1 =======================
2 IRQ-flags state tracing
3 =======================
5 :Author: started by Ingo Molnar <mingo@redhat.com>
7 The "irq-flags tracing" feature "traces" hardirq and softirq state, in
8 that it gives interested subsystems an opportunity to be notified of
9 every hardirqs-off/hardirqs-on, softirqs-off/softirqs-on event that
10 happens in the kernel.
12 CONFIG_TRACE_IRQFLAGS_SUPPORT is needed for CONFIG_PROVE_SPIN_LOCKING
13 and CONFIG_PROVE_RW_LOCKING to be offered by the generic lock debugging
14 code. Otherwise only CONFIG_PROVE_MUTEX_LOCKING and
15 CONFIG_PROVE_RWSEM_LOCKING will be offered on an architecture - these
16 are locking APIs that are not used in IRQ context. (the one exception
17 for rwsems is worked around)
19 Architecture support for this is certainly not in the "trivial"
20 category, because lots of lowlevel assembly code deal with irq-flags
21 state changes. But an architecture can be irq-flags-tracing enabled in a
22 rather straightforward and risk-free manner.
24 Architectures that want to support this need to do a couple of
25 code-organizational changes first:
27 - add and enable TRACE_IRQFLAGS_SUPPORT in their arch level Kconfig file
29 and then a couple of functional changes are needed as well to implement
30 irq-flags-tracing support:
32 - in lowlevel entry code add (build-conditional) calls to the
33 trace_hardirqs_off()/trace_hardirqs_on() functions. The lock validator
34 closely guards whether the 'real' irq-flags matches the 'virtual'
35 irq-flags state, and complains loudly (and turns itself off) if the
36 two do not match. Usually most of the time for arch support for
37 irq-flags-tracing is spent in this state: look at the lockdep
38 complaint, try to figure out the assembly code we did not cover yet,
39 fix and repeat. Once the system has booted up and works without a
40 lockdep complaint in the irq-flags-tracing functions arch support is
42 - if the architecture has non-maskable interrupts then those need to be
43 excluded from the irq-tracing [and lock validation] mechanism via
44 lockdep_off()/lockdep_on().
46 In general there is no risk from having an incomplete irq-flags-tracing
47 implementation in an architecture: lockdep will detect that and will
48 turn itself off. I.e. the lock validator will still be reliable. There
49 should be no crashes due to irq-tracing bugs. (except if the assembly
50 changes break other code by modifying conditions or registers that