| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
* Makefile.am: Install tapset/README too.
* Makefile.in: Regenerated.
2006-11-02 Frank Ch. Eigler <fche@elastic.org>
* README: Add a blurb against running these scripts directly.
|
|
|
|
| |
modify tapsets/LKET/register_events to change the event_desc strings
|
|
|
|
|
| |
* context.stp: Added proper handling of statement
* probe in probefunc() for IA64 (bz# 3423)
|
|
|
|
| |
user registered events
|
|
|
|
|
|
|
| |
check whether rq_disk is empty
tapset/scheduler.stp
revert changes
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
add more variables
* tapset/LKET/register_event.stp:
be sync with scsi trace hooks
* tapset/LKET/scsi.stp:
avoid reference to struct members in
embedded c. Use stap variables as more as possible.
break scsi_info into four variables.
delete cmd_pid from all logging functions
|
|
|
|
| |
skip __switch_to on x86_64
|
|
|
|
|
|
|
|
|
|
|
|
| |
* context.stp (print_backtrace): Pass in new
kretprobe instance arg.
(backtrace): Ditto.
(is_return): Rewrite.
(stack_size): New.
(stack_used): New.
(stack_unused): New.
(called_addr): New.
(caller): New.
|
|
|
|
|
|
|
|
|
|
|
| |
bugfix to avoid refer to NULL pointer
* hookid_defs.stp, ioscheduler.stp, register_event.stp:
break the old addevent.ioscheduler.elv_next_request into
addevent.ioscheduler.elv_next_request.entry and
addevent.ioscheduler.elv_next_request.return.
Now log_ioscheduler_tracedata_* will use the
variables defined in generic ioscheduler tapset
|
|
|
|
| |
make sys_request_key optional on i386
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
document signal syscall
hookiddefs.stp:
some changes to the hookid:
Now the hookid of a return type event(addevent.*.return)
should be an even number and its value should be the
corresponding entry event hookid +1.
The hookid of a non-return type event should be an odd
number.
Add the hookid of HOOKID_REGEVTDESC
register_event.stp:
add register_evt_desc() to register the description of an event.
Change the register_sys_event() to add the an argument
representing event description
Allow to use empty format string and empty field name when
calling register_event().
Change all calling to register_sys_event() to add an
argument which represents the event description.
Change register_evt_desc() and register_event() to add an argument
representing the event description string, which could help
facilitate the error debugging.
iosyscall.stp:
bugfix to add the variable GroupID and hookID
nfs.stp:
bugfix to add some omitted probe definitions.
nfs_proc.stp:
bugfix of the wrong calling to _lket_trace() for
nfs_proc.remove.entry.
|
|
|
|
|
| |
* context.stp (probefunc): Return empty string for
userspace (< PAGE_OFFSET) pointers.
|
|
|
|
| |
* test/run.sh: Changed 'stpd' references to 'staprun'.
|
|
|
|
| |
* tskdispatch.stp: Changed 'stpd' reference to 'staprun'.
|
|
|
|
| |
Change to per_cpu_ptr(Stp_pbuf, smp_processor_id()) to get the per-cpu printing buffer
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
PR 3233
./
* stapfuncs.5.in: Document new gettimeofday_ns.
runtime/
* time.c (stp_time_t): Use ns for the base time, and freq is now kHz.
(__stp_estimate_cpufreq): Compute kHz instead of MHz.
(__stp_time_timer_callback, __stp_init_time): Compute base in ns.
(__stp_time_cpufreq_callback): Record kHz instead of MHz.
(_stp_init_time): Record kHz, and disable preemption to avoid a race
in the cpufreq notifier.
(_stp_gettimeofday_ns): Converted from _stp_gettimeofday_us.
tapset/
* timestamp.stp (gettimeofday_ns): New function
(gettimeofday_us, gettimeofday_ms, gettimeofday_s):
Use gettimeofday_ns as the base unit.
testsuite/
* buildok/timestamp.stp: add gettimeofday_ns test.
|
|
|
|
|
|
|
|
|
| |
Be sync with the runtime printing changes made by Martin
lket/b2a/lket_b2a.c, lket/b2a/lket_b2a.h:
Be sync with the recent runtime printing changes made by Martin.
