Whilst writing testcases to exercise the VMA API, some oddities came to
light, such as i915_gem_obj_lookup_or_create(). Joonas suggested
i915_vma_instance() as a neat replacement, so rename them, move them to
i915_vma.c and add some kerneldoc as a sugary bonus.
s/i915_gem_obj_to_vma/i915_vma_lookup/
s/i915_gem_obj_lookup_or_create_vma/i915_vma_instance/
Suggested-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170116152131.18089-2-chris@chris-wilson.co.uk
With atomic plane states we are able to track an allocation right from
preparation, during use and through to the final free after being
swapped out for a new plane. We can couple the VMA we pin for the
framebuffer (and its rotation) to this lifetime and avoid all the clumsy
lookups in between.
v2: Remove residual vma on plane cleanup (Chris)
v3: Add a description for the vma destruction in
intel_plane_destroy_state (Maarten)
References: https://bugs.freedesktop.org/show_bug.cgi?id=98829
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170116152131.18089-1-chris@chris-wilson.co.uk
Acked-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
The constants MAX_COMPLETIONS and MSG_QUEUE_SIZE control
the number of messages that can be outstanding to each client
before the system as a whole stalls. If the numbers are too
small then unnecessary thread switching will occur while
waiting for a (potentially low priority) client thread to
consume some data; badly written clients can even lead to
deadlock.
For services that carry many short messages, 16 messages can
represent a very small amount of data. Since the resources
are small - 16 bytes for a completion, 4 bytes for a message
pointer - increase the limits so they are unlikely to be hit
except in exceptional circumstances.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
An issue was observed when flushing openmax components
which generate a large number of messages returning
buffers to host.
We occasionally found a duplicate message from 16
messages prior, resulting in a buffer returned twice.
So fix the issue by adding more memory barriers.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Service callbacks are not allowed to return an error. The internal
callback that delivers events and messages to user tasks does not
enqueue them if the service is closing, but this is not an error
and should not be reported as such.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The dequeue_pending flag wasn't protected by a spinlock in the
service_callback. So fix this to make it safe.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This reduces the memory dump size to a sufficient value of 16 bytes.
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Proper byte order was completely disregarded for multi byte data shared
between AP and module (and APB1). Fix this.
Signed-off-by: Vaibhav Agarwal <vaibhav.agarwal@linaro.org>
Signed-off-by: Vaibhav Agarwal <vaibhav.sr@gmail.com>
Acked-by: Mark Greer <mgreer@animalcreek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dummy codec register were initially added while populating dummy codec
mixer controls until module topology parser was available. Now, these
dummy registers are nowhere used and thus can be safely removed.
Since ASoC framework requires a valid callback for both read & write
register APIS, currently empty placeholders are kept to avoid panic.
Later, register mapping logic can be defined:
1. Assuming fixed number of maximum modules connected and register bits
corresponds to basic info of each module OR
2. With a logic to dynamically grow register_cache_size based on codec
modules added/removed.
Signed-off-by: Vaibhav Agarwal <vaibhav.agarwal@linaro.org>
Signed-off-by: Vaibhav Agarwal <vaibhav.sr@gmail.com>
Acked-by: Mark Greer <mgreer@animalcreek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Uninitialized variable sig_bits was used while configuring stream params
for codec module. These params are used to configure PCM settings for
APBridgeA.
Usually, this is dependent on codec capability and thus populated via
codec dai_driver definition. In our case, it is fixed to 16 based on the
data format, container supported.
Signed-off-by: Vaibhav Agarwal <vaibhav.agarwal@linaro.org>
Signed-off-by: Vaibhav Agarwal <vaibhav.sr@gmail.com>
Acked-by: Mark Greer <mgreer@animalcreek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mixer control->info call back function checks for -ve values to rebase
min and max values. However, le32 variable is used to fetch values from
GB module FW. Thus negative value checking is not required. Fix this!!
Signed-off-by: Vaibhav Agarwal <vaibhav.sr@gmail.com>
Acked-by: Mark Greer <mgreer@animalcreek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
checkpatch.pl warning:
Symbolic permissions are not preferred. Consider using octal permissions.
Signed-off-by: Igor Pylypiv <igor.pylypiv@gmail.com>
Reviewed-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
len was declared unsigned int where we use an int
Fix sparse (-Wtypesign) issues:
drivers/staging/rtl8188eu/core/rtw_ieee80211.c:1081:88: warning: incorrect type in argument 3 (different signedness)
drivers/staging/rtl8188eu/core/rtw_ieee80211.c:1081:88: expected int *len
drivers/staging/rtl8188eu/core/rtw_ieee80211.c:1081:88: got unsigned int *<noident>
drivers/staging/rtl8188eu/core/rtw_ieee80211.c:1091:86: warning: incorrect type in argument 3 (different signedness)
drivers/staging/rtl8188eu/core/rtw_ieee80211.c:1091:86: expected int *len
drivers/staging/rtl8188eu/core/rtw_ieee80211.c:1091:86: got unsigned int *<noident>
Signed-off-by: Pierre-Yves Kerbrat <pkerbrat@free.fr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
wpa_ielen was declared u32 when we actually use an int
Fix sparse (-Wtypesign) issues:
drivers/staging/rtl8188eu/core/rtw_ieee80211.c:1008:60: warning: incorrect type in argument 2 (different signedness)
drivers/staging/rtl8188eu/core/rtw_ieee80211.c:1008:60: expected int *wpa_ie_len
drivers/staging/rtl8188eu/core/rtw_ieee80211.c:1008:60: got unsigned int *<noident>
drivers/staging/rtl8188eu/core/rtw_ieee80211.c:1021:69: warning: incorrect type in argument 2 (different signedness)
Signed-off-by: Pierre-Yves Kerbrat <pkerbrat@free.fr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix the following warnings:
Block comments should align the * on each line
Signed-off-by: Abdul Rauf <abdulraufmujahid@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
