summaryrefslogtreecommitdiffstats
path: root/runtime
Commit message (Collapse)AuthorAgeFilesLines
...
* PR7016: Check free memory before allocating buffers.Masami Hiramatsu2008-11-132-0/+26
|
* PR6964: configurably revert to signal-based "stap -c CMD" startup ↵Frank Ch. Eigler2008-11-122-29/+69
| | | | synchronization
* reduce transport debug verbosityFrank Ch. Eigler2008-11-122-2/+6
|
* PR5274 - Added s390 task_finder support.David Smith2008-11-042-0/+51
| | | | | | | 2008-11-04 David Smith <dsmith@redhat.com> PR 5274. * syscall.h: Added s390 task_finder support.
* PR5274 - Add s390 arch support for uprobes longjump handling.David Smith2008-11-042-0/+18
| | | | | | | | 2008-11-04 David Smith <dsmith@redhat.com> From Jim Keniston, PR 5274. * uprobes/uprobes_s390.h: Add s390 arch support for uprobes longjump handling.
* PR6028 / BZ468143: ia64 fetch-register speedupMasami Hiramatsu2008-11-042-2/+16
|
* Add a version of uprobes that works with the 2.6.27 version of utrace.Jim Keniston2008-11-039-2/+4085
|
* PR6964: temporary workaround for fedora utrace kernel bug 467568Wenji Huang2008-10-282-0/+25
|
* PR6967: fix s390x compilation error re on_each_cpuFrank Ch. Eigler2008-10-282-2/+24
|
* PR5274: uretprobes fixes, belated commitSrikar Dronamraju2008-10-289-4/+112
|
* Fix hrtimer.expires name change.Josh Stone2008-10-242-0/+10
| | | | | | The field is renamed to _expires starting in 2.6.28-rc1, and it's not meant to be directly manipulated. Instead hrtimer users are supposed to use hrtimer_get_expires and hrtimer_set_expires.
* BZ467652: support old-style stap -c "FOO > BAR" usage by backing down to ↵Frank Ch. Eigler2008-10-231-7/+27
| | | | sh -c if needed.
* PR6923: Make on_each_cpu() autoconf test work on old kernels.Jim Keniston2008-10-171-1/+7
|
* Fix compilation warning of uninitialized value in gcc 3.x.Wenji Huang2008-10-172-1/+5
|
* fix wordexp error handling, for invalid shell-like stap -c "command > file"Frank Ch. Eigler2008-10-152-3/+13
| | | | Reported-By: Wade Mealing <wmealing@redhat.com>
* PR4886: weaken build-id failure semanticsFrank Ch. Eigler2008-10-075-49/+50
|
* Add workaround for fedora 9 2.6.26 kernels to utrace_compatibility.h.Mark Wielaard2008-10-062-1/+26
|
* Add uprobes.txt.Jim Keniston2008-10-061-0/+675
|
* PR4886: check build-id if able.Wenji Huang2008-10-065-1/+101
| | | | This provides sanity check of debuginfo file based on build-id. Many cases are considered, whether build-id exists in debuginfo file or not, whether module is loaded or not, whether build-id exists in runtime kernel/module. It will do work when LD >= 2.18 and kernel >=2.6.23, otherwise no check.
* Merge branch 'master' of ssh://kenistoj@sources.redhat.com/git/systemtapJim Keniston2008-10-032-1/+17
|\
| * Remove debug output from warning in procfs.Mark Wielaard2008-10-042-2/+6
| |
| * procfs dir_entry count warning on removal depends on kernel version.Mark Wielaard2008-10-042-3/+15
| |
* | PR 6850Jim Keniston2008-10-031-28/+148
|/ | | | | Fix handling of fork with uretprobe_instances outstanding. Add regression test.
* fix bugs in add_bad_addr_entryTim Moore2008-10-022-2/+8
|
* Fix race condition in addr-map; simplify allocation logicTim Moore2008-09-302-46/+58
|
* Fixed __STP_ATTACHED_TASK_EVENTS problem caused by last commit.David Smith2008-09-262-1/+5
| | | | | | | 2008-09-26 David Smith <dsmith@redhat.com> * task_finder.c (__STP_ATTACHED_TASK_EVENTS): Removed UTRACE_STOP, which isn't needed anymore.
