% [kernel] [k] __reserve_bp_slot
The culprit here is task_bp_pinned(), which has a runtime complexity of
O(#tasks) due to storing all task breakpoints in the same list and
iterating through that list looking for a matching task. Clearly, this
does not scale to thousands of tasks.
Instead, make use of the "rhashtable" variant "rhltable" which stores
multiple items with the same key in a list. This results in average
runtime complexity of O(1) for task_bp_pinned().
With the optimization, the benchmark shows:
| $> perf bench -r 30 breakpoint thread -b 4 -p 64 -t 64
| # Running 'breakpoint/thread' benchmark:
| # Created/joined 30 threads with 4 breakpoints and 64 parallelism
| Total time: 0.208 [sec]
|
| 108.422396 usecs/op
| 6939.033333 usecs/op/cpu
On this particular setup that's a speedup of ~1135x.
While one option would be to make task_struct a breakpoint list node,
this would only further bloat task_struct for infrequently used data.
Furthermore, after all optimizations in this series, there's no evidence
it would result in better performance: later optimizations make the time
spent looking up entries in the hash table negligible (we'll reach the
theoretical ideal performance i.e. no constraints).
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20220829124719.675715-5-elver@google.com
Clean up headers:
- Remove unused <linux/kallsyms.h>
- Remove unused <linux/kprobes.h>
- Remove unused <linux/module.h>
- Remove unused <linux/smp.h>
- Add <linux/export.h> for EXPORT_SYMBOL_GPL().
- Add <linux/mutex.h> for mutex.
- Sort alphabetically.
- Move <linux/hw_breakpoint.h> to top to test it compiles on its own.
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20220829124719.675715-4-elver@google.com
Add KUnit test for hw_breakpoint constraints accounting, with various
interesting mixes of breakpoint targets (some care was taken to catch
interesting corner cases via bug-injection).
The test cannot be built as a module because it requires access to
hw_breakpoint_slots(), which is not inlinable or exported on all
architectures.
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20220829124719.675715-2-elver@google.com
We had historically not checked that genlmsghdr.reserved
is 0 on input which prevents us from using those precious
bytes in the future.
One use case would be to extend the cmd field, which is
currently just 8 bits wide and 256 is not a lot of commands
for some core families.
To make sure that new families do the right thing by default
put the onus of opting out of validation on existing families.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com> (NetLabel)
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull more hotfixes from Andrew Morton:
"Seventeen hotfixes. Mostly memory management things.
Ten patches are cc:stable, addressing pre-6.0 issues"
* tag 'mm-hotfixes-stable-2022-08-28' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:
.mailmap: update Luca Ceresoli's e-mail address
mm/mprotect: only reference swap pfn page if type match
squashfs: don't call kmalloc in decompressors
mm/damon/dbgfs: avoid duplicate context directory creation
mailmap: update email address for Colin King
asm-generic: sections: refactor memory_intersects
bootmem: remove the vmemmap pages from kmemleak in put_page_bootmem
ocfs2: fix freeing uninitialized resource on ocfs2_dlm_shutdown
Revert "memcg: cleanup racy sum avoidance code"
mm/zsmalloc: do not attempt to free IS_ERR handle
binder_alloc: add missing mmap_lock calls when using the VMA
mm: re-allow pinning of zero pfns (again)
vmcoreinfo: add kallsyms_num_syms symbol
mailmap: update Guilherme G. Piccoli's email addresses
writeback: avoid use-after-free after removing device
shmem: update folio if shmem_replace_page() updates the page
mm/hugetlb: avoid corrupting page->mapping in hugetlb_mcopy_atomic_pte
Pull audit fix from Paul Moore:
"Another small audit patch, this time to fix a bug where the return
codes were not properly set before the audit filters were run,
potentially resulting in missed audit records"
* tag 'audit-pr-20220826' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/audit:
audit: move audit_return_fixup before the filters
Since audit_proctitle is generated at syscall exit time, its value is
used immediately and cached for the next syscall. Since this is the
case, then only clear it at task exit time. Otherwise, there is no
point in caching the value OR bearing the overhead of regenerating it.
