On Thu, 2006-02-23 at 17:50 +0800, erich wrote: > Dear Arjan van de Ven, > > The following contex is coming from comment of Christoph Hellwig. > > - msi should be a module options if at all, but defintitly not > a config options > > #ifdef CONFIG_SCSI_ARCMSR_MSI > if (!pci_enable_msi(pci_device)) > pACB->acb_flags |= ACB_F_HAVE_MSI; > #endif > > I make an option config for prevent some mainboards hang up if arcmsr enable > msi function. > Areca RAID controller is bridged hardware. > There were a lots of mainboards had wrong IRQ routing table issue with it. > If somebody meet this issue and people can enable msi function to fix its > hardware bug. > But unfortunately I found some mainboards will hang up if I always enable > this function in my lab. > To avoid this issue, I do an option for this case. yes the reason for making this optional is clear, and Christoph also understands that. However the idea that Christoph is proposing is to not make it a compile time option, but a runtime option. Compile-time is not very flexible, especially not for linux distributions. Making it a module option means it becomes runtime behavior, and the user can load the module like modprobe aerca msi=0 and msi gets turned off. No need to recompile anything! That has many advantages over a more inflexible (from the user view) compiletime-only option. - : send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
