Commit graph

377,253 commits

Author SHA1 Message Date
Jesse Larrew
176bbf1424 powerpc/pseries: Update NUMA VDSO information when updating CPU maps
The following patch adds vdso_getcpu_init(), which stores the NUMA node for
a cpu in SPRG3:

Commit 18ad51dd34 ("powerpc: Add VDSO version of getcpu") adds
vdso_getcpu_init(), which stores the NUMA node for a cpu in SPRG3.

This patch ensures that this information is also updated when the NUMA
affinity of a cpu changes.

Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:24 +10:00
Nathan Fontenot
30c05350c3 powerpc/pseries: Use stop machine to update cpu maps
The new PRRN firmware feature allows CPU and memory resources to be
transparently reassigned across NUMA boundaries. When this happens, the
kernel must update the node maps to reflect the new affinity information.

Although the NUMA maps can be protected by locking primitives during the
update itself, this is insufficient to prevent concurrent accesses to these
structures. Since cpumask_of_node() hands out a pointer to these
structures, they can still be modified outside of the lock. Furthermore,
tracking down each usage of these pointers and adding locks would be quite
invasive and difficult to maintain.

The approach used is to make a list of affected cpus and call stop_machine
to have the update routine run on each of the affected cpus allowing them
to update themselves. Each cpu finds itself in the list of cpus and makes
the appropriate updates. We need to have each cpu do this for themselves to
handle calls to vdso_getcpu_init() added in a subsequent patch.

Situations like these are best handled using stop_machine(). Since the NUMA
affinity updates are exceptionally rare events, this approach has the
benefit of not adding any overhead while accessing the NUMA maps during
normal operation.

Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:24 +10:00
Jesse Larrew
5d88aa85c0 powerpc/pseries: Update CPU maps when device tree is updated
Platform events such as partition migration or the new PRRN firmware
feature can cause the NUMA characteristics of a CPU to change, and these
changes will be reflected in the device tree nodes for the affected
CPUs.

This patch registers a handler for Open Firmware device tree updates
and reconfigures the CPU and node maps whenever the associativity
changes. Currently, this is accomplished by marking the affected CPUs in
the cpu_associativity_changes_mask and allowing
arch_update_cpu_topology() to retrieve the new associativity information
using hcall_vphn().

Protecting the NUMA cpu maps from concurrent access during an update
operation will be addressed in a subsequent patch in this series.

Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:23 +10:00
Nathan Fontenot
8002b0c54b powerpc/pseries: Update numa.c to use updated firmware_has_feature()
Update the numa code to use the updated firmware_has_feature() when checking
for type 1 affinity.

Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:22 +10:00
Nathan Fontenot
f0ff7eb483 powerpc/pseries: Update firmware_has_feature() to check architecture vector 5 bits
The firmware_has_feature() function makes it easy to check for supported
features of the hypervisor. This patch extends the capability of
firmware_has_feature() to include checking for specified bits
in vector 5 of the architecture vector as reported in the device tree.

As part of this the #defines used for the architecture vector are re-defined
such that each option has the index into vector 5 and the feature bit encoded
into it. This makes checking for architecture bits when initiating data
for firmware_has_feature much easier.

Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:22 +10:00
Nathan Fontenot
43c0ea6053 powerpc/pseries: Use ARRAY_SIZE to iterate over firmware_features_table array
When iterating over the entries in firmware_features_table we only need
to go over the actual number of entries in the array instead of declaring
it to be bigger and checking to make sure there is a valid entry in every
slot.

This patch removes the FIRMWARE_MAX_FEATURES #define and replaces the
array looping with the use of ARRAY_SIZE().

Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:21 +10:00
Nathan Fontenot
530b5e1475 powerpc/pseries: Move architecture vector definitions to prom.h
As part of handling of PRRN events we need to check vector 5 of the
architecture vector bits reported in the device tree to ensure PRRN event
handling is enabled. To do this firmware_has_feature() is updated (in a
subsequent patch) to make this check vector 5 bits. To avoid having to
re-define bits in the architecture vector the bit definitions are moved
to prom.h.

Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:21 +10:00
Jesse Larrew
49c68a8518 powerpc/pseries: Add PRRN RTAS event handler
A PRRN event is signaled via the RTAS event-scan mechanism, which
returns a Hot Plug Event message "fixed part" indicating "Platform
Resource Reassignment". In response to the Hot Plug Event message,
we must call ibm,update-nodes to determine which resources were
reassigned and then ibm,update-properties to obtain the new affinity
information about those resources.