Fixes: 12c5e81d3f ("audit: prepare audit_context for use in calling contexts beyond syscalls")
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Be explicit in checking the struct audit_context "context" member enum
value rather than assuming the order of context enum values.
Fixes: 12c5e81d3f ("audit: prepare audit_context for use in calling contexts beyond syscalls")
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
The pid member of struct audit_context is never used. Remove it.
The audit_reset_context() comment about unconditionally resetting
"ctx->state" should read "ctx->context".
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Cgroup id is user provided datum hence extend its return domain to
include possible error reason (similar to cgroup_get_from_fd()).
This change also fixes commit d4ccaf58a8 ("bpf: Introduce cgroup
iter") that would use NULL instead of proper error handling in
d4ccaf58a8 ("bpf: Introduce cgroup iter").
Additionally, neither of: fc_appid_store, bpf_iter_attach_cgroup,
mem_cgroup_get_from_ino (callers of cgroup_get_from_fd) is built without
CONFIG_CGROUPS (depends via CONFIG_BLK_CGROUP, direct, transitive
CONFIG_MEMCG respectively) transitive, so drop the singular definition
not needed with !CONFIG_CGROUPS.
Fixes: d4ccaf58a8 ("bpf: Introduce cgroup iter")
Signed-off-by: Michal Koutný <mkoutny@suse.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Cgroup ids are resolved in the global scope. That may be needed sometime
(in future) but currently it violates virtual view provided through
cgroup namespaces.
There are currently following users of the resolution:
- fc_appid_store
- bpf_iter_attach_cgroup
- mem_cgroup_get_from_ino
None of the is a called on behalf of kernel but the resolution is made
with proper userspace context, hence the default to current->nsproxy
makes sens. (This doesn't rule out cgroup_get_from_id with cgroup NS
parameter in the future.)
Since cgroup ids are defined on v2 hierarchy only, we simply check
existence in the cgroup namespace by looking at ancestry on the default
hierarchy.
Fixes: 6b658c4863 ("scsi: cgroup: Add cgroup_get_from_id()")
Signed-off-by: Michal Koutný <mkoutny@suse.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
cgroup_get_from_path() is not widely used function. Its callers presume
the path is resolved under cgroup namespace. (There is one caller
currently and resolving in init NS won't make harm (netfilter). However,
future users may be subject to different effects when resolving
globally.)
Since, there's currently no use for the global resolution, modify the
existing function to take cgroup NS into account.
Fixes: a79a908fd2 ("cgroup: introduce cgroup namespaces")
Signed-off-by: Michal Koutný <mkoutny@suse.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
There are several places in the kernel where wait_on_bit is not followed
by a memory barrier (for example, in drivers/md/dm-bufio.c:new_read).
On architectures with weak memory ordering, it may happen that memory
accesses that follow wait_on_bit are reordered before wait_on_bit and
they may return invalid data.
Fix this class of bugs by introducing a new function "test_bit_acquire"
that works like test_bit, but has acquire memory ordering semantics.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Acked-by: Will Deacon <will@kernel.org>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Daniel borkmann says:
====================
The following pull-request contains BPF updates for your *net* tree.
We've added 11 non-merge commits during the last 14 day(s) which contain
a total of 13 files changed, 61 insertions(+), 24 deletions(-).
The main changes are:
1) Fix BPF verifier's precision tracking around BPF ring buffer, from Kumar Kartikeya Dwivedi.
2) Fix regression in tunnel key infra when passing FLOWI_FLAG_ANYSRC, from Eyal Birger.
3) Fix insufficient permissions for bpf_sys_bpf() helper, from YiFei Zhu.
4) Fix splat from hitting BUG when purging effective cgroup programs, from Pu Lehui.
5) Fix range tracking for array poke descriptors, from Daniel Borkmann.
6) Fix corrupted packets for XDP_SHARED_UMEM in aligned mode, from Magnus Karlsson.
