Since 'struct timespec' members (time_t and long) are both 32bit on
32bit systems, and std::chrono::{seconds,nanoseconds}::rep are both
>32bit, timespec members assigned in DurationToTimeSpec() can have a
negative value, especially when WaitForProperty() is called with the
default timeout value which is std::chrono::milliseconds::max().
Regarding functionality, passing a negative value to
__system_property_wait() is okay because WaitForProperty() still
waits for the property value (so unit tests are passing), but while
WaitForProperty() does that, the function, to be more exact,
SystemProperties::Wait() in bionic/, consumes ~100% of CPU time. This
happens because SystemProperties::Wait() which implements
__system_property_wait() has a tight while-loop with a __futex_wait()
call, and the futex call immediately returns EINVAL when the timespec
passed in has a negative value.
With this CL, WaitForProperty() will never pass a negative timespec
to __system_property_wait(), and therefore the __futex_wait() call
in bionic works as expected without consuming too much CPU time even
on 32bit systems.
Bug: None
Test: libbase_test32 still passes
Test: strace no longer shows repeated EINVALs from __futex_wait
Change-Id: Id1834fac8cd2876b02dbe4479bf3d3eda2fa7da1
This is needed if they will ever handle ro. properties that have
values longer than 92 characters.
Bug: 23102347
Bug: 34954705
Test: read and write properties with value length > 92 characters
Change-Id: I44aa135c97ec010f12162c30f743387810ae2c5d
std::chrono doesn't handle integer overflow, so using
std::chrono::milliseconds::max() to indicate an infinite timeout is
not handled well in the current code. It causes an 'absolute_timeout'
earlier in time than 'now' and causes the associated WaitForProperty*
functions to return immediately.
Also, any duration_cast from relative_timeout to nanoseconds would
cause the same issue, as it would overflow in the conversion and
result in an invalid results.
This change prevents any duration_casts of relative_timeout to
nanoseconds and replaces the logic to wait on an absolute timeout with
logic that compares the time elapsed to the provided relative timeout.
This change also includes a test that std::chrono::milliseconds::max()
does not return immediately and that negative values do return immediately.
Test: Boot bullhead + libbase_test
Change-Id: I335bfa7ba71e86c20119a0ed46014cad44361162
- unlike base::WaitForProperty, which waits for specific value to
be set, this one only waits until the property is created.
bug: 35178781
Test: added unit test
Change-Id: Idbf98c2152fe768357302f6b69310c55305f5d54
Also adapt libcutils to the bionic change that was necessary for this.
Bug: http://b/35201172
Test: ran tests
Change-Id: I72a98b70b03d23e958b46778b505fbd5c86c32ae
Makes it easier to write correct code in a world where the maximum
property key/value lengths change.
Bug: http://b/23102347
Test: libbase_test64
Change-Id: I100f00904221bbcef9e8786a4e6e30428039bb49