The PRRN event-scan RTAS message contains only the "fixed part" with
the "Type" field set to the value 160 and no Extended Event Log. The
four-byte Extended Event Log Length field is re-purposed (since no
Extended Event Log message is included) to pass the "scope" parameter
that causes the ibm,update-nodes to return the nodes affected by the
specific resource reassignment.

This patch adds a handler for RTAS events. The function
pseries_devicetree_update() (from mobility.c) is used to make the
ibm,update-nodes/ibm,update-properties RTAS calls. Updating the NUMA maps
(handled by a subsequent patch) will require significant processing,
so pseries_devicetree_update() is called from an asynchronous workqueue
to allow event processing to continue.

PRRN RTAS events on pseries systems are rare events that have to be
initiated from the HMC console for the system by an IBM tech. This allows
us to assume that these events are widely spaced. Additionally, all work
on the queue is flushed before handling any new work to ensure we only have
one event in flight being handled at a time.

Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:20 +10:00
Nathan Fontenot
2e9b7b02a3 powerpc/pseries: Correct buffer parsing in update_dt_node()
Correct parsing of the buffer returned from ibm,update-properties. The first
element is a length and the path to the property which is slightly different
from the list of properties in the buffer so we need to specifically
handle this.

Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:19 +10:00
Nathan Fontenot
762ec15707 powerpc/pseries: Expose pseries devicetree_update()
Newer firmware on Power systems can transparently reassign platform resources
(CPU and Memory) in use. For instance, if a processor or memory unit is
predicted to fail, the platform may transparently move the processing to an
equivalent unused processor or the memory state to an equivalent unused
memory unit. However, reassigning resources across NUMA boundaries may alter
the performance of the partition. When such reassignment is necessary, the
Platform Resource Reassignment Notification (PRRN) option provides a
mechanism to inform the Linux kernel of changes to the NUMA affinity of
its platform resources.

When rtasd receives a PRRN event, it needs to make a series of RTAS
calls (ibm,update-nodes and ibm,update-properties) to retrieve the
updated device tree information. These calls are already handled in the
pseries_devicetree_update() routine used in partition migration.

This patch exposes pseries_devicetree_update() to make it accessible
to other pseries routines, this patch also updates pseries_devicetree_update()
to take a 32-bit scope parameter. The scope value, which was previously hard
coded to 1 for partition migration, is used for the RTAS calls
ibm,update-nodes/properties to update the device tree.

Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:19 +10:00
Michael Neuling
3e96ca7f00 powerpc: Fix hardware IRQs with MMU on exceptions when HV=0
POWER8 allows us to take interrupts with the MMU on.  This gives us a
second set of vectors offset at 0x4000.

Unfortunately when coping these vectors we missed checking for MSR HV
for hardware interrupts (0x500).  This results in us trying to use
HSRR0/1 when HV=0, rather than SRR0/1 on HW IRQs

The below fixes this to check CPU_FTR_HVMODE when patching the code at
0x4500.

Also we remove the check for CPU_FTR_ARCH_206 since relocation on IRQs
are only available in arch 2.07 and beyond.

Thanks to benh for helping find this.

Signed-off-by: Michael Neuling <mikey@neuling.org>
CC: <stable@vger.kernel.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:18 +10:00
Michael Neuling
8c2a381734 powerpc/power8: Fix secondary CPUs hanging on boot for HV=0
In __restore_cpu_power8 we determine if we are HV and if not, we return
before setting HV only resources.

Unfortunately we forgot to restore the link register from r11 before
returning.

This will happen on boot and with secondary CPUs not coming online.

This adds the missing link register restore.

Signed-off-by: Michael Neuling <mikey@neuling.org>
CC: <stable@vger.kernel.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:17 +10:00
Michael Neuling
29ce3c5073 powerpc: Add isync to copy_and_flush
In __after_prom_start we copy the kernel down to zero in two calls to
copy_and_flush.  After the first call (copy from 0 to copy_to_here:)
we jump to the newly copied code soon after.

Unfortunately there's no isync between the copy of this code and the
jump to it.  Hence it's possible that stale instructions could still be
in the icache or pipeline before we branch to it.

We've seen this on real machines and it's results in no console output
after:
  calling quiesce...
  returning from prom_init

The below adds an isync to ensure that the copy and flushing has
completed before any branching to the new instructions occurs.

