From mboxrd@z Thu Jan 1 00:00:00 1970 From: Marek Vasut Date: Fri, 27 Jul 2012 03:11:04 +0200 Subject: [U-Boot] [U-Boot-DM] List of offending drivers In-Reply-To: References: <201207270118.19524.marex@denx.de> Message-ID: <201207270311.04978.marex@denx.de> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Dear Graeme Russ, > Hi Marek, > > On Fri, Jul 27, 2012 at 9:18 AM, Marek Vasut wrote: > > Hello, > > > > Check the following list, it's the list of drivers scattered (misplaced) > > across the tree. The list is not complete and might be inaccurate. But > > it should give a good impression of what I'm going to break soon: > > > > arch/x86/cpu/sc520/sc520_pci.c > > board/eNET/eNET_pci.c > > Hmm, I'm wondering what where the line between 'driver' and 'arch/board > specific driver glue' is? That's why I said the list isn't exactly precise. > How was this list generated? By hard manual labor (=slavework). > You seem to have missed sc520_ssi.c and sc520_timer.c I wonder if we should move the timer drivers ... maybe to drivers/timer/ ? > board/eNET/eNET_pci.c only contains: > > pci_enet_fixup_irq() - Board specific configuration of PCI interrupt lines. > This is a platform function which is specified when the board initialises > the PCI driver. Note that this function calls pci_sc520_set_irq() which is > located on arch/x86/cpu/sc520/sc520_pci.c (see below) Ok, so this one should be left out, I didn't properly examine them all. Just wanted to share the list, the examination will follow this weekend, when I start moving them. > pci_init_board() - One line wrapper for pci_sc520_init() which should get > dropped once the driver model and init sequence (if that gets looked at > again) refactoring > > pci_set_regions() - Configure the board-specific PCI memory and I/O regions > > arch/x86/cpu/sc520/sc520_pci.c contains: > pci_sc520_set_irq() - A support function for board-specific PCI interrupt > line configuration > pci_sc520_init() - SC520 specific PCI driver initialisation > > So neither of these files are 'drivers' per-se. They are really just > initialisation and platform specific support functions. How do these fit > into the new driver model? You can supply a pointer to that function to some "pci" driver I guess ... Pavel? > Regards, > > Graeme Best regards, Marek Vasut