Commit Graph

2746 Commits

Author SHA1 Message Date
Paul Mundt
028c5d5d59 Merge branch 'sh/stable-updates' 2010-02-15 14:49:37 +09:00
Paul Mundt
4b505db9c4 sh64: fix tracing of signals.
This follows the parisc change to ensure that tracehook_signal_handler()
is aware of when we are single-stepping in order to ptrace_notify()
appropriately. While this was implemented for 32-bit SH, sh64 neglected
to make use of TIF_SINGLESTEP when it was folded in with the 32-bit code,
resulting in ptrace_notify() never being called.

As sh64 uses all of the other abstractions already, this simply plugs in
the thread flag in the appropriate enable/disable paths and fixes up the
tracehook notification accordingly. With this in place, sh64 is brought
in line with what 32-bit is already doing.

Reported-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-15 14:17:45 +09:00
Paul Mundt
19f6b8b44e sh64: fix up memory offset calculation.
The linker script offsets were broken by the recent 29/32-bit
integration, so this fixes it up for sh64.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-12 15:41:45 +09:00
Paul Mundt
b0f3ae03ac sh: Isolate uncached mapping support.
This splits out the uncached mapping support under its own config option,
presently only used by 29-bit mode and 32-bit + PMB. This will make it
possible to optionally add an uncached mapping on sh64 as well as booting
without an uncached mapping for 32-bit.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-12 15:40:00 +09:00
Paul Mundt
a4dad4c75c sh: update sdk7786 defconfig.
This plugs in USB and PCI and other bits for SDK7786.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-10 16:06:42 +09:00
Paul Mundt
7578a4c625 sh: Fix up multi-resource mapping for SH7786 PCIe.
This reworks some of the SH7786 PCIe initialization code to dynamically
setup and size the various resource windows, as opposed to the original
code that simply wired in a couple of them statically.

At the same time, we tidy up the initialization code a bit, kill off some
read-only register twiddling that was gleaned from the bus analyzer, and
also propagate the physical slot/channel mapping.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-10 16:00:58 +09:00
Magnus Damm
801cd56e3e sh: break out enable/reparent div4 clocks on sh7723
Break out sh7723 div4 clocks for SIU and IRDA as
reparent / enable clocks. Similar to the SIU clock
patch for sh7722 by Guennadi.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-09 18:24:31 +09:00
Magnus Damm
3844eadcfd sh: sh7724/Ecovec24/KFR2R09/MS7724SE SDHI vector merge
Merge the SDHI vectors in the sh7724 INTC table
and update the SDHI platform data for Ecovec24,
KFR2R09 and MS7724SE.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-09 18:24:30 +09:00
Magnus Damm
e3e80046e0 sh: sh7723/AP325 SDHI vector merge
Merge the SDHI vectors in the sh7723 INTC table
and update the SDHI platform data for AP325.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-09 18:24:30 +09:00
Magnus Damm
8d9adabac3 sh: sh7722/Migo-R SDHI vector merge
Merge the SDHI vectors in the sh7722 INTC table
and update the SDHI platform data for Migo-R.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-09 18:24:29 +09:00
Paul Mundt
7561f2dd39 sh: Fix up SH7786 PCI resource definitions.
This adds in some of the missing memory resources for channels 1/2 and
gets the code building again for the recent changes.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-08 16:36:56 +09:00
Paul Mundt
13fd7aeb9a Merge branches 'sh/dwarf-unwinder', 'sh/g3-prep' and 'sh/stable-updates' 2010-02-08 11:48:10 +09:00
Paul Mundt
2e18e04798 Merge branch 'sh/dmaengine'
Conflicts:
	arch/sh/drivers/dma/dma-sh.c
2010-02-08 11:34:03 +09:00
Matt Fleming
858918b77b sh: Optimise FDE/CIE lookup by using red-black trees
Now that the DWARF unwinder is being used to provide perf callstacks
unwinding speed is an issue. It is no longer being used in exceptional
circumstances where we don't care about runtime performance, e.g. when
panicing, so it makes sense improve performance is possible.

With this patch I saw a 42% improvement in unwind time when calling
return_address(1). Greater improvements will be seen as the number of
levels unwound increases as each unwind is now cheaper.

Note that insertion time has doubled but that's just the price we pay
for keeping the trees balanced. However, this is a one-time cost for
kernel boot/module load and so the improvements in lookup time dominate
the extra time we spend keeping the trees balanced.

