> AFAICT, the situation with drivers/ide is that it happens to work (by
> luck) well enough for most people that it's been ignored until now. And
> when it doesn't work, the result is a hard, difficult-to-debug hang on
> resume that is rarely diagnosed because the standard
> try-it-with-all-modules-unloaded trick doesn't cover permanent
> non-unloadable ide drivers. Or because the ide layer eventually resets
> the bus without hanging machine and the user just thinks the resume is
> "slow".
> 
> I gather that there are two important phases to ide suspend: the ide
> buses (meaning the drives), and the ide controller (usually on the pci
> bus) that gives access to the ide bus.  My first patch was NACKed by
> Alan because I was doing IDE drive resume at the pci ide controller
> level.

The IDE code has a suspend implementation that works fine for cases
where

 - No host protected area is used. That is the disk doesn't need to be
reconfigured in any way but the usual speed settings
 - The controller driver has code to bring itself back to some sane
timings (PIO0 is fine) on resume
before the generic ide code starts the resume process.

Which means it works fine for PowerMac and a bunch of embedded stuffs,
but tends to have issues with x86 peecees...

As Alan pointed out at KS, the proper fix would be to use the ACPI
provided bits on these though while I wrote the current IDE
suspend/resume implementation, I will not do that as I have 0 knowledge
of ACPI and no x86 hardware to test.
 
Cheers,
Ben.



-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Suspend-devel mailing list
Suspend-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/suspend-devel

Reply via email to