On Wed, Jun 21, 2017 at 10:39:31AM +0200, Linus Walleij wrote: > On Thu, Jun 8, 2017 at 4:13 PM, Lorenzo Pieralisi > <lorenzo.pieralisi@xxxxxxx> wrote: > > I pulled this into the Gemini tree: > > [1] git://git.kernel.org/pub/scm/linux/kernel/git/lpieralisi/linux.git pci/pci-fixup-irqs-removal-v2 > Thank you Linus for helping me out. > And then this happens: > > OF: PCI: host bridge /soc/pci@50000000 ranges: > OF: PCI: IO 0x50000000..0x500fffff -> 0x00000000 > OF: PCI: MEM 0x58000000..0x5fffffff -> 0x58000000 > Unable to handle kernel NULL pointer dereference at virtual address 00000070 > pgd = c0004000 > [00000070] *pgd=00000000 > Internal error: Oops: 17 [#1] PREEMPT ARM > CPU: 0 PID: 254 Comm: kworker/0:1 Not tainted 4.12.0-rc3+ #888 > Hardware name: Gemini (Device Tree) > Workqueue: events deferred_probe_work_func > task: c38c3a40 task.stack: c390a000 > PC is at faraday_pci_probe+0x274/0x65c > LR is at __irq_put_desc_unlock+0x18/0x70 > pc : [<c0232648>] lr : [<c004be48>] psr: 80000013 I think I see what it is. faraday_pci_irq_handler() relies on struct faraday_pci->bus to access config space and with this series it can't be set till pci_scan_root_bus_bridge() returns. If that's the case we will have to update the config accessors in faraday_pci_irq_handler() to raw ones (ie ones that do not depend on the struct pci_bus), code is already there, patch coming, sorry about this but it is extremely complicated for me to untangle these drivers inner workings. Thanks, Lorenzo > sp : c390bdf0 ip : c3424274 fp : c3ac2ee0 > r10: c3f77718 r9 : c3ad49b0 r8 : c3907800 > r7 : 00000004 r6 : c390be18 r5 : c3907810 r4 : c3ad4810 > r3 : 00000000 r2 : 00000000 r1 : 00000001 r0 : c3ac2de0 > Flags: Nzcv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none > Control: 0000397f Table: 03a9c000 DAC: 00000053 > Process kworker/0:1 (pid: 254, stack limit = 0xc390a190) > Stack: (0xc390bdf0 to 0xc390c000) > bde0: c0534a60 c3ad49b0 00000003 c024cb68 > be00: 00000001 20000013 00000000 50000000 00000007 000020dd c3ae2680 c3ae2740 > be20: c07c861f c0134328 00000000 c3ad5fa0 c3910a50 c3910a50 00000001 00000000 > be40: 00000000 c01344ec c3910a50 c3910a50 00000001 00000000 c3910a50 c39d76e0 > be60: c3ad5fa0 c05f6f58 c3910a50 c3907810 c3907810 c07cba88 fffffdfb 00000003 > be80: 00000000 c07c5040 c386e420 c02bbbe0 c3907810 c081c6c4 00000000 c07cba88 > bea0: 00000003 c02ba9dc 00000000 c390bee0 c02bab30 00000001 c07bf660 00000000 > bec0: c07c5040 c02b93dc c382829c c39886f4 c3907810 c3907844 c3907810 c02ba674 > bee0: c3907810 00000001 c390bf0c c3907810 c07d3a80 c3907810 c07d38e4 c02b961c > bf00: c3907810 c07d38d0 c07d38d0 c02ba210 c07d38ec c386e420 00000000 c3f6b100 > bf20: c07bf660 c002dac0 00000008 c07bf660 c07bf660 c386e438 c390a000 c07bf674 > bf40: 00000008 c07bf660 c07c5040 c002e044 c390a000 00000001 00000000 00000000 > bf60: c3901518 c3901500 ffffe000 c38e2700 00000000 c3901518 c386e420 c002ddb4 > bf80: c3839ee4 c00335ec c390a000 c38e2700 c003350c 00000000 00000000 00000000 > bfa0: 00000000 00000000 00000000 c000a270 00000000 00000000 00000000 00000000 > bfc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 > bfe0: 00000000 00000000 00000000 00000000 00000013 00000000 ffffffff ffffffff > [<c0232648>] (faraday_pci_probe) from [<c02bbbe0>] > (platform_drv_probe+0x50/0xb4) > [<c02bbbe0>] (platform_drv_probe) from [<c02ba9dc>] > (driver_probe_device+0x248/0x2dc) > [<c02ba9dc>] (driver_probe_device) from [<c02b93dc>] > (bus_for_each_drv+0x68/0x98) > [<c02b93dc>] (bus_for_each_drv) from [<c02ba674>] (__device_attach+0xa8/0x110) > [<c02ba674>] (__device_attach) from [<c02b961c>] (bus_probe_device+0x88/0x90) > [<c02b961c>] (bus_probe_device) from [<c02ba210>] > (deferred_probe_work_func+0x54/0x80) > [<c02ba210>] (deferred_probe_work_func) from [<c002dac0>] > (process_one_work+0x13c/0x430) > [<c002dac0>] (process_one_work) from [<c002e044>] (worker_thread+0x290/0x5f4) > [<c002e044>] (worker_thread) from [<c00335ec>] (kthread+0xe0/0x120) > [<c00335ec>] (kthread) from [<c000a270>] (ret_from_fork+0x14/0x24) > Code: e59401b0 e3700a01 8a00001d e59421ac (e5d23070) > ---[ end trace 1e313b1ea11e101c ]--- > > I'm trying to figure out what is causing it. > > Yours, > Linus Walleij
