linux-toradex.git/arch/arm, branch v3.10.5

ARM: 7722/1: zImage: Convert 32bits memory size and address from ATAG to 64bits DTB

2013-08-04T08:50:38+00:00

commit faefd550c45d8d314e8f260f21565320355c947f upstream.

When CONFIG_ARM_APPENDED_DTB is selected, if the bootloader provides
an ATAG_MEM it replaces the memory size and the memory address in the
memory node of the device tree. In the case of a system which can
handle more than 4GB, the memory node cell size is 4: each data
(memory size and memory address) are 64 bits and then use 2 cells.

The current code in atags_to_fdt.c made the assumption of a cell size
of 2 (one cell for the memory size and one cell for the memory
address), this leads to an improper write of the data and ends with a
boot hang.

This patch writes the memory size and the memory address on the memory
node in the device tree depending of the size of the memory node (32
bits or 64 bits).

It has been tested in the 2 cases:
- with a dtb using skeleton.dtsi
- and with a dtb using skeleton64.dtsi

Signed-off-by: Gregory CLEMENT 
Acked-by: Nicolas Pitre 
Signed-off-by: Russell King 
Cc: Willy Tarreau 
Signed-off-by: Greg Kroah-Hartman

ARM: footbridge: fix overlapping PCI mappings

2013-07-28T23:30:17+00:00

commit 6287e7319870ec949fb809e4eb4154c2b05b221f upstream.

Commit 8ef6e6201b26cb9fde79c1baa08145af6aca2815 (ARM: footbridge: use
fixed PCI i/o mapping) broke booting on my netwinder.  Before that,
everything boots fine.  Since then, it crashes on boot.

With earlyprintk, I see it BUG-ing like so:
kernel BUG at lib/ioremap.c:27!
Internal error: Oops - BUG: 0 [#1] ARM
...
[] (ioremap_page_range+0x128/0x154) from [] (dc21285_setup+0xd0/0x114)
[] (dc21285_setup+0xd0/0x114) from [] (pci_common_init+0xa0/0x298)
[] (pci_common_init+0xa0/0x298) from [] (netwinder_pci_init+0xc/0x18)
[] (netwinder_pci_init+0xc/0x18) from [] (do_one_initcall+0xb4/0x180)
...

Russell points out it's because of overlapping PCI mappings that was
added with the aforementioned commit.  Rob thought the code would re-use
the static mapping, but that turns out to not be the case and instead
hits the BUG further down.

After deleting this hunk as suggested by Russel, the system boots up fine
again and all my PCI devices work (IDE, ethernet, the DC21285).

Signed-off-by: Mike Frysinger 
Acked-by: Rob Herring 
Signed-off-by: Olof Johansson 
Signed-off-by: Greg Kroah-Hartman

ARM: S3C24XX: Add missing clkdev entries for s3c2440 UART

2013-07-28T23:30:16+00:00

commit d817468c4b2892b9468e2a0c92116e38a3a61370 upstream.

This patch restores serial port operation which has been broken since
commit 60e93575476f ("serial: samsung: enable clock before clearing
pending interrupts during init")

That commit only uncovered the real issue which was missing clkdev
entries for the "uart" clocks on S3C2440. It went unnoticed so far
because return value of clk API calls were not being checked at all
in the samsung serial port driver.

This patch should be backported to at least 3.10 stable kernel, since
the serial port has not been working on s3c2440 since 3.10-rc5.

Signed-off-by: Sylwester Nawrocki 
Cc: Chander Kashyap 
[on S3C2440 SoC based Mini2440 board]
Tested-by: Sylwester Nawrocki 
Reviewed-by: Tomasz Figa 
Tested-by: Juergen Beisert 
Signed-off-by: Kukjin Kim 
Signed-off-by: Greg Kroah-Hartman

ARM: mm: fix boot on SA1110 Assabet

2013-07-22T01:21:35+00:00

commit 319e0b4f02f73983c03a2ca38595fc6367929edf upstream.

Commit 83db0384 (mm/ARM: use common help functions to free reserved
pages) broke booting on the Assabet by trying to convert a PFN to
a virtual address using the __va() macro.  This macro takes the
physical address, not a PFN.  Fix this.

