[#4195] [ltp] LTP test case gettimeofday02 fails sometimes in smp kernel

Submitted By: Vivi Li

Open Date

2008-06-24 23:13:55 Close Date

2008-08-11 23:04:59

Priority:

Medium High Assignee:

Graf Yang

Status:

Closed Fixed In Release:

N/A

Found In Release:

N/A Release:

Category:

N/A Board:

N/A

Processor:

N/A Silicon Revision:

Is this bug repeatable?:

Yes Resolution:

Fixed

Uboot version or rev.:

Toolchain version or rev.:

08r1.5-11

App binary format:

N/A

Summary: [ltp] LTP test case gettimeofday02 fails sometimes in smp kernel

Details:

Ltp test case gettimeofday02 sometimes fails in smp kernel.

It always passed once then failed once.

This bug happens with all options of wt_slab_buddy, wt_slab_np2 and wt_slob

This ltp case is not compiled with -mstack-check-l1 option.

root:/> gettimeofday02

gettimeofday02 0 INFO : checking if gettimeofday is monotonous, takes 30s

gettimeofday02 1 FAIL : Time is going backwards: old 1167609708.883994 vs new 1167609708.880065!

root:/>

root:/> gettimeofday02

gettimeofday02 0 INFO : checking if gettimeofday is monotonous, takes 30s

gettimeofday02 1 PASS : gettimeofday monotonous in 30 seconds

root:/>

Follow-ups

--- Graf Yang 2008-07-15 03:49:28

Fixed, when codes running on CoreB calling do_gettimeofday() or gettimeoffset(),

we call smp_call_function() to let CoreA call these functions and return the

value.

To prevent the rare case,

https://blackfin.uclinux.org/gf/project/uclinux-dist/tracker/?action=TrackerItemEdit&tracker_item_id=4241

, I add some compare code at the end of do_gettimeofday()

--- Vivi Li 2008-07-22 03:19:22

OK now. Close it. Thanks!

--- Mike Frysinger 2008-07-30 06:17:17

erm, re-open because it seems to cause [#4283]

--- Graf Yang 2008-08-01 07:40:12

Fixed, we use GPTIMER0 to send tick, instead of the core timer.

--- Vivi Li 2008-08-12 00:05:03

Ok now. Close it.

Thanks!

Files

Changes

Commits

Dependencies

Duplicates

Associations