Hi Bin. On 19 January 2016 at 02:15, Bin Meng <bmeng...@gmail.com> wrote: > Hi Simon, > > On Tue, Jan 19, 2016 at 11:39 AM, Simon Glass <s...@chromium.org> wrote: >> This series adds an interrupt driver for x86. Since different platforms >> can implement this in their own way, we no-longer need the platform-specific >> weak function. We can also dispense with the arch_misc_init() call in some >> cases. >> >> Changes in v2: >> - Rebase on top of updated SPI flash series >> >> Simon Glass (6): >> dm: x86: Create a driver for x86 interrupts >> dm: x86: Set up interrupt routing from interrupt_init() >> dm: x86: Add a common PIRQ init function >> dm: x86: quark: Add an interrupt driver >> dm: x86: queensbay: Add an interrupt driver >> dm: x86: Drop the weak cpu_irq_init() function >> > > I've tested this series, but oops, it broke queensbay and quark. Sorry > I did not catch this during previous review :( > > The issue is with: > > Both queensbay and quark irq driver's probe will call > create_pirq_routing_table() in the end, where it calls: > > node = fdtdec_next_compatible(blob, 0, COMPAT_INTEL_IRQ_ROUTER); > > but COMPAT_INTEL_IRQ_ROUTER is not the string for queensbay and quark. > Maybe we need pass the compatible string as the parameter to > create_pirq_routing_table()?
I think we can drop this with a few more patches. I sent a v3 for you to try/review. Regards, Simon _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot