Messages in this thread | | | Date | Mon, 2 Dec 2013 13:44:51 -0800 | From | Andrew Morton <> | Subject | Re: [PATCH] parport: parport_pc: remove double PCI ID for NetMos |
| |
On Thu, 28 Nov 2013 11:35:11 +0100 Sebastian Andrzej Siewior <bigeasy@linutronix.de> wrote:
> On 11/27/2013 10:53 PM, Andrew Morton wrote: > >> While here I also drop the entry for titan_1284p2 which is the same as > >> netmos_9815. > > > > Does this change have any runtime-visible effects? > > Titan p2 used to specify class + subclass which is now replaced with > PCI_ID_ANY, I doubt this makes any difference at all. > > The only runtime visible change I notice is with KERN_DEBUG being > visible. The card details are the same, it always prints > > |parport1: PC-style at 0xc120 (0xc128), irq 11, using FIFO > [PCSPP,TRISTATE,COMPAT,ECP] > > However the ID that was used to detect the card is different: > > | PCI parallel port detected: 1415:c118, I/O at 0xc120(0xc128), IRQ 11 > > vs > > | PCI parallel port detected: 9710:9805, I/O at 0xc120(0xc128), IRQ 11 > > The wrong id print is printed for all devices which come after > PCI_DEVICE_ID_OXSEMI_PCIe840_G in the id table because the code that > prints the ID table assumes one id per entry. > > So we could > - drop printk > - use id->vendor + id->device instead > - do the last item + remove parport_pc_pci_cards and pass the data > struct as data. > - nothing > > any preferences? >
Well if the kernel is now printing the incorrect ID, we should fix that. The message should match the discovered hardware, and it should match whatever `lspci' says is there.
| |