Re: [microblaze-uclinux] Power-fail handling

["Jim Law" <jlaw@xxxxxxxxxxxxx>]

Hi again John,

Thanks for the place to look.  I'm in the 2.4 kernel, and it looks like 
emergency_sync() doesn't exist in linux/fs.h there.  But I followed your 
clue to look in drivers/char/sysrq.c and I see something that looks like a 
earlier attempt at a "do_emergency_sync()" in the driver itself.  I'll study 
this for a while to see if I can figure out how this function gets 
registered to be called back by the bdflush kernel thread.  Is this likely 
to be interrupt-safe in your opinion?

Thanks,

Jim


----- Original Message ----- 
From: "John Williams" <jwilliams@xxxxxxxxxxxxxx>
To: <microblaze-uclinux@xxxxxxxxxxxxxx>
Sent: Sunday, October 19, 2008 10:09 PM
Subject: Re: [microblaze-uclinux] Power-fail handling


> Hi Jim,
>
> Jim Law wrote:
>
> > We have a SystemAce CardFlash with a single FAT partition.  Our custom 
> > board gives an interrupt when the AC power begins to fail, allowing us at 
> > least 20ms of time before on-board power begins to go away.  I'm trying 
> > to find a reliable way to flush anything that might be in uClinux's 
> > buffers to the CardFlash when we get that interrupt. I believe sync() 
> > does what we want, but I'm having issues getting there fast enough.  We 
> > catch the power fail interrupt just fine.  Then we throw a signal that 
> > gets caught by a handler in our application.  Our application priority 
> > has been set so we catch the signal reliably quickly in the handler, and 
> > so far we're running less than 1ms into our available time. But it seems 
> > like the call to sync() does not return reliably in the time available, 
> > even if we know there is nothing pending to write to the CardFlash.
> >  I don't know if I've chosen a silly way of doing this, or if there is 
> > some other side-effect or function of sync() that might be taking the 
> > time.
>
> You might try calling emergency_sync() directly from your interrupt 
> handler in the driver, that's how they do it in e.g. drivers/char/sysrq.c
>
> Your driver will have to #include <linux/fs.h> to get the function 
> prototype.
>
> Let us know if this helps, it's a useful capability.
>
> Cheers,
>
> John
>
> ___________________________
> microblaze-uclinux mailing list
> microblaze-uclinux@xxxxxxxxxxxxxx
> Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
> Mailing List Archive : 
> http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/

___________________________
microblaze-uclinux mailing list
microblaze-uclinux@xxxxxxxxxxxxxx
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/