[PATCH 0/19]: RPAPHP pci hotplug cleanup patchbomb

Kristen Carlson Accardi kristen.c.accardi at intel.com
Fri Apr 6 01:33:48 EST 2007


On Tue, 3 Apr 2007 11:34:14 -0500
linas at austin.ibm.com (Linas Vepstas) wrote:

> 
> Hi Kristen,
> 
> Please queue these cleanup patches for 2.6.22.

Ok - will do.

> 
> This is a collection of very small, mostly trite, patches that clean up 
> various bits and pieces of the RPAPHP hotplug code. They eliminate
> almost 10% of the code, while making almost no funcional change. 
> There are a few bugfixes to various error paths, and one memleak fix. 
> Some documentation is added.  The result is, I beleive, slightly more 
> readable, easier to understand code. In particular, the enable/disable 
> add/remove code paths are now more obviously symmetrical in thier function.
> 
> --linas
> 
> p.s. some more simplifcation is possible: one could probably merge
> __enable_slot() and rpaphp_enable_slot() with a bit of elbow grease,
> and the asymmetric pairing of rpaphp_deregister_slot()  with
> rpaphp_add_slot() as "opposites" of each other still bugs me.
> 
> I'm also irked that dlpar_pci_add_bus() is quite similar to 
> pcibios_add_pci_devices() which is quite similar to init_phb_dynamic()
> and think that these should be refactored so that they are more clearly
> orthogonal to one another. Just right now, I'm not planning on doing 
> anything about this, at least, not without prodding.
> 
> --linas 
> 

Thanks for the cleanups - most of the drivers in the hotplug tree badly
need cleaning up, but... 

Kristen



More information about the Linuxppc-dev mailing list