Some clean up work before adding the support of dumping data into MySQL
Bug fix of segment fault when an event is not registered.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
tapset/rpc.stp:
Some changes and more comments of RPC tapset
tapset/LKET/rpc.stp:
New trace hooks for RPC activities on client, server and scheduler sides
tapset/LKET/register_event.stp:
Add the register_sys_event() calls for new RPC trace hooks
tapset/LKET/hookid_defs.stp:
Add the definitions of RPC trace hooks
|
|
|
|
|
| |
* logging.stp (stp_print_binary): New function.
* string.stp: Fix docs.
|
|
|
|
|
|
|
|
| |
Make all integers passed into _lket_trace() casted into int64_t explicitly.
Delete calling to filter_by_pid()
* nfs*.stp:
Change addevent.*.entries to addevent.*.entry
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
document nfs trace hooks
tapset/nfsd.stp, tapset/nfs_proc.stp, tapset/nfs.stp:
Change NFS file identifier. Change NFS version data type
from String to INT. Add version variable to nfs_open
and nfs_release.
LKET/iosyscall.stp:
change log_iosyscall_sync() be sync with different timing methods.
LKET/utils.stp:
add reset_maxaction() since there are too many functions called inside "probe begin".
LKET/nfsd.stp, LKET/nfs_proc.stp, LKET/nfs.stp:
New trace hooks for NFS, including both NFS client and server side activities.
LKET/register_event.stp:
Add the register event calls for NFS trace hooks.
|
| |
|
|
|
|
| |
some changes to arguments and comments of signal tapset
|
|
|
|
| |
sched_clock()
|
| |
|
|
|
|
| |
to avoid compiling error
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New tapset from Thomas Zanussi(trz@us.ibm.com) to probe vfs layer activities.
nfs.stp:
New tapset from Li Xuepeng(xuepengl@cn.ibm.com) to probe nfs file operations
and nfs address space operations on client side.
nfs_proc.stp:
New tapset from Li Xuepeng to probe some nfs RPC procedure stub functions
on client side.
nfsd.stp:
New tapset from Li Xuepeng to probe nfs server side activities, including
some RPC procedure stub functions, nfsd dispatch routine, and nfsd_* functions
|
|
|
|
| |
eliminates the global stap variable to syscall tapset for faster speed
|
|
|
|
| |
get_sigset()
|
|
|
|
|
| |
* string.stp (substr): Rewrite. Make the 3rd parameter
be the length.
|
| |
|
|
|
|
|
|
|
|
| |
* signal.stp: Changes on behalf of Manoj S Pattabhiraman:
1. As per the suggestions, i have removed the argstr from the
probe points.
2. Added some checks to find whether the signals generated are
USER or Kernel Mode in signal_handle probe.
|
| |
|
|
|
|
|
| |
* ioblock.stp: Added safety checks for __bio_ino() and
__bio_start_sect().
|
| |
|
|
|
|
|
| |
* ioblock.stp: Merged io.stp from Tom Zanussi (IBM) into existing
ioblock.stp. Removed/Renamed duplicate variables and probes.
|
|
|
|
|
|
|
|
|
|
|
|
| |
* stapprobes.5.in, testsuite/buildok/memory.stp: move pagefault
to vm.* namespace
tapset/
* memory.stp: move pagefault to vm.* namespace, and add many other
virtual-memory themed probes.
tapset/LKET/
* pagefault.stp: move pagefault to vm.* namespace
|
|
|
|
|
|
|
|
|
| |
* tapset/signal.stp: Create a new tapset that addresses process signals.
Much of this was contributed by Manoj Pattabhiraman (IBM).
* tapset/process.stp: Remove aliases that now belong in signal tapset
* examples/small_demos/proc_snoop.stp,
testsuite/buildok/process_test.stp: Rename process.signal_* to
new signal.* tapset.
|
|
|
|
|
| |
* syscalls.stp: Fixed typo in syscall.kexec_load argument
handling.
|
|
|
|
|
|
| |
* tapset/context.stp (probemod): New function.
* stapfuncs.5.in: Document it.
* testsuite/buildok/probemod.stp: Test it.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
* context.stp: Removed returnval() function. Use $return
instead.
* return.stp: Removed. Use $return instead.
|
| |
|
|
|
|
|
|
|
|
| |
update lket man page to refect the new tapset search path
syscalls.stp: make sys_acct, sys_add_key, sys_keyctl and sys_modify_ldt optional on ppc64 since they are weak symbol.
ppc64/syscalls.stp: remove syscall.acct, syscall.add_key and syscall.umask since they are already defined in syscalls.stp
|
| |
|