percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Specify Kconfig dependencies, and include warnings for building as a module
udlfb is dependent on FB_DEFERRED_IO, FB_SYS_*, and FB_MODE_HELPERS
Because many kernels do not include defio (which cannot be built
as a module), yet users want to be able to build udlfb as a module later,
udlfb has ifdefs and these dependency warnings to help udlfb build with or
without certain dependencies, but also print warnings for any lost function.
Even though this kind of flexibility isn't common, we've gotten feedback
from a significant portion of users that they were frustrated without it.
Signed-off-by: Bernie Thompson <bernie@plugable.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Add support for fbdev mmap clients who don't send damage ioctls
Because DisplayLink devices are out on the other end of usb, their
"framebuffer" is just normal system memory. So memory mapped writes
don't automatically trigger anything. So up to this point, standard
fbdev clients who rely on mmap() will get an unchanging screen.
This patch makes udlfb a client of Jaya Kumar's defio framework - which sets
up page fault triggers, and those faults are accumulated and sent to udlfb
on a defferred basis, to process as damage notifications for the framebuffer.
Because this involves more overhead than a notification directly from
the application (e.g. just passing on X DAMAGE extension notifications),
a sysfs attribute is provided by udlfb to control defio support.
/sys/class/graphics/fb*/use_defio - writing a "0" to this file before
calling mmap() causes defio to not be initialized - instead udlfb
will rely on getting damage notifications directly through the damage ioctl.
There are unsolved rendering problems with defio (horizontal dead regions
on framebuffer, that accumulate over time) which still needs a fix.
Signed-off-by: Bernie Thompson <bernie@plugable.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Rework rendering for improved performance
Approx 10-20% gain avg across several "benchmarks": x11perf, gtkperf, glxgears
Moves from a single pre-alloc'd urb protected by a long-held mutex
To a list of (4) pre-alloc'd urbs which can be dispatched asynchonously
Improved rendering algorithm to hardware with lower CPU consumption,
fewer system memory accesses, and slightly higher compression.
Better scalability to multiple processors, especially with
multiple framebuffers active.
Signed-off-by: Bernie Thompson <bernie@plugable.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Rework probe to use refcounts and std functions
Because the different parts of the driver (usb, fbdev) tear down
in different orders, the driver previously could crash accessing
data that had already been freed. Refcounting system used to handle.
Reworked probe to make use of refcounts, set mode using std fbops,
and set up sysfs and pre-allocated urbs.
Signed-off-by: Bernie Thompson <bernie@plugable.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Add sysfs controls for edid and performance metrics
There are 8 new files exposed in /sys/class/graphics/fb*
edid - returns 128 byte edid blog, suitable for parsing with parse-edid
metrics_bytes_identical
metrics_bytes_rendered
metrics_bytes_sent
metrics_cpu_kcycles_used
metrics_misc
and metrics_reset, which resets all perf metrics to zero
The 6 perf metrics are of type atomic_t.
So these metrics return precise results for short benchmarks, but
any test approx a minute or longer runtime may roll over.
Signed