linux-toradex.git/arch/powerpc, branch v4.9.121

powerpc/embedded6xx/hlwd-pic: Prevent interrupts from being handled by Starlet

2018-08-03T05:55:21+00:00

[ Upstream commit 9dcb3df4281876731e4e8bff7940514d72375154 ]

The interrupt controller inside the Wii's Hollywood chip is connected to
two masters, the "Broadway" PowerPC and the "Starlet" ARM926, each with
their own interrupt status and mask registers.

When booting the Wii with mini[1], interrupts from the SD card
controller (IRQ 7) are handled by the ARM, because mini provides SD
access over IPC. Linux however can't currently use or disable this IPC
service, so both sides try to handle IRQ 7 without coordination.

Let's instead make sure that all interrupts that are unmasked on the PPC
side are masked on the ARM side; this will also make sure that Linux can
properly talk to the SD card controller (and potentially other devices).

If access to a device through IPC is desired in the future, interrupts
from that device should not be handled by Linux directly.

[1]: https://github.com/lewurm/mini

Signed-off-by: Jonathan Neuschäfer 
Signed-off-by: Michael Ellerman 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

bpf: powerpc64: pad function address loads with NOPs

2018-08-03T05:55:19+00:00

[ Upstream commit 4ea69b2fd623dee2bbc77d3b6b7d8c0924e2026a ]

For multi-function programs, loading the address of a callee
function to a register requires emitting instructions whose
count varies from one to five depending on the nature of the
address.

Since we come to know of the callee's address only before the
extra pass, the number of instructions required to load this
address may vary from what was previously generated. This can
make the JITed image grow or shrink.

To avoid this, we should generate a constant five-instruction
when loading function addresses by padding the optimized load
sequence with NOPs.

Signed-off-by: Sandipan Das 
Signed-off-by: Daniel Borkmann 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

powerpc/8xx: fix invalid register expression in head_8xx.S

2018-08-03T05:55:19+00:00

[ Upstream commit e4ccb1dae6bdef228d729c076c38161ef6e7ca34 ]

New binutils generate the following warning

  AS      arch/powerpc/kernel/head_8xx.o
arch/powerpc/kernel/head_8xx.S: Assembler messages:
arch/powerpc/kernel/head_8xx.S:916: Warning: invalid register expression

This patch fixes it.

Signed-off-by: Christophe Leroy 
Signed-off-by: Michael Ellerman 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

powerpc/powermac: Mark variable x as unused

2018-08-03T05:55:19+00:00

[ Upstream commit 5a4b475cf8511da721f20ba432c244061db7139f ]

Since the value of x is never intended to be read, declare it with gcc
attribute as unused. Fix warning treated as error with W=1:

  arch/powerpc/platforms/powermac/bootx_init.c:471:21: error: variable ‘x’ set but not used [-Werror=unused-but-set-variable]

Suggested-by: Christophe Leroy 
Signed-off-by: Mathieu Malaterre 
Signed-off-by: Michael Ellerman 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

powerpc/powermac: Add missing prototype for note_bootable_part()

2018-08-03T05:55:18+00:00

[ Upstream commit f72cf3f1d49f2c35d6cb682af2e8c93550f264e4 ]

Add a missing prototype for function `note_bootable_part` to silence a
warning treated as error with W=1:

  arch/powerpc/platforms/powermac/setup.c:361:12: error: no previous prototype for ‘note_bootable_part’ [-Werror=missing-prototypes]

Suggested-by: Christophe Leroy 
Signed-off-by: Mathieu Malaterre 
Signed-off-by: Michael Ellerman 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

powerpc/chrp/time: Make some functions static, add missing header include

2018-08-03T05:55:18+00:00

[ Upstream commit b87a358b4a1421abd544c0b554b1b7159b2b36c0 ]

Add a missing include .

These functions can all be static, make it so. Fix warnings treated as
errors with W=1:

  arch/powerpc/platforms/chrp/time.c:41:13: error: no previous prototype for ‘chrp_time_init’ [-Werror=missing-prototypes]
  arch/powerpc/platforms/chrp/time.c:66:5: error: no previous prototype for ‘chrp_cmos_clock_read’ [-Werror=missing-prototypes]
  arch/powerpc/platforms/chrp/time.c:74:6: error: no previous prototype for ‘chrp_cmos_clock_write’ [-Werror=missing-prototypes]
  arch/powerpc/platforms/chrp/time.c:86:5: error: no previous prototype for ‘chrp_set_rtc_time’ [-Werror=missing-prototypes]
  arch/powerpc/platforms/chrp/time.c:130:6: error: no previous prototype for ‘chrp_get_rtc_time’ [-Werror=missing-prototypes]

Signed-off-by: Mathieu Malaterre 
Signed-off-by: Michael Ellerman 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

powerpc/32: Add a missing include header

2018-08-03T05:55:18+00:00

[ Upstream commit c89ca593220931c150cffda24b4d4ccf82f13fc8 ]

The header file  was missing from the includes. Fix the
following warning, treated as error with W=1:

  arch/powerpc/kernel/pci_32.c:286:6: error: no previous prototype for ‘sys_pciconfig_iobase’ [-Werror=missing-prototypes]

Signed-off-by: Mathieu Malaterre 
Signed-off-by: Michael Ellerman 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

powerpc/64s: Fix compiler store ordering to SLB shadow area

2018-08-03T05:55:14+00:00

[ Upstream commit 926bc2f100c24d4842b3064b5af44ae964c1d81c ]

The stores to update the SLB shadow area must be made as they appear
in the C code, so that the hypervisor does not see an entry with
mismatched vsid and esid. Use WRITE_ONCE for this.

GCC has been observed to elide the first store to esid in the update,
which means that if the hypervisor interrupts the guest after storing
to vsid, it could see an entry with old esid and new vsid, which may
possibly result in memory corruption.

Signed-off-by: Nicholas Piggin 
Signed-off-by: Michael Ellerman 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

powerpc/eeh: Fix use-after-release of EEH driver

2018-08-03T05:55:14+00:00

[ Upstream commit 46d4be41b987a6b2d25a2ebdd94cafb44e21d6c5 ]

Correct two cases where eeh_pcid_get() is used to reference the driver's
module but the reference is dropped before the driver pointer is used.

In eeh_rmv_device() also refactor a little so that only two calls to
eeh_pcid_put() are needed, rather than three and the reference isn't
taken at all if it wasn't needed.

Signed-off-by: Sam Bobroff 
Signed-off-by: Michael Ellerman 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

powerpc/fadump: Unregister fadump on kexec down path.

2018-07-03T09:23:08+00:00

commit 722cde76d68e8cc4f3de42e71c82fd40dea4f7b9 upstream.

Unregister fadump on kexec down path otherwise the fadump registration
in new kexec-ed kernel complains that fadump is already registered.
This makes new kernel to continue using fadump registered by previous
kernel which may lead to invalid vmcore generation. Hence this patch
fixes this issue by un-registering fadump in fadump_cleanup() which is
called during kexec path so that new kernel can register fadump with
new valid values.

Fixes: b500afff11f6 ("fadump: Invalidate registration and release reserved memory for general use.")
Cc: stable@vger.kernel.org # v3.4+
Signed-off-by: Mahesh Salgaonkar 
Signed-off-by: Michael Ellerman 
Signed-off-by: Greg Kroah-Hartman