public inbox for kernel-testers@vger.kernel.org
 help / color / mirror / Atom feed
* SYNC - takes VERY  long complete  ( system freezes) !
@ 2017-04-18  8:43 Vellemans, Noel
       [not found] ` <D24E577EA1E37B4788A3BFBC418F14A70750D5DD-bUMf7dFpk4BvAH7Z6HRyZmXnswh1EIUO@public.gmane.org>
  0 siblings, 1 reply; 2+ messages in thread
From: Vellemans, Noel @ 2017-04-18  8:43 UTC (permalink / raw)
  To: kernel-testers-u79uwXL29TY76Z2rM5mHXA@public.gmane.org

Hi all,

I've been upgrading to a recent kernel ( for one of our boards).  { Running on a 1- Ghz ARM single core CPU} 
And I'm having an issue that keeps me busy for some days.
"sync" seems to take VERY long to complete. ( the whole system seems to freeze, so that finally the WATDOG triggers a reset) 

There is no disk activity ( logs are stored on tmpfs ) ... .!
 
On a recent 4.4.61  kernel  the time measured for the "sync"  seems often to be more than 5 seconds ( up to 30 seconds). 
# /usr/bin/time  sync;

real    0m 8.01s
user    0m 0.00s
sys     0m 0.01s

{ fyi: system load with 'TOP' is less than 2%,   I've been trying 4.4.43 and 4.4.61 .. both have the same 'behavoir'}


On an older 2.6.35.x kernel ( the kernel has been rebuild)  I have this behavior.
# /usr/bin/time  sync;

real    0m 0.01s
user    0m 0.00s
sys     0m 0.01s


I'm using  same root-filesystem, for both builds (the only change is  another (upgraded) kernel version) -> must be kernel related.
I'm using the same file systems , I've been trying ext2,ubifs, even FAT .  

Old-2.6.35.x-Kernel : sync - time very good ( acceptable)
NEW-4.4.x-Kernel   : sync  - time .. BAD .. takes way to long.  !!! 

Any clue what could be the cause , or any hints that could explain ( and solve ?) this ..


Best Regards
Noel

^ permalink raw reply	[flat|nested] 2+ messages in thread

* re: SYNC - takes VERY  long complete  ( system freezes) !
       [not found] ` <D24E577EA1E37B4788A3BFBC418F14A70750D5DD-bUMf7dFpk4BvAH7Z6HRyZmXnswh1EIUO@public.gmane.org>
@ 2017-04-19 11:12   ` Hubert Chaumette
  0 siblings, 0 replies; 2+ messages in thread
From: Hubert Chaumette @ 2017-04-19 11:12 UTC (permalink / raw)
  To: Vellemans Noel,
	kernel-testers-u79uwXL29TY76Z2rM5mHXA@public.gmane.org

> Any clue what could be the cause, or any hints that could explain ( and solve ?) this ..

Assuming the regression is 100% reproducible, your best bet is IMO to do a (git) bisect between the two kernels versions.
Regards,

Hubert Chaumette

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2017-04-19 11:12 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2017-04-18  8:43 SYNC - takes VERY long complete ( system freezes) ! Vellemans, Noel
     [not found] ` <D24E577EA1E37B4788A3BFBC418F14A70750D5DD-bUMf7dFpk4BvAH7Z6HRyZmXnswh1EIUO@public.gmane.org>
2017-04-19 11:12   ` Hubert Chaumette

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox