GetSigtrampRange DCHECKs always on 10.11 |
||
Issue description10.11.6 (15G1611) I always get [33270:1295:0912/153905.448096:FATAL:native_stack_sampler_mac.cc(339)] Check failed: info.start_ip == address (140735348958457 vs. 140735348958480) 0 libbase.dylib 0x00000001231ee04e base::debug::StackTrace::StackTrace(unsigned long) + 174 1 libbase.dylib 0x00000001231ee10d base::debug::StackTrace::StackTrace(unsigned long) + 29 2 libbase.dylib 0x00000001231ec3bc base::debug::StackTrace::StackTrace() + 28 3 libbase.dylib 0x000000012328bcdf logging::LogMessage::~LogMessage() + 479 4 libbase.dylib 0x0000000123289645 logging::LogMessage::~LogMessage() + 21 5 libbase.dylib 0x00000001233bf53f base::(anonymous namespace)::GetSigtrampRange(unsigned long*, unsigned long*) + 447 on every launch. It's 0x7FFF807B64F9 vs 0x7FFF807B6510.
,
Sep 12 2017
dwarfdump says: start_addr: 0x0000000000002510 __sigtramp So it seems like libunwind is falling down somewhere
,
Sep 13 2017
For posterity: 10.11 indeed seems to be pulling info for __sigunaltstack. Adding 1 to sigtramp's address to avoid the boundary condition seems to work, so that's the planned workaround.
,
Sep 14 2017
Forgot to add this bug to https://chromium-review.googlesource.com/c/chromium/src/+/665501 Should be fixed at ToT
,
Sep 14 2017
Confirmed. |
||
►
Sign in to add a comment |
||
Comment 1 by lgrey@chromium.org
, Sep 12 2017