[lldb] [test] Synchronize before the breakpoint in fork testsWe set breakpoint on child_func, so synchronization inside it is toolate to guarantee ordering between the parent output and childbrea
[lldb] [test] Synchronize before the breakpoint in fork testsWe set breakpoint on child_func, so synchronization inside it is toolate to guarantee ordering between the parent output and childbreakpoint. Split the function in two, and perform synchronizationbefore the breakpoint.Differential Revision: https://reviews.llvm.org/D109591
show more ...
[lldb] [test] Add synchronization fix Subprocess test flakinessAdd synchronization routines to ensure that Subprocess tests outputin a predictable order, and all test strings are output before the
[lldb] [test] Add synchronization fix Subprocess test flakinessAdd synchronization routines to ensure that Subprocess tests outputin a predictable order, and all test strings are output before the teststerminate.Differential Revision: https://reviews.llvm.org/D109495
Reland "[lldb] [Process] Watch for fork/vfork notifications" for FreeBSDThe original commit was reverted because of the problems it introducedon Linux. However, FreeBSD should not be affected, so
Reland "[lldb] [Process] Watch for fork/vfork notifications" for FreeBSDThe original commit was reverted because of the problems it introducedon Linux. However, FreeBSD should not be affected, so restore that partand we will address Linux separately.While at it, remove the dbreg hack as the underlying issue has beenfixed in the FreeBSD kernel and the problem is unlikely to happenin real life use anyway.Differential Revision: https://reviews.llvm.org/D98822
Revert "[lldb] [Process] Watch for fork/vfork notifications" and associated followupsThis commit has caused the following tests to be flaky:TestThreadSpecificBpPlusCondition.pyTestExitDuringExpre
Revert "[lldb] [Process] Watch for fork/vfork notifications" and associated followupsThis commit has caused the following tests to be flaky:TestThreadSpecificBpPlusCondition.pyTestExitDuringExpression.pyThe exact cause is not known yet, but since both tests deal withthreads, my guess is it has something to do with the tracking ofcreation of new threads (which the commit touches upon).This reverts the following commits:d01bff8cbdc98fb8751f7bf10af19b47ae5c445d,ba62ebc48e8c424ce3a78ba01acda679d536dd47,e761b6b4c58d4f7ae1073d925d7cb321d68ee93a,a345419ee03095c8cdfbe1c2728467c4da8fa0a4.
[lldb] [Process] Watch for fork/vfork notificationsWatch for fork(2)/vfork(2) (also fork/vfork-style clone(2) on Linux)notifications and explicitly detach the forked child process, and addinitial
[lldb] [Process] Watch for fork/vfork notificationsWatch for fork(2)/vfork(2) (also fork/vfork-style clone(2) on Linux)notifications and explicitly detach the forked child process, and addinitial tests for these cases. The code covers FreeBSD, Linuxand NetBSD process plugins. There is no new user-visible functionalityprovided -- this change lays foundations over subsequent work on forksupport.Differential Revision: https://reviews.llvm.org/D98822