Signed-off-by: Michael Neuling <mikey@neuling.org>
CC: <stable@vger.kernel.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:17 +10:00
Michael Neuling
2171364d1a powerpc: Add HWCAP2 aux entry
We are currently out of free bits in AT_HWCAP. With POWER8, we have
several hardware features that we need to advertise.

Tested on POWER and x86.

Signed-off-by: Michael Neuling <michael@neuling.org>
Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-04-26 16:08:16 +10:00
Dave Airlie
36d9b1541c Merge branch 'drm-nouveau-next' of git://anongit.freedesktop.org/git/nouveau/linux-2.6 into drm-next
"Nothing overly exciting here aside from calim's fermi/kepler vram
compression patches.  The rest is misc fixes I gathered from the list.

Most of the stuff from me is fixing issues that have come up from the
work on kepler PM, as well as a commit moving all the old-school
modesetting out of the way (no code changes here).  There's other
patches to go on top of that, but, it'll have to wait until I can rip
out the old PM code, it's a bit tangled."

* 'drm-nouveau-next' of git://anongit.freedesktop.org/git/nouveau/linux-2.6: (38 commits)
  drm/nouveau/fifo: implement channel creation event generation
  drm/nouveau/core: allow non-maskable events
  drm/nouveau/timer: allow alarms to be cancelled
  drm/nouveau/device: tweak the device/subdev relationship a little
  drm/nouveau/device: enable proper constructor/destructor
  drm/nouveau/device: have engine object initialised before creation
  drm/nouveau/device: convert to engine, rather than subdev
  drm/nv50-/disp: use self as parent for subobjects
  drm/nv50-/fifo: use parent as self for subobjects
  drm/nv20-nv30/gr: use parent as self for subobjects
  drm/nvc0-/gr: use self as parent for subobjects
  drm/nv04-nv40/instmem: use self as parent for subobjects
  drm/nv04-nv40/vm: use self as parent for subobjects
  drm/nv50-/bar: use self as parent for subobjects
  drm/nv04-nv40/instmem: remove parent deref hack
  drm/nouveau/i2c: remove parent deref hack
  drm/nouveau/core: rebase object ref/use counts after ctor/init/fini events
  drm/nv50/disp: inform core when we're not creating a new context
  drm/nouveau/therm: send some messages to debug level
  drm/nve0/gr: add handling for a bunch of PGRAPH traps
  ...
2013-04-26 15:42:02 +10:00
Ben Skeggs
893e90c554 drm/nouveau/fifo: implement channel creation event generation
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:17 +10:00
Ben Skeggs
10eeaf123d drm/nouveau/core: allow non-maskable events
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:16 +10:00
Ben Skeggs
6d1d1cc97b drm/nouveau/timer: allow alarms to be cancelled
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:15 +10:00
Ben Skeggs
10caad339c drm/nouveau/device: tweak the device/subdev relationship a little
Fixes not-in-use engines not having their reset() method called on
resume.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:14 +10:00
Ben Skeggs
066a5d0938 drm/nouveau/device: enable proper constructor/destructor
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:12 +10:00
Ben Skeggs
9aecbada75 drm/nouveau/device: have engine object initialised before creation
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:11 +10:00
Ben Skeggs
dded35dee3 drm/nouveau/device: convert to engine, rather than subdev
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:10 +10:00
Ben Skeggs
2ecda48b36 drm/nv50-/disp: use self as parent for subobjects
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:08 +10:00
Ben Skeggs
f50c805488 drm/nv50-/fifo: use parent as self for subobjects
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:07 +10:00
Ben Skeggs
a3e6789a54 drm/nv20-nv30/gr: use parent as self for subobjects
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:06 +10:00
Ben Skeggs
617a6cbd7c drm/nvc0-/gr: use self as parent for subobjects
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:05 +10:00
Ben Skeggs
1409d90f24 drm/nv04-nv40/instmem: use self as parent for subobjects
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:03 +10:00
Ben Skeggs
be1e8e16ec drm/nv04-nv40/vm: use self as parent for subobjects
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:02 +10:00
Ben Skeggs
f83145ecd7 drm/nv50-/bar: use self as parent for subobjects
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:01 +10:00
Ben Skeggs
b5795c77e5 drm/nv04-nv40/instmem: remove parent deref hack
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:38:00 +10:00
Ben Skeggs
d395f1e4c5 drm/nouveau/i2c: remove parent deref hack
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:58 +10:00
Ben Skeggs
db91d68c9b drm/nouveau/core: rebase object ref/use counts after ctor/init/fini events
This is intended to support named (with a handle, etc) objects having
children that don't have an outside reference.

