It turns out that clang can emit code where the sp is saved in the r7
register on arm. Unfortunately, a lot of our syscalls overwrite that
value while the syscall is executing, so unwinding through that syscall
fails.
Update the syscall generation code to add unwinding information for
these uses.
Bug: 28411713
Change-Id: Id51a5d43c058631bb50834527f93fa299bbcf759
Changes affecting future time stamps
America/Caracas switches from -0430 to -04 on 2016-05-01 at 02:30.
(Thanks to Alexander Krivenyshev for the heads-up.)
Asia/Magadan switches from +10 to +11 on 2016-04-24 at 02:00.
(Thanks to Alexander Krivenyshev and Matt Johnson.)
New zone Asia/Tomsk, split off from Asia/Novosibirsk. It covers
Tomsk Oblast, Russia, which switches from +06 to +07 on 2016-05-29
at 02:00. (Thanks to Stepan Golosunov.)
Changes affecting past time stamps
New zone Europe/Kirov, split off from Europe/Volgograd. It covers
Kirov Oblast, Russia, which switched from +04/+05 to +03/+04 on
1989-03-26 at 02:00, roughly a year after Europe/Volgograd made
the same change. (Thanks to Stepan Golosunov.)
Russia and nearby locations had daylight-saving transitions on
1992-03-29 at 02:00 and 1992-09-27 at 03:00, instead of on
1992-03-28 at 23:00 and 1992-09-26 at 23:00. (Thanks to Stepan
Golosunov.)
Many corrections to historical time in Kazakhstan from 1991
through 2005. (Thanks to Stepan Golosunov.) Replace Kazakhstan's
invented time zone abbreviations with numeric abbreviations.
Bug: 28240832
Change-Id: If11e29f064f074f8b5812d397996322d007ab1d2
android_net_res_stats_get_info_for_net returns the current name servers, search
domains, parameters and stats for the given network ID.
android_net_res_stats_aggregate provides statistics such as errors counts from
the raw stats data reported by android_net_res_stats_get_info_for_net.
android_net_res_stats_get_usable_servers uses the data returned by
android_net_res_stats_aggregate to determine which of the servers are
considered valid or broken by the resolver.
BUG: 25731675
Change-Id: I6059b68e5e8b809027a4d3135f6081588bee8a7d
Instead of keeping a sentinel after nameservers[], nsaddrinfo[] and
nstats[], store the server count in the structure, freeing up memory and
eliminating the need to enumerate the server count every time
_resolv_is_nameservers_equal_locked() is invoked.
Also increase MAXNS from 3 to 4.
BUG: 28153323
Change-Id: I11a7257af695157c9e32019cd00c67b535b63c75
(cherry picked from commit fff356786f)
I changed the output from "pointer" to "allocation" for allocation
mismatches. I forgot to update the tests.
Bug: 28218530
(cherry picked from commit 97fd7d95f5)
Change-Id: I4076e4ddb5d2385e786e3ebca187d1e84e3c172c
- Change the printed name from pointer to allocation to be explicit about
what is wrong.
- Change the signal to be SIGRTMAX - 19 instead of SIGRTMIN. This should
prevent problems if we have to reserve other real time signals.
Bug: 28218530
(cherry picked from commit ea26b33248)
Change-Id: I707d5200c1804bdbf45fc46de0f2565016c48a82
Changes affecting future time stamps
Azerbaijan no longer observes DST. (Thanks to Steffen Thorsen.)
Chile reverts from permanent to seasonal DST. (Thanks to Juan
Correa for the heads-up, and to Tim Parenti for corrections.)
Guess that future transitions are August's and May's second
Saturdays at 24:00 mainland time. Also, call the period from
2014-09-07 through 2016-05-14 daylight saving time instead of
standard time, as that seems more appropriate now.
Changes affecting past time stamps
Europe/Kaliningrad and Europe/Vilnius changed from +03/+04 to
+02/+03 on 1989-03-26, not 1991-03-31. Europe/Volgograd changed
from +04/+05 to +03/+04 on 1988-03-27, not 1989-03-26.
(Thanks to Stepan Golosunov.)
Bug: 27826144
Change-Id: I4353fa075e6cc0f277a72c928442a63835ec4d97
linker didn't remove link to the soinfo from shared
namespaces on soinfo_unload, because it didn't keep
record of all namespaces the library is added to.
This change adds test for this and also fixes the
problem by introducing list of secondary namespaces
to soinfo, which is used to remove soinfo in
soinfo::remove_all_links().