Signed-off-by: Russell King 
Signed-off-by: Greg Kroah-Hartman

ARM: shmobile: emev2 GIO3 resource fix

2013-07-22T01:21:35+00:00

commit 1eb14ea1e6bcd11d6d0ba937fc39808bb4d3453e upstream.

Fix GIO3 base addresses for EMEV2.

This bug was introduced by 088efd9273b5076a0aead479aa31f1066d182b3e
("mach-shmobile: Emma Mobile EV2 GPIO support V3") which was included in v3.5.

Signed-off-by: Magnus Damm 
Signed-off-by: Simon Horman 
Signed-off-by: Greg Kroah-Hartman

ARM: shmobile: r8a73a4: Fix resources for SCIFB0

2013-07-22T01:21:35+00:00

commit f820b60582f75e73e83b8505d7e48fe59770f558 upstream.

Fix base address and IRQ resources associated with SCIFB0.

This bug was introduced by e481a528901d0cd18b5b5fcbdc55207ea3b6ef68
("ARM: shmobile: r8a73a4 SCIF support V3") which was included in v3.10.

Signed-off-by: Takanari Hayama 
Acked-by: Magnus Damm 
[ horms+renesas@verge.net.au: Add information about commit and version
  this bug was added in ]
Signed-off-by: Simon Horman 
Signed-off-by: Greg Kroah-Hartman

ARM: dts: imx: cpus/cpu nodes dts updates

2013-07-22T01:21:34+00:00

commit 7925e89f54fc49bcd1e73f0a65c4a3eb35b9cfb1 upstream.

This patch updates the in-kernel dts files according to the latest cpus
and cpu bindings updates for ARM.

Signed-off-by: Lorenzo Pieralisi 
Acked-by: Shawn Guo 
Signed-off-by: Greg Kroah-Hartman

ARM: 7778/1: smp_twd: twd_update_frequency need be run on all online CPUs

2013-07-22T01:21:34+00:00

commit cbbe6f82b489e7ceba4ad7c833bd3a76cd0084cb upstream.

When the local timer freq changed, the twd_update_frequency function
should be run all the CPUs include itself, otherwise, the twd freq will
not get updated and the local timer will not run correcttly.

smp_call_function will run functions on all other CPUs, but not include
himself, this is not correct,use on_each_cpu instead to fix this issue.

Acked-by: Linus Walleij 
Cc: Linus Walleij 
Cc: Rob Herring 
Cc: Shawn Guo 
Cc: Arnd Bergmann 
Acked-by: Shawn Guo 
Signed-off-by: Jason Liu 
Signed-off-by: Russell King 
Signed-off-by: Greg Kroah-Hartman

ARM: 7769/1: Cortex-A15: fix erratum 798181 implementation

2013-07-22T01:21:34+00:00

commit 0d0752bca1f9a91fb646647aa4abbb21156f316c upstream.

Looking into the active_asids array is not enough, as we also need
to look into the reserved_asids array (they both represent processes
that are currently running).

Also, not holding the ASID allocator lock is racy, as another CPU
could schedule that process and trigger a rollover, making the erratum
workaround miss an IPI.

Exposing this outside of context.c is a little ugly on the side, so
let's define a new entry point that the erratum workaround can call
to obtain the cpumask.

Acked-by: Will Deacon 
Acked-by: Catalin Marinas 
Signed-off-by: Marc Zyngier 
Signed-off-by: Russell King 
Signed-off-by: Greg Kroah-Hartman

ARM: 7768/1: prevent risks of out-of-bound access in ASID allocator

2013-07-22T01:21:34+00:00

commit b8e4a4740fa2b17c0a447b3ab783b3dc10702e27 upstream.

On a CPU that never ran anything, both the active and reserved ASID
fields are set to zero. In this case the ASID_TO_IDX() macro will
return -1, which is not a very useful value to index a bitmap.

Instead of trying to offset the ASID so that ASID #1 is actually
bit 0 in the asid_map bitmap, just always ignore bit 0 and start
the search from bit 1. This makes the code a bit more readable,
and without risk of OoB access.

Acked-by: Will Deacon 
Acked-by: Catalin Marinas 
Reported-by: Catalin Marinas 
Signed-off-by: Marc Zyngier 
Signed-off-by: Russell King 
Signed-off-by: Greg Kroah-Hartman