This will replace the various hacks around the place where subdev
objects have children, and have to manually drop the self-refs so
that they can be destroyed etc when all the outside refs have gone.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:57 +10:00
Ben Skeggs
43e6e51c03 drm/nv50/disp: inform core when we're not creating a new context
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:56 +10:00
Ben Skeggs
bdd4e843fa drm/nouveau/therm: send some messages to debug level
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:55 +10:00
Christoph Bumiller
bf3d8165e4 drm/nve0/gr: add handling for a bunch of PGRAPH traps
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:53 +10:00
Ben Skeggs
0fa9061ae8 drm/nouveau/mc: handle irq-related setup ourselves
We need to be able to process interrupts before the DRM code is able to
actually enable them, set it up ourselves.  Also, it's less convoluted
to *not* use the DRM wrappers it appears...

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:52 +10:00
Ben Skeggs
1a64634255 drm/nv04/disp: hide all the cruft away in its own little hole
It'd be pretty awesome if someone would care enough to port this all
properly to a class interface, perhaps submitting a command stream to
the core via a sw object on PFIFO (emulating how EVO works basically,
and also what nvidia have done forever..)..

But, this seems unlikely given how old this hardware is now, so, lets
just hide it away.

There's a heap of other bits and pieces laying around that are still
tangled.  I'll (re)move them in pieces.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:51 +10:00
Ben Skeggs
b9a3140ce8 drm/nouveau/bios: add missing newline on IO*_OR opcode debugging
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:50 +10:00
Ben Skeggs
28ec70f7fb drm/nouveau/bios: suppress some parser errors when dry-running scripts
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:48 +10:00
Ben Skeggs
3f196a045e drm/nve0: magic up some support for GF117
Seen in the wild, don't have the hardware but this hacks things up to
treat it the same as GF119 for now.

Should be relatively safe, I'd be very surprised if anything major
changed outside of PGRAPH.  PGRAPH (3D etc) is disabled by default
however until it's confirmed working.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:47 +10:00
Roy Spliet
8cb303a85b drm/nvc0-/ltcg: Fix build on 32-bit platforms (v2)
v2: read, don't assume.. *puts on brown paper bag*

Signed-off-by: Roy Spliet <r.spliet@student.tudelft.nl>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:46 +10:00
Florian Scholz
cade2413e9 drm/nouveau/drm: adding support for backlight control of GT525M (NVC0)
This patch adds support for the backlight control of the NVIDIA GT
525M, which identifies itself as a member of the NVC0 family.

v2. Extended to handle Kepler too

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:45 +10:00
Christoph Bumiller
142c21b8d4 drm/nouveau/drm: bump the driver version to 1.1.1 to report new features
Signed-off-by: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:43 +10:00
Christoph Bumiller
9685482385 drm/nvc0/gr: add software methods to control some MP regs
Signed-off-by: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:42 +10:00
Christoph Bumiller
7e22e71e80 drm/nvc0-: support NOUVEAU_GETPARAM_GRAPH_UNITS
Signed-off-by: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:41 +10:00
Christoph Bumiller
e30441adb9 drm/nvc0-/ltcg: implement VRAM compression
Signed-off-by: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:40 +10:00
Ben Skeggs
de7b7d59d5 drm/nouveau/ttm: allow tiled memtype on system memory buffer objects
Compression not supported, and will be silently dropped.  Original G80
can't handle this either and requires LINEAR memtype, though it's still
possible to correctly texture and m2mf to/from these objects anyway.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:38 +10:00
Ben Skeggs
7092a8dd5c drm/nve0/grctx: initial attempt at unhardcoding yet more magic
Not sure about the (gpc_nr == 1) condition, it's probably wrong but for
all the examples I've seen so far it matches what NVIDIA end up poking.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:37 +10:00
Martin Peres
ba366c25bc drm/nv50/therm: implement temperature reading
Signed-off-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:36 +10:00
Martin Peres
2f4573679a drm/nouveau/therm: split the nv50 and nv84 code
This is needed because temperature management on nv50 can be enabled and it
looks about the same as nv40.

Signed-off-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-04-26 15:37:35 +10:00