Signed-off-by: Matt Fleming <matt@console-pimps.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-08 11:29:15 +09:00
Matt Fleming
1af0b2fc67 sh: Remove superfluous setup_frame_reg call
There's no need to setup the frame pointer again in
call_handle_tlbmiss. The frame pointer will already have been setup in
handle_interrupt.

Signed-off-by: Matt Fleming <matt@console-pimps.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-08 10:47:11 +09:00
Matt Fleming
944a343861 sh: Don't continue unwinding across interrupts
Unfortunately, due to poor DWARF info in current toolchains, unwinding
through interrutps cannot be done reliably. The problem is that the
DWARF info for function epilogues is wrong.

Take this standard epilogue sequence,

80003cc4:       e3 6f           mov     r14,r15
80003cc6:       26 4f           lds.l   @r15+,pr
80003cc8:       f6 6e           mov.l   @r15+,r14
						<---- interrupt here
80003cca:       f6 6b           mov.l   @r15+,r11
80003ccc:       f6 6a           mov.l   @r15+,r10
80003cce:       f6 69           mov.l   @r15+,r9
80003cd0:       0b 00           rts

If we take an interrupt at the highlighted point, the DWARF info will
bogusly claim that the return address can be found at some offset from
the frame pointer, even though the frame pointer was just restored. The
worst part is if the unwinder finds a text address at the bogus stack
address - unwinding will continue, for a bit, until it finally comes
across an unexpected address on the stack and blows up.

The only solution is to stop unwinding once we've calculated the
function that was executing when the interrupt occurred. This PC can be
easily calculated from pt_regs->pc.

Signed-off-by: Matt Fleming <matt@console-pimps.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-08 10:47:04 +09:00
Matt Fleming
1dca56f138 sh: Setup frame pointer in handle_exception path
In order to allow the DWARF unwinder to unwind through exceptions we
need to setup the frame pointer register (r14).

Signed-off-by: Matt Fleming <matt@console-pimps.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-08 10:46:53 +09:00
Matt Fleming
142698282c sh: Correct the offset of the return address in ret_from_exception
The address that ret_from_exception and ret_from_irq will return to is
found in the stack slot for SPC, not PR. This error was causing the
DWARF unwinder to pick up the wrong return address on the stack and then
unwind using the unwind tables for the wrong function.

While I'm here I might as well add CFI annotations for the other
registers since they could be useful when unwinding.

Signed-off-by: Matt Fleming <matt@console-pimps.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-08 10:46:46 +09:00
Guennadi Liakhovetski
cfefe99795 sh: implement DMA_SLAVE capability in SH dmaengine driver
Tested to work with a SIU ASoC driver on sh7722 (migor).

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-08 09:40:26 +09:00
Guennadi Liakhovetski
623b4ac4bf sh: fix Transfer Size calculation in both DMA drivers
Both the original arch/sh/drivers/dma/dma-sh.c and the new SH dmaengine drivers
do not take into account bits 3:2 of the Transfer Size field in the CHCR
register, besides, bit-field defines set bit 2, but the mask only passes bits
1:0 through. TS_16BLK and TS_32BLK macros are bogus too. This patch fixes all
these issues for sh7722 and sh7724, other CPUs stay unchanged and might need to
be fixed too.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-08 09:40:24 +09:00
Paul Mundt
b06ede84dd sh: Fix up hp6xx build.
With the sparseirq conversion there was a stray irq_desc reference left
over, this tidies it up and brings the demuxer in line with what the
solution engine boards are doing.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-05 17:45:14 +09:00
Paul Mundt
3b0be1a4f2 sh: Fix an off-by-1 in SH7780 PCIC memory resource mapping.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-05 16:11:25 +09:00
Paul Mundt
9ad62ec4f7 sh: Fix up early PCI PERR/SERR IRQ handling.
This adds support for handling early PERR/SERR triggering in between
controller registration and the initial bus scan. Buggy cards end up
asserting these as soon as the M66EN scan is undertaken, resulting in
an early crash.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-03 16:46:20 +09:00
Paul Mundt
37feecb0ae sh: Disable generic IRQ probing.
IRQ autoprobing hasn't actually worked for us at all since very early in
2.6, but no one seems to have noticed given that none of the drivers
that use it see much testing.