7) Fix NULL pointer splat in BPF sockmap sk_msg_recvmsg(), from Liu Jian.
8) Add READ_ONCE() to bpf_jit_limit when reading from sysctl, from Kuniyuki Iwashima.
9) Add BPF selftest lru_bug check to s390x deny list, from Daniel Müller.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
bpf_cgroup_iter_order is globally visible but the entries do not have
CGROUP prefix. As requested by Andrii, put a CGROUP in the names
in bpf_cgroup_iter_order.
This patch fixes two previous commits: one introduced the API and
the other uses the API in bpf selftest (that is, the selftest
cgroup_hierarchical_stats).
I tested this patch via the following command:
test_progs -t cgroup,iter,btf_dump
Fixes: d4ccaf58a8 ("bpf: Introduce cgroup iter")
Fixes: 88886309d2 ("selftests/bpf: add a selftest for cgroup hierarchical stats collection")
Suggested-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Hao Luo <haoluo@google.com>
Link: https://lore.kernel.org/r/20220825223936.1865810-1-haoluo@google.com
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Hsin-Wei reported a KASAN splat triggered by their BPF runtime fuzzer which
is based on a customized syzkaller:
BUG: KASAN: slab-out-of-bounds in bpf_int_jit_compile+0x1257/0x13f0
Read of size 8 at addr ffff888004e90b58 by task syz-executor.0/1489
CPU: 1 PID: 1489 Comm: syz-executor.0 Not tainted 5.19.0 #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
1.13.0-1ubuntu1.1 04/01/2014
Call Trace:
<TASK>
dump_stack_lvl+0x9c/0xc9
print_address_description.constprop.0+0x1f/0x1f0
? bpf_int_jit_compile+0x1257/0x13f0
kasan_report.cold+0xeb/0x197
? kvmalloc_node+0x170/0x200
? bpf_int_jit_compile+0x1257/0x13f0
bpf_int_jit_compile+0x1257/0x13f0
? arch_prepare_bpf_dispatcher+0xd0/0xd0
? rcu_read_lock_sched_held+0x43/0x70
bpf_prog_select_runtime+0x3e8/0x640
? bpf_obj_name_cpy+0x149/0x1b0
bpf_prog_load+0x102f/0x2220
? __bpf_prog_put.constprop.0+0x220/0x220
? find_held_lock+0x2c/0x110
? __might_fault+0xd6/0x180
? lock_downgrade+0x6e0/0x6e0
? lock_is_held_type+0xa6/0x120
? __might_fault+0x147/0x180
__sys_bpf+0x137b/0x6070
? bpf_perf_link_attach+0x530/0x530
? new_sync_read+0x600/0x600
? __fget_files+0x255/0x450
? lock_downgrade+0x6e0/0x6e0
? fput+0x30/0x1a0
? ksys_write+0x1a8/0x260
__x64_sys_bpf+0x7a/0xc0
? syscall_enter_from_user_mode+0x21/0x70
do_syscall_64+0x3b/0x90
entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7f917c4e2c2d
The problem here is that a range of tnum_range(0, map->max_entries - 1) has
limited ability to represent the concrete tight range with the tnum as the
set of resulting states from value + mask can result in a superset of the
actual intended range, and as such a tnum_in(range, reg->var_off) check may
yield true when it shouldn't, for example tnum_range(0, 2) would result in
00XX -> v = 0000, m = 0011 such that the intended set of {0, 1, 2} is here
represented by a less precise superset of {0, 1, 2, 3}. As the register is
known const scalar, really just use the concrete reg->var_off.value for the
upper index check.
