x86: Don't include linux/irq.h from asm/hardirq.h
authorNicolai Stange <nstange@suse.de>
Sun, 29 Jul 2018 10:15:33 +0000 (12:15 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 15 Aug 2018 16:14:52 +0000 (18:14 +0200)
commit8574df1a8741f6cce1f2fbdd921b07adeec8d932
treec1364da6136da6bfe06b46568d048aa883bac1ed
parente371c92e168df9c0713bd4085fbb8501d88b297a
x86: Don't include linux/irq.h from asm/hardirq.h

commit 447ae316670230d7d29430e2cbf1f5db4f49d14c upstream

The next patch in this series will have to make the definition of
irq_cpustat_t available to entering_irq().

Inclusion of asm/hardirq.h into asm/apic.h would cause circular header
dependencies like

  asm/smp.h
    asm/apic.h
      asm/hardirq.h
        linux/irq.h
          linux/topology.h
            linux/smp.h
              asm/smp.h

or

  linux/gfp.h
    linux/mmzone.h
      asm/mmzone.h
        asm/mmzone_64.h
          asm/smp.h
            asm/apic.h
              asm/hardirq.h
                linux/irq.h
                  linux/irqdesc.h
                    linux/kobject.h
                      linux/sysfs.h
                        linux/kernfs.h
                          linux/idr.h
                            linux/gfp.h

and others.

This causes compilation errors because of the header guards becoming
effective in the second inclusion: symbols/macros that had been defined
before wouldn't be available to intermediate headers in the #include chain
anymore.

A possible workaround would be to move the definition of irq_cpustat_t
into its own header and include that from both, asm/hardirq.h and
asm/apic.h.

However, this wouldn't solve the real problem, namely asm/harirq.h
unnecessarily pulling in all the linux/irq.h cruft: nothing in
asm/hardirq.h itself requires it. Also, note that there are some other
archs, like e.g. arm64, which don't have that #include in their
asm/hardirq.h.

Remove the linux/irq.h #include from x86' asm/hardirq.h.

Fix resulting compilation errors by adding appropriate #includes to *.c
files as needed.

Note that some of these *.c files could be cleaned up a bit wrt. to their
set of #includes, but that should better be done from separate patches, if
at all.

Signed-off-by: Nicolai Stange <nstange@suse.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
[dwmw2: More fixes for EFI and Xen in 4.9]
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
28 files changed:
arch/x86/include/asm/dmi.h
arch/x86/include/asm/hardirq.h
arch/x86/include/asm/kvm_host.h
arch/x86/kernel/apic/apic.c
arch/x86/kernel/apic/htirq.c
arch/x86/kernel/apic/io_apic.c
arch/x86/kernel/apic/msi.c
arch/x86/kernel/apic/vector.c
arch/x86/kernel/fpu/core.c
arch/x86/kernel/ftrace.c
arch/x86/kernel/hpet.c
arch/x86/kernel/i8259.c
arch/x86/kernel/irq.c
arch/x86/kernel/irq_32.c
arch/x86/kernel/irq_64.c
arch/x86/kernel/irqinit.c
arch/x86/kernel/kprobes/core.c
arch/x86/kernel/kprobes/opt.c
arch/x86/kernel/smpboot.c
arch/x86/kernel/time.c
arch/x86/mm/fault.c
arch/x86/mm/kaiser.c
arch/x86/platform/efi/efi_64.c
arch/x86/platform/efi/quirks.c
arch/x86/platform/intel-mid/device_libs/platform_mrfld_wdt.c
arch/x86/xen/enlighten.c
arch/x86/xen/setup.c
drivers/pci/host/pci-hyperv.c