yenta_socket is the odd one out, and that depends on PCI IRQs which are
fixed on all SH platforms anyways. Consequently, turning off autoprobing
fixes up crashes triggered by yenta_socket and at least gets it working
again on r7785rp.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-02 19:17:42 +09:00
Paul Mundt
4b842c8571 sh: enable sparseirq for highlander and r2d.
highlander and r2d are the only remaining ones that were blocking
sparseirq being turned on by default, but it turns out that they already
work fine with it by virtue of register_intc_controller(). As such, we
can kill off the dependencies and turn it on by default.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-02 19:15:17 +09:00
Paul Mundt
deb9b22b89 sh: mach-dreamcast: Convert to sparseirq.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-02 18:01:55 +09:00
Paul Mundt
6eb6f98396 sh: hd6446x: Convert to sparseirq.
Follows the se7724 change and converts the hd64461 IRQ handling to
sparseirq.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-02 17:48:17 +09:00
Paul Mundt
c7a967273a sh: mach-se: Convert SH7724 solution engine FPGA to sparseirq.
This uses the new create_irq_nr() to build up the FPGA's desired virtual
IRQ mapping and permits us to finally flip on sparseirq for this board.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-02 17:38:04 +09:00
Guennadi Liakhovetski
b4f74767a0 sh: add high impedance mode management for SIUA pins on sh7722
This improves power management for the SIUA controller on sh7722. Similar
patches might be desired for other SIU-enabled SH platforms.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-02 13:02:30 +09:00
Paul Mundt
9d3f1881ab Merge branch 'sh/stable-updates' 2010-02-02 11:33:45 +09:00
Marek Skuczynski
bc10e875d4 sh: Fix access to released memory in clk_debugfs_register_one()
Signed-off-by: Marek Skuczynski <mareksk7@gmail.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-02 11:32:23 +09:00
Marek Skuczynski
00b3e0a2e0 sh: Fix access to released memory in dwarf_unwinder_cleanup()
Signed-off-by: Marek Skuczynski <mareksk7@gmail.com>
Acked-by: Matt Fleming <matt@console-pimps.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-02 11:32:22 +09:00
Paul Mundt
b6c58b1d98 sh: Improved multi-resource handling for SH7780 PCI.
The SH7780 PCI controller supports 3 different ranges of PCI memory in
addition to its PCI I/O window. In the case of 29-bit mode, only 2 memory
windows are supported, while in 32-bit mode all 3 are visible. This
attempts to make the resource handling completely dynamic and to permit
platforms to map in as many apertures as they can handle.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-01 20:01:50 +09:00
Paul Mundt
ef407beefb sh: Hook up ERR/PERR/SERR detection for SH7780 PCI host controllers.
These were never handled before, so implement some common infrastructure
to support them, then make use of that in the SH7780-specific code. In
practice there is little here that can not be generalized for SH4 parts,
which will be an incremental change as the 7780/7751 code is gradually
unified.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-01 16:39:46 +09:00
Paul Mundt
bcf39352eb sh: Handle PCI controller resource conflicts.
register_pci_controller() can fail, but presently is a void function.
Change this over to an int so that we can bail early before continuing on
with post-registration initialization (such as throwing the controller in
to 66MHz mode in the case of the SH7780 host controller).

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-01 13:11:25 +09:00
Paul Mundt
85b59f5bb2 sh: Enable PCI66 support for SH7780 host controller.
This adds some helper glue for scanning the bus and determining if all
of the devices are 66MHz capable or not before flipping on 66MHz mode.
This isn't quite to spec, but it's fairly consistent with what other
embedded controllers end up having to do.