Fixes: d2e4c1e6c2 ("bpf: Constant map key tracking for prog array pokes")
Reported-by: Hsin-Wei Hung <hsinweih@uci.edu>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Shung-Hsi Yu <shung-hsi.yu@suse.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/r/984b37f9fdf7ac36831d2137415a4a915744c1b6.1661462653.git.daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Precision markers need to be propagated whenever we have an ARG_CONST_*
style argument, as the verifier cannot consider imprecise scalars to be
equivalent for the purposes of states_equal check when such arguments
refine the return value (in this case, set mem_size for PTR_TO_MEM). The
resultant mem_size for the R0 is derived from the constant value, and if
the verifier incorrectly prunes states considering them equivalent where
such arguments exist (by seeing that both registers have reg->precise as
false in regsafe), we can end up with invalid programs passing the
verifier which can do access beyond what should have been the correct
mem_size in that explored state.
To show a concrete example of the problem:
0000000000000000 <prog>:
0: r2 = *(u32 *)(r1 + 80)
1: r1 = *(u32 *)(r1 + 76)
2: r3 = r1
3: r3 += 4
4: if r3 > r2 goto +18 <LBB5_5>
5: w2 = 0
6: *(u32 *)(r1 + 0) = r2
7: r1 = *(u32 *)(r1 + 0)
8: r2 = 1
9: if w1 == 0 goto +1 <LBB5_3>
10: r2 = -1
0000000000000058 <LBB5_3>:
11: r1 = 0 ll
13: r3 = 0
14: call bpf_ringbuf_reserve
15: if r0 == 0 goto +7 <LBB5_5>
16: r1 = r0
17: r1 += 16777215
18: w2 = 0
19: *(u8 *)(r1 + 0) = r2
20: r1 = r0
21: r2 = 0
22: call bpf_ringbuf_submit
00000000000000b8 <LBB5_5>:
23: w0 = 0
24: exit
For the first case, the single line execution's exploration will prune
the search at insn 14 for the branch insn 9's second leg as it will be
verified first using r2 = -1 (UINT_MAX), while as w1 at insn 9 will
always be 0 so at runtime we don't get error for being greater than
UINT_MAX/4 from bpf_ringbuf_reserve. The verifier during regsafe just
sees reg->precise as false for both r2 registers in both states, hence
considers them equal for purposes of states_equal.
If we propagated precise markers using the backtracking support, we
would use the precise marking to then ensure that old r2 (UINT_MAX) was
within the new r2 (1) and this would never be true, so the verification
would rightfully fail.
The end result is that the out of bounds access at instruction 19 would
be permitted without this fix.
Note that reg->precise is always set to true when user does not have
CAP_BPF (or when subprog count is greater than 1 (i.e. use of any static
or global functions)), hence this is only a problem when precision marks
need to be explicitly propagated (i.e. privileged users with CAP_BPF).
A simplified test case has been included in the next patch to prevent
future regressions.
Fixes: 457f44363a ("bpf: Implement BPF ring buffer and verifier support for it")
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20220823185300.406-2-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Enable bpf programs to make use of rstat to collect cgroup hierarchical
stats efficiently:
- Add cgroup_rstat_updated() kfunc, for bpf progs that collect stats.
- Add cgroup_rstat_flush() sleepable kfunc, for bpf progs that read stats.
- Add an empty bpf_rstat_flush() hook that is called during rstat
flushing, for bpf progs that flush stats to attach to. Attaching a bpf
prog to this hook effectively registers it as a flush callback.
Signed-off-by: Yosry Ahmed <yosryahmed@google.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Hao Luo <haoluo@google.com>
Link: https://lore.kernel.org/r/20220824233117.1312810-4-haoluo@google.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cgroup_iter is a type of bpf_iter. It walks over cgroups in four modes:
- walking a cgroup's descendants in pre-order.
- walking a cgroup's descendants in post-order.
- walking a cgroup's ancestors.
- process only the given cgroup.
When attaching cgroup_iter, one can set a cgroup to the iter_link
created from attaching. This cgroup is passed as a file descriptor
or cgroup id and serves as the starting point of the walk. If no
cgroup is specified, the starting point will be the root cgroup v2.
For walking descendants, one can specify the order: either pre-order or
post-order. For walking ancestors, the walk starts at the specified
cgroup and ends at the root.
One can also terminate the walk early by returning 1 from the iter
program.
Note that because walking cgroup hierarchy holds cgroup_mutex, the iter
program is called with cgroup_mutex held.
Currently only one session is supported, which means, depending on the
volume of data bpf program intends to send to user space, the number
of cgroups that can be walked is limited. For example, given the current
buffer size is 8 * PAGE_SIZE, if the program sends 64B data for each
cgroup, assuming PAGE_SIZE is 4kb, the total number of cgroups that can
be walked is 512. This is a limitation of cgroup_iter. If the output
data is larger than the kernel buffer size, after all data in the
kernel buffer is consumed by user space, the subsequent read() syscall
will signal EOPNOTSUPP. In order to work around, the user may have to
update their program to reduce the volume of data sent to output. For
example, skip some uninteresting cgroups. In future, we may extend
bpf_iter flags to allow customizing buffer size.
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Hao Luo <haoluo@google.com>
Link: https://lore.kernel.org/r/20220824233117.1312810-2-haoluo@google.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Pull another cgroup fix from Tejun Heo:
"Commit 4f7e723643 ("cgroup: Fix threadgroup_rwsem <->
cpus_read_lock() deadlock") required the cgroup
core to grab cpus_read_lock() before invoking ->attach().
Unfortunately, it missed adding cpus_read_lock() in
cgroup_attach_task_all(). Fix it"
* tag 'cgroup-for-6.0-rc2-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup: Add missing cpus_read_lock() to cgroup_attach_task_all()
Currently, verifier verifies callback functions (sync and async) as if
they will be executed once, (i.e. it explores execution state as if the
function was being called once). The next insn to explore is set to
start of subprog and the exit from nested frame is handled using
curframe > 0 and prepare_func_exit. In case of async callback it uses a
customized variant of push_stack simulating a kind of branch to set up
custom state and execution context for the async callback.
While this approach is simple and works when callback really will be
executed only once, it is unsafe for all of our current helpers which
are for_each style, i.e. they execute the callback multiple times.
A callback releasing acquired references of the caller may do so
multiple times, but currently verifier sees it as one call inside the
frame, which then returns to caller. Hence, it thinks it released some
reference that the cb e.g. got access through callback_ctx (register
filled inside cb from spilled typed register on stack).
Similarly, it may see that an acquire call is unpaired inside the
callback, so the caller will copy the reference state of callback and
then will have to release the register with new ref_obj_ids. But again,
the callback may execute multiple times, but the verifier will only
account for acquired references for a single symbolic execution of the
callback, which will cause leaks.
Note that for async callback case, things are different. While currently
we have bpf_timer_set_callback which only executes it once, even for
multiple executions it would be safe, as reference state is NULL and
check_reference_leak would force program to release state before
BPF_EXIT. The state is also unaffected by analysis for the caller frame.
Hence async callback is safe.
Since we want the reference state to be accessible, e.g. for pointers
loaded from stack through callback_ctx's PTR_TO_STACK, we still have to
copy caller's reference_state to callback's bpf_func_state, but we
enforce that whatever references it adds to that reference_state has
been released before it hits BPF_EXIT. This requires introducing a new
callback_ref member in the reference state to distinguish between caller
vs callee references. Hence, check_reference_leak now errors out if it
sees we are in callback_fn and we have not released callback_ref refs.
Since there can be multiple nested callbacks, like frame 0 -> cb1 -> cb2
etc. we need to also distinguish between whether this particular ref
belongs to this callback frame or parent, and only error for our own, so
we store state->frameno (which is always non-zero for callbacks).
In short, callbacks can read parent reference_state, but cannot mutate
it, to be able to use pointers acquired by the caller. They must only
undo their changes (by releasing their own acquired_refs before
BPF_EXIT) on top of caller reference_state before returning (at which
point the caller and callback state will match anyway, so no need to
copy it back to caller).
Fixes: 69c087ba62 ("bpf: Add bpf_for_each_map_elem() helper")
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20220823013125.24938-1-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
