[lsan] [aarch64] Fix calculating TLS area

[spetrovic]

Summary

This patch fixes calculating begin of TLS area. For example in pthread_setspecific function memory is allocated (calloc) and pointer on that memory is stored in TLS area, before this patch LSAN was skipping that part of TLS area (because begin of TLS area was not calculated correctly). Because of this memory leak was reported. I'm referencing on libc (# define THREAD_SELF \ ((struct pthread *)__builtin_thread_pointer () - 1)) in calculating begin of TLS area. This patch fixes make-check tests: TestCases/use_tls_pthread_specific_dynamic.cc, use_tls_pthread_specific_static.cc.

[peter.smith]

This looks good to me, with the caveat that I'm not an expert in this area. I've checked that descr_addr will match the glibc THREAD_SELF macro for AARCH64 so this looks sensible.

compiler-rt/trunk/lib/sanitizer_common/sanitizer_linux_libcdep.cc

# elif defined(__mips__)
   // TCB and the size of pthread_descr.
   const uptr kTlsTcbOffset = 0x7000;
   uptr thread_pointer;
   asm volatile(".set push;\
                 .set mips64r2;\
                 rdhwr %0,$29;\
                 .set pop" : "=r" (thread_pointer));
   descr_addr = thread_pointer - kTlsTcbOffset - TlsPreTcbSize();
-# elif defined(__aarch64__) || defined(__s390__)
+# elif defined(__aarch64__)
+  descr_addr = reinterpret_cast<uptr>(__builtin_thread_pointer()) -
+                                      ThreadDescriptorSize();
+# elif defined(__s390__)
   descr_addr = reinterpret_cast<uptr>(__builtin_thread_pointer());
 # elif defined(__powerpc64__)
   // PPC64LE uses TLS variant I. The thread pointer (in GPR 13)
   // points to the end of the TCB + 0x7000. The pthread_descr structure is
   // immediately in front of the TCB. TlsPreTcbSize() includes the size of the
   // TCB and the size of pthread_descr.
   const uptr kTlsTcbOffset = 0x7000;
   uptr thread_pointer;