Re: Kexec, DMA, and SMP

[Corey Minyard <minyard@acm.org>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Eric W. Biederman wrote:

|Corey Minyard <cminyard@mvista.com> writes:
|
|>|
|>|(So adding a special mode to the power management code may
|>|be too much overhead. Besides, sometimes, you can just pull
|>|a reset line, and don't have to do anything even remotely
|>|related to power management.)
|>
|>True, I didn't mean the high-level power management code directly.  
But the
|>PCI API defines a suspend operation that could take a special mode for 
this.
|
|
|The generic device api has a shutdown method for this.  And in the non 
panic
|case we use it.  Not a lot of devices have it implemented but it exists.
|
|And except that it doesn't have a restriction that it can't block is pretty
|much what you want.

That's a pretty big restriction.  Plus, you can't claim spinlocks.

The panic shutdown is different from an orderly shutdown.  What the 
current shutdown does is probably not what you want.

|
|>Or maybe a new field in the PCI structure (and equivalent for other 
things, if
|>there are any).  But the suspend and resume operations should at least 
give
|>a good idea where its needed and how to use it.
|
|
|The API is already done...

The API is not done for panics.  There's no call that has the proper 
semantics.

|
|
|We just don't trust the dying kernel enough to use it during a panic.

I don't understand this.  If you can't trust a dying kernel to properly 
shut down devices, how can you trust it to boot a new kernel?  And (much 
worse) if you don't shut down the devices, how can you trust the new 
kernel to execute properly?  I know there are levels of trust here, but 
I'd much rather have the kernel lockup during the reboot than have a 
chance of a new kernel booting that could behave incorrectly.  In 
general, the chance of behaving incorrectly is MUCH worse than a sure 
lockup, especially in systems that must be reliable.

- -Corey
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQE+T7rOIXnXXONXERcRAksfAJ9kVRD2S9OK5siBqAPMkbfi2iS2fgCeM3hw
Fjp2LXiNEURU+HNrByOGVBQ=
=5sxh
-----END PGP SIGNATURE-----