linux-toradex.git/arch/powerpc/sysdev/mpic_u3msi.c, branch v3.10.94

powerpc/MSI: Fix race condition in tearing down MSI interrupts

2015-10-22T21:37:52+00:00

commit e297c939b745e420ef0b9dc989cb87bda617b399 upstream.

This fixes a race which can result in the same virtual IRQ number
being assigned to two different MSI interrupts.  The most visible
consequence of that is usually a warning and stack trace from the
sysfs code about an attempt to create a duplicate entry in sysfs.

The race happens when one CPU (say CPU 0) is disposing of an MSI
while another CPU (say CPU 1) is setting up an MSI.  CPU 0 calls
(for example) pnv_teardown_msi_irqs(), which calls
msi_bitmap_free_hwirqs() to indicate that the MSI (i.e. its
hardware IRQ number) is no longer in use.  Then, before CPU 0 gets
to calling irq_dispose_mapping() to free up the virtal IRQ number,
CPU 1 comes in and calls msi_bitmap_alloc_hwirqs() to allocate an
MSI, and gets the same hardware IRQ number that CPU 0 just freed.
CPU 1 then calls irq_create_mapping() to get a virtual IRQ number,
which sees that there is currently a mapping for that hardware IRQ
number and returns the corresponding virtual IRQ number (which is
the same virtual IRQ number that CPU 0 was using).  CPU 0 then
calls irq_dispose_mapping() and frees that virtual IRQ number.
Now, if another CPU comes along and calls irq_create_mapping(), it
is likely to get the virtual IRQ number that was just freed,
resulting in the same virtual IRQ number apparently being used for
two different hardware interrupts.

To fix this race, we just move the call to msi_bitmap_free_hwirqs()
to after the call to irq_dispose_mapping().  Since virq_to_hw()
doesn't work for the virtual IRQ number after irq_dispose_mapping()
has been called, we need to call it before irq_dispose_mapping() and
remember the result for the msi_bitmap_free_hwirqs() call.

The pattern of calling msi_bitmap_free_hwirqs() before
irq_dispose_mapping() appears in 5 places under arch/powerpc, and
appears to have originated in commit 05af7bd2d75e ("[POWERPC] MPIC
U3/U4 MSI backend") from 2007.

Fixes: 05af7bd2d75e ("[POWERPC] MPIC U3/U4 MSI backend")
Reported-by: Alexey Kardashevskiy 
Signed-off-by: Paul Mackerras 
Signed-off-by: Michael Ellerman 
Signed-off-by: Greg Kroah-Hartman

powerpc: Convert to new irq_* function names

2011-03-29T12:48:12+00:00

Scripted with coccinelle.

Signed-off-by: Thomas Gleixner

powerpc: mpic irq_data conversion.

2011-03-10T00:03:56+00:00

Signed-off-by: Lennert Buytenhek 
Signed-off-by: Benjamin Herrenschmidt

pci: Convert msi to new irq_chip functions

2010-10-12T14:53:34+00:00

Signed-off-by: Thomas Gleixner 
Reviewed-by: Ingo Molnar 
Acked-by: Jesse Barnes 
Cc: Benjamin Herrenschmidt 
Cc: "David S. Miller" 
Cc: Tony Luck 
Cc: Russell King

powerpc: Fix MSI support on U4 bridge PCIe slot

2009-12-18T03:55:43+00:00

On machines using the Apple U4 bridge (AKA IBM CPC945) PCIe interface such
as the latest generation G5 machines x16 slot or the x16 slot of the
PowerStation, MSIs are currently broken (and will oops when enabling).

This fixes the oops and implements proper support for those. Instead of
using the PCIe <-> HT bridge conversion, on such slots we need to use
a bunch of magic registers in the bridge as the MSI target, encoding
the interrupt number in the low bits of the address itself

Signed-off-by: Benjamin Herrenschmidt

powerpc: Fixup last users of irq_chip->typename

2009-11-24T03:32:45+00:00

The typename member of struct irq_chip was kept for migration purposes
and is obsolete since more than 2 years. Fix up the leftovers.

Signed-off-by: Thomas Gleixner 
Cc: Benjamin Herrenschmidt 
Cc: linuxppc-dev@ozlabs.org
Acked-by: Geoff Levand 
Signed-off-by: Benjamin Herrenschmidt

powerpc: Convert the MPIC MSI code to use msi_bitmap

2008-08-20T06:34:58+00:00

This affects the U3 MSI code as well as the PASEMI MSI code.  We keep
some of the MPIC routines as helpers, and also the U3 best-guess
reservation logic.  The rest is replaced by the generic code.

And a few printk format changes due to hwirq type change.

Signed-off-by: Michael Ellerman 
Signed-off-by: Paul Mackerras

[POWERPC] mpic_u3msi: Failed allocation unnoticed

2008-05-14T12:31:25+00:00

bitmap_find_free_region(), called by mpic_msi_alloc_hwirqs() may return
signed, but hwirq is unsigned.  A failed allocation remains unnoticed.

Signed-off-by: Roel Kluin <12o3l@tiscali.nl>
Signed-off-by: Paul Mackerras

[POWERPC] Inline u3msi_compose_msi_msg()

2007-10-02T23:11:42+00:00

In the MPIC U3 MSI code, we call u3msi_compose_msi_msg() once for each MSI.
This is overkill, as the address is per pci device, not per MSI. So setup
the address once, and just set the data per MSI.

Signed-off-by: Michael Ellerman 
Acked-by: Benjamin Herrenschmidt 
Signed-off-by: Paul Mackerras

[POWERPC] Simplify error logic in u3msi_setup_msi_irqs()

2007-10-02T23:11:32+00:00

u3msi_setup_msi_irqs() doesn't need to call teardown() itself,
the generic code will do this for us as long as we return a non
zero value.

Signed-off-by: Michael Ellerman 
Acked-by: Benjamin Herrenschmidt 
Signed-off-by: Paul Mackerras