On 03/02/2012 01:29 PM, David Daney wrote: > On 03/02/2012 11:07 AM, Grant Likely wrote: >> On Fri, 02 Mar 2012 10:03:58 -0800, David >> Daney<david.daney@xxxxxxxxxx> wrote: >>> On 03/02/2012 06:22 AM, Rob Herring wrote: >>> [...] >>>>> diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig >>>>> index ce30e2f..01344ae 100644 >>>>> --- a/arch/mips/Kconfig >>>>> +++ b/arch/mips/Kconfig >>>>> @@ -1432,6 +1432,7 @@ config CPU_CAVIUM_OCTEON >>>>> select WEAK_ORDERING >>>>> select CPU_SUPPORTS_HIGHMEM >>>>> select CPU_SUPPORTS_HUGEPAGES >>>>> + select IRQ_DOMAIN >>>> >>>> IIRC, Grant has a patch cued up that enables IRQ_DOMAIN for all of >>>> MIPS. >>>> >>> >>> Indeed, I now see it in linux-next. I will remove this one. >>> >>>>> help >>>>> The Cavium Octeon processor is a highly integrated chip >>>>> containing >>>>> many ethernet hardware widgets for networking tasks. The >>>>> processor >>>>> diff --git a/arch/mips/cavium-octeon/octeon-irq.c >>>>> b/arch/mips/cavium-octeon/octeon-irq.c >>>>> index bdcedd3..e9f2f6c 100644 >>>>> --- a/arch/mips/cavium-octeon/octeon-irq.c >>>>> +++ b/arch/mips/cavium-octeon/octeon-irq.c >>> [...] >>>>> +static void __init octeon_irq_set_ciu_mapping(unsigned int irq, >>>>> + unsigned int line, >>>>> + unsigned int bit, >>>>> + struct irq_domain *domain, >>>>> struct irq_chip *chip, >>>>> irq_flow_handler_t handler) >>>>> { >>>>> + struct irq_data *irqd; >>>>> union octeon_ciu_chip_data cd; >>>>> >>>>> irq_set_chip_and_handler(irq, chip, handler); >>>>> - >>>>> cd.l = 0; >>>>> cd.s.line = line; >>>>> cd.s.bit = bit; >>>>> >>>>> irq_set_chip_data(irq, cd.p); >>>>> octeon_irq_ciu_to_irq[line][bit] = irq; >>>>> + >>>>> + irqd = irq_get_irq_data(irq); >>>>> + irqd->hwirq = line<< 6 | bit; >>>>> + irqd->domain = domain; >>>> >>>> I think the domain code will set these. >>> >>> It is my understanding that the domain code only does this for: >>> >>> o irq_domain_add_legacy() >>> >>> o irq_create_direct_mapping() >>> >>> o irq_create_mapping() >>> >>> We use none of those. So I do it here. >>> >>> If there is a better way, I am open to suggestions. >> >> irq_create_mapping is called by irq_create_of_mapping() which is >> in turn called by irq_of_parse_and-map(). irq_domain always >> manages the hwirq and domain values. Driver code cannot manipulate >> them manually. >> > > I really must be missing something. > > Given: > > 1) I must have a mapping between hwirq and irq that I control so that > non-OF code using the OCTEON_IRQ_* constants continues to work. Those defines are what you need to work to get rid of. > 2) irq_create_mapping() will allocate a random irq value if none is > already assigned to the hwirq. > > Therefore: To avoid having random irq values assigned, I must manually > assign them. > So you should be using legacy domain if you need to maintain fixed hwirq to linux irq numbers. "linear" is a bit confusing as it doesn't mean linear 1:1 irq number assignment, but linear search. Ultimately, for DT boot you should use of_irq_init to scan the dts, and then create a linear domain for each interrupt controller node. You may need to decide on linear vs. legacy at runtime based on having a DT node pointer or not. Rob