RE: BUG/PATCH? mmcblk devices don't suspend properly.

2012-04-05 Thread Hebbar, Gururaja
I contacted the original author of the patch (http://comments.gmane.org/gmane.linux.kernel.mmc/9168) And he has posted the 2nd version of the commit. [PATCH V2]mmc: remove MMC bus legacy suspend/resume method MMC bus is using legacy suspend/resume method, which is not compatible if runtime pm c

RE: BUG/PATCH? mmcblk devices don't suspend properly.

2012-04-04 Thread Hebbar, Gururaja
On Tue, Apr 03, 2012 at 02:13:54, Per Forlin wrote: > On Fri, Mar 30, 2012 at 7:25 AM, Sujit Reddy Thumma > wrote: > > On 3/30/2012 8:17 AM, NeilBrown wrote: > >> > >> ...snip... ...snip... > >> > >> The net effect is that - as I observed - mmc_bus_suspend is never called. > >> > >> I added line

Re: BUG/PATCH? mmcblk devices don't suspend properly.

2012-04-02 Thread Per Forlin
On Fri, Mar 30, 2012 at 7:25 AM, Sujit Reddy Thumma wrote: > On 3/30/2012 8:17 AM, NeilBrown wrote: >> >> >> I've been experimenting with removing the call to sys_sync() in >> enter_state().  For me (with verbose debugging and syslog running) >> this causes a noticeable delay when entering suspend

Re: BUG/PATCH? mmcblk devices don't suspend properly.

2012-03-29 Thread Sujit Reddy Thumma
On 3/30/2012 8:17 AM, NeilBrown wrote: I've been experimenting with removing the call to sys_sync() in enter_state(). For me (with verbose debugging and syslog running) this causes a noticeable delay when entering suspend. Removing it should not affect correctness as there is no locking the en

BUG/PATCH? mmcblk devices don't suspend properly.

2012-03-29 Thread NeilBrown
I've been experimenting with removing the call to sys_sync() in enter_state(). For me (with verbose debugging and syslog running) this causes a noticeable delay when entering suspend. Removing it should not affect correctness as there is no locking the ensure that no other process writes out dat