Bug: http://b/28115950
Change-Id: Ifbf6e54f92fa6e88f86b6a8dd6dc22d4553afd22
(cherry picked from commit aca299ac47)
Linker was not unloading libraries marked with RTLD_NODELETE
even when there was a error on load. For example when one of
dt_needed libraries was not found.
Bug: http://b/27911891
Change-Id: Ibc0bf110aa41fbfdc9f327ad8c364f2e3600af71
Collect statistics about DNS query success state and delay. Ignore
servers that have been tried at least five times and have a success rate
of < 0.25. Retry these servers once every 180s.
Bug: 25731675
(cherry picked from commit 6b3f0d65f2)
Change-Id: Id989404b14fd885fb31a5a4de36a4186be1c9ad1
Built for fugu this was working fine, but built for generic x86/x86-64
the compiler was (a) optimizing out all the stack writes and (b) inserting
enough padding on x86-64 for the canary to be safely out of the way.
While here, let's tighten up this test so we test that it's sufficient to
only overwrite the buffer by one byte...
Bug: http://b/27815668
Change-Id: I80a646de4b30fd5c78df20fdaa7e3eb163585caf
Using single quotes can be more convenient -- because there's no need to
escape -- but the vast majority of existing quoting was done with double
quotes, so move in that direction.
Bug: http://b/27524191
Change-Id: I8f58593120e4c8aeaec45e16fc0178e9c162454d
Previously, the implementation of setjmp on x86_64 claimed that
sigprocmask would write to two longs' worth of bytes.
Bug: http://b/27856501
Change-Id: I9f32b40ac773a0cd91a976aace5bfba6e67fb0f8
(cherry picked from commit c244fcb8a3)
This patch uses __kernel_vsyscall instead of "int 0x80"
as the syscall entry point. AT_SYSINFO points to
an adapter to mask the arch specific difference and gives a
performance boost on i386 architecture.
Bug: http://b/27533895
Change-ID: Ib340c604d02c6c25714a95793737e3cfdc3fc5d7
Signed-off-by: Mingwei Shi <mingwei.shi@intel.com>
(cherry picked from commit be91052932)
The DNS search string contains zeros as domain separator. The resolver
code erroneously used strlcpy(), which resulted in only the first domain
to be copied. The code uses pointers into this string to access the
individual domains. Since the structure is zero-initialized, this bug only
resulted in zero-length domains instead of accessing unitialized memory.
BUG: 27312811
Change-Id: Ia9d066c405dfcc5e82d6766d93ead2ce574e7b0d
(cherry picked from commit 0967fc7e59)
* changes:
Do not add linker the _r_debug.r_map on link_image
Improvements to dynamic linker messages
Move gdb support functions to a separate file
Use insert_link_map_into_debug_map for executable
Update a comment for init_linker_info_for_gdb
Initialize linker link_map for gdb directly
Handle no longer is a pointer to soinfo of
a corresponding library. This is done to
prevent access to linker internal fields.
Bug: http://b/25593965
Change-Id: I62bff0d0e5b2dc842e6bf0babb30fcc4c000be24
(cherry picked from commit d88e1f3501)
Error messages now include executable name,
also linker doesn't abort if called on itself.
Bug: http://b/27533895
Change-Id: Ia17dd540919544b28062ed71751a1d79ee865206
(cherry picked from commit 9f0a6954c6)
Move gdb support functions and variables to
linker_gdb_support.h/cpp
Bug: http://b/27533895
Change-Id: I96c6592a7055715b18f1137367470fe80987263f
(cherry picked from commit 6b788eeff2)
Use insert_link_map_into_debug_map to insert the main
executable's link_map to r_debug
Bug: http://b/27533895
Change-Id: I0eacb3f030ea3eb16ed50ad2011d604beece2d03
(cherry picked from commit f3064e4bc7)
Remove unnecessary construction of soinfo when
initializing linker link_map for gdb.
Bug: http://b/27533895
Change-Id: Idf32cee56309aa9c9cf260efbd17a9deae9a756b
(cherry picked from commit 8d22dd53fe)
pthread_barrier_smoke test uses WaitUntilThreadSleep() to wait until
BarrierTestHelper threads sleep in pthread_barrier_wait(). But this
is flaky as there a two futex_wait places in pthread_barrier_wait.
This patch modifies this test to avoid using WaitUntilThreadSleep().
Bug: 27780937
Bug: 27811377
Change-Id: I4c36b82cce9345d5088f8854b289dc5bf7a08e8c
(cherry picked from commit 81d2797e33)
"I have none of these capabilities" is a legitimate response, and
distinct from "I don't know what my capabilities are". It's the latter
I meant to test for.
Bug: http://b/27810985
(cherry picked from commit d4c9112490)
Change-Id: I4093948039a658e926e9b5e28d19b1344c70bd0e
When using libhoudini to run arm code on x86 platforms, we can't
assume the main thread allocates local variables at the stack
declared by kernel.
(cherry picked from commit 61e4d461e5)
Bug: http://b/27815309
Change-Id: Ic714a8e633a0bb6e85b38ee0e01e3439d0a0cea5
The bionic change merged down from upstream but the
external/icu one didn't.
A replacement change will be submitted along with one for
external/icu.
This reverts commit 7bcee86edc.
Change-Id: I0dd28e02231e6ebba5b61131b22889a1268a0675
Changes affecting future time stamps
New zones Europe/Astrakhan and Europe/Ulyanovsk for Astrakhan and
Ulyanovsk Oblasts, Russia, both of which will switch from +03 to +04 on
2016-03-27 at 02:00 local time. They need distinct zones since their
post-1970 histories disagree. New zone Asia/Barnaul for Altai Krai and
Altai Republic, Russia, which will switch from +06 to +07 on the same date
and local time. Also, Asia/Sakhalin moves from +10 to +11 on 2016-03-27
at 02:00. (Thanks to Alexander Krivenyshev for the heads-up, and to
Matt Johnson and Stepan Golosunov for followup.)
As a trial of a new system that needs less information to be made up,
the new zones use numeric time zone abbreviations like "+04"
instead of invented abbreviations like "ASTT".
Haiti will not observe DST in 2016. (Thanks to Jean Antoine via
Steffen Thorsen.)
Palestine's spring-forward transition on 2016-03-26 is at 01:00, not 00:00.
(Thanks to Hannah Kreitem.) Guess future transitions will be March's last
Saturday at 01:00, not March's last Friday at 24:00.
Changes affecting past time stamps
Europe/Chisinau observed DST during 1990, and switched from +04 to
+03 at 1990-05-06 02:00, instead of switching from +03 to +02.
(Thanks to Stepan Golosunov.)
1991 abbreviations in Europe/Samara should be SAMT/SAMST, not
KUYT/KUYST. (Thanks to Stepan Golosunov.)
The update produced (apparently benign) warnings from zic as suggested here:
http://mm.icann.org/pipermail/tz-announce/2016-March/000036.html
Bug: 27656428
Change-Id: I31c7805b6fdb458c3ee9e695b3dff7bd003d8cab
Changes affecting future time stamps
New zones Europe/Astrakhan and Europe/Ulyanovsk for Astrakhan and
Ulyanovsk Oblasts, Russia, both of which will switch from +03 to +04 on
2016-03-27 at 02:00 local time. They need distinct zones since their
post-1970 histories disagree. New zone Asia/Barnaul for Altai Krai and
Altai Republic, Russia, which will switch from +06 to +07 on the same date
and local time. Also, Asia/Sakhalin moves from +10 to +11 on 2016-03-27
at 02:00. (Thanks to Alexander Krivenyshev for the heads-up, and to
Matt Johnson and Stepan Golosunov for followup.)
As a trial of a new system that needs less information to be made up,
the new zones use numeric time zone abbreviations like "+04"
instead of invented abbreviations like "ASTT".
Haiti will not observe DST in 2016. (Thanks to Jean Antoine via
Steffen Thorsen.)
Palestine's spring-forward transition on 2016-03-26 is at 01:00, not 00:00.
(Thanks to Hannah Kreitem.) Guess future transitions will be March's last
Saturday at 01:00, not March's last Friday at 24:00.
Changes affecting past time stamps
Europe/Chisinau observed DST during 1990, and switched from +04 to
+03 at 1990-05-06 02:00, instead of switching from +03 to +02.
(Thanks to Stepan Golosunov.)
1991 abbreviations in Europe/Samara should be SAMT/SAMST, not
KUYT/KUYST. (Thanks to Stepan Golosunov.)
The update produced (apparently benign) warnings from zic as suggested here:
http://mm.icann.org/pipermail/tz-announce/2016-March/000036.html
Bug: 27656428
Change-Id: Ibbf7237344c21d22a0f781f547fbc95b46623ece
This is important for performance of 32-bit ARM crypto code in apps.
Bug: http://b/27750752
(cherry picked from commit ebb4895c68)
Change-Id: Iec69fe57b93da0480ef3bbd4c9e7882c8aa55011