* New utrace fixes.David Smith2008-09-253-29/+121
| | | | | | | | | | | | | | | | 2008-09-25 David Smith <dsmith@redhat.com> * task_finder.c (__stp_utrace_attach): Added action flag to know to request the thread to be stopped or not. (stap_utrace_attach): Now just calls __stp_utrace_attach(). (__stp_utrace_task_finder_target_quiesce): Handles utrace_set_events() errors properly. * utrace_compatibility.h (enum utrace_resume_action): Added utrace_resume_action enum. (utrace_control): Added UTRACE_STOP support. (utrace_engine_put): New. (utrace_barrier): New.
* Fix PR 6903 by checking for the real uid, not the euid.David Smith2008-09-182-1/+7
| | | | | | | | 2008-09-18 David Smith <dsmith@redhat.com> PR 6903. * staprun_funcs.c (check_permissions): Instead of checking the effective uid, check the real uid for root permissions.
* PR 6487, 6504: robustify __stp_print_flushFrank Ch. Eigler2008-09-175-3/+145
|
* PR6879: uprobe support for 0x64/0x65 segment prefix instructionsJim Keniston2008-09-173-12/+15
|
* fix bspcache regression typoFrank Ch. Eigler2008-09-121-1/+1
|
* PR6028: use ip instead of pp for probe identification.Masami Hiramatsu2008-09-122-4/+10
|
* BZ 6755 fix. Wait until all callbacks are finished.David Smith2008-09-122-7/+70
| | | | | | | | | | | | | | | | | | 2008-09-12 David Smith <dsmith@redhat.com> BZ 6755 * task_finder.c: Added 'inuse' count to know when handlers are still running. (__stp_utrace_task_finder_report_clone): If state isn't correct, detach. Increase 'inuse' count when starting, decrement when exiting. (__stp_utrace_task_finder_report_exec): Ditto. (__stp_utrace_task_finder_target_death): Ditto. (__stp_utrace_task_finder_target_quiesce): Increase 'inuse' count when starting, decrement when exiting. (__stp_utrace_task_finder_target_syscall_entry): Ditto. (__stp_utrace_task_finder_target_syscall_exit): Ditto. (stap_stop_task_finder): Wait until all callbacks are finished.
* backtrace fix for i386 with neither frame pointers nor dwarf unwindingFrank Ch. Eigler2008-09-103-4/+14
|
* Rename _stp_module module_base output to dwarf_module_base and document.Mark Wielaard2008-09-103-4/+14
|
* Enable usage of STP_USE_DWARF_UNWINDER for i386 and arm for debugging purposes.Mark Wielaard2008-09-102-0/+7
|
* Record module base and use it for adjusting start locations in dwarf unwinder.Mark Wielaard2008-09-103-3/+33
|
* Feed dwarf unwinder data through _stp_mod_sec_lookup (), but disabled for now.Mark Wielaard2008-09-104-4/+15
|
* Add new function _stp_mod_sec_lookup extracted from _stp_kallsyms_lookup.Mark Wielaard2008-09-102-23/+37
|
* Utrace on ia64 fast fetch-register supportMasami Hiramatsu2008-09-094-17/+38
|
* Add ia64 utrace supportMasami Hiramatsu2008-09-092-1/+50
|
* Correct this_section_offset calculation in _stp_kallsyms_lookup.Mark Wielaard2008-09-092-1/+6
|
* PR1288: runtime functions for avoiding certain addressesTim Moore2008-09-084-94/+326
|
* PR6445: documentation, mopping upFrank Ch. Eigler2008-09-062-0/+12
|
* "stap -c" cleanup part 2 - use wordexp / execvp instead of "sh -c" to ↵Frank Ch. Eigler2008-09-061-28/+30
| | | | execute cmd string
* task_finder <-> target_pid coupling; staprun/target_cmd fork under ptrace ↵Frank Ch. Eigler2008-09-052-300/+335
| | | | control
* remove capabilities logic, cont'dFrank Ch. Eigler2008-09-054-91/+73
|
* remove capability logicFrank Ch. Eigler2008-09-051-166/+0
| | | | | It was only barely beneficial anyway, since some crucial capabilities were never permanently dropped.
* PR4225: post-merge changelogsFrank Ch. Eigler2008-09-011-0/+5
|
* Merge commit 'origin/master' into pr4225Frank Ch. Eigler2008-09-012-46/+82
|\ | | | | | | | | | | | | | | | | | | | | * commit 'origin/master': Fix memory access error in nfs.proc.read_setup, nfs.proc.write_setup and nfs.proc.commit_setup Update NEWS regarding the systemtap client/server. New man page for the systemtap client/server and related utilities. Repopulate symbol/type info. Pushed quiesce logic down into the task_finder layer. Added bug 6841 fix utrace syscall test. PR6731: improve listing mode to list variables