Scanning code cribbed from the MIPS txx9 PCI code.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-01 13:01:42 +09:00
Paul Mundt
aee4467b5c sh: Fix up large system memory handling for SH7780 PCI.
For systems that have more than 512MB we need to set up an additional
mapping, this fixes up the rounding to the next power of two and splits
out the mapping accordingly between the two local bus mapping windows.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-02-01 11:33:22 +09:00
Paul Mundt
31a090ae06 sh: Fix up asm/hw_breakpoint.h header check.
Presently headers_check complains about linux/kdebug.h being unexported,
so just bump the __KERNEL__ ifdef up, as per the x86 change.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-30 11:29:56 +09:00
Paul Mundt
97b19778ee sh: mach-se: Fix up irq_desc reference.
The irq_desc needs to be accessed with irq_to_desc(), this fixes up a
build error with irq_desc being undefined.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-30 11:04:38 +09:00
Paul Mundt
396c56a9c6 sh: Kill off broken type 1 PCI config access checks.
The host controllers only support type 1, so there's not much else to
test for. Some of the older controllers also supported type 2 accesses,
but we've never supported those, and likely never will. Beyond that, the
P1SEG test is meaningless for 32-bit mode, so rather than refactoring it,
just kill the type 1 test off completely.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-30 01:41:21 +09:00
Linus Torvalds
221af7f87b Split 'flush_old_exec' into two functions
'flush_old_exec()' is the point of no return when doing an execve(), and
it is pretty badly misnamed.  It doesn't just flush the old executable
environment, it also starts up the new one.

Which is very inconvenient for things like setting up the new
personality, because we want the new personality to affect the starting
of the new environment, but at the same time we do _not_ want the new
personality to take effect if flushing the old one fails.

As a result, the x86-64 '32-bit' personality is actually done using this
insane "I'm going to change the ABI, but I haven't done it yet" bit
(TIF_ABI_PENDING), with SET_PERSONALITY() not actually setting the
personality, but just the "pending" bit, so that "flush_thread()" can do
the actual personality magic.

This patch in no way changes any of that insanity, but it does split the
'flush_old_exec()' function up into a preparatory part that can fail
(still called flush_old_exec()), and a new part that will actually set
up the new exec environment (setup_new_exec()).  All callers are changed
to trivially comply with the new world order.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-01-29 08:22:01 -08:00
Paul Mundt
320e68da59 sh: support PCI domains.
Newer SH parts are now commonly shipping with multiple controllers, so
we wire up PCI domain support to deal with them. Shamelessly cloned from
the MIPS implementation.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-29 22:38:13 +09:00
Paul Mundt
ac8ab54a8e sh: Bail out early on PCI resource conflicts.
Presently we just call in to request_resource() for the ioport and iomem
resources without checking for errors. This has already hidden a couple
of bugs, so add some error handling in for good measure.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-29 22:22:27 +09:00
Paul Mundt
a45635dfb0 sh: Reworked SH7780 PCI initialization.
This consolidates the PCI initialization code for all of the pci-sh7780
users, and sets up the memory window dynamically as opposed to using
hardcoded window positions.

A number of bugs were fixed at the same time, including the PIO handling
and master abort timeout settings being incorrect.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-29 22:19:04 +09:00
Paul Mundt
9762528f37 sh: Kill off deprecated fixed PCI memory window accessors.
This kills off the deprected fixed memory range accessors for
the cases of non-translatable ioremapping.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-29 16:14:29 +09:00
Paul Mundt
abec86a802 sh: Kill off superfluous sh_bios references from compressed loader.
The VBR value needs to be initialized in order for any of the bios
trapping to work, and as we don't do that until slightly later in the
boot process, this has no real chance of working.

The only reason for using this in the past was due to the fact that early
serial initialization wasn't possible, but now that early sh-sci is
supported by everyone, there's no longer any reason to support the
sh_bios hacks in the first place. Given that, we just kill off the
references completely.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-28 18:26:37 +09:00
Paul Mundt
d627a2ebd1 sh: Fix up the ioremap_fixed() build for nommu.
arch/sh/kernel/setup.c:455: error: implicit declaration of function 'ioremap_fixed_init'

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-28 18:17:29 +09:00
Paul Mundt
849593591c sh: Fix up read-only variable assignment in pcibios_align_resource().
arch/sh/drivers/pci/pci.c:167: error: assignment of read-only location '*res'

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-28 18:15:05 +09:00
Paul Mundt
621266bdf4 sh: Quiet noisy coherent DMA unmapping on R2D.
IRQs are re-enabled at a later stage when doing the unmapping on R2D via
the sm501 USB coherent DMA, resulting in the irqs_disabled() check
producing considerable noise for this configuration. Just kill off the
check, which was blindly copied from x86 anyways.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-28 15:55:37 +09:00
Paul Mundt
f655f5e956 sh: mach-titan: Kill off unused PIO port mangling.
Nothing is using this, kill it off. Fixing up access sizes can be done
with trapped I/O for anyone wanting to make use of this for devices that
need it, everything else is already pure MMIO.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2010-01-28 15:08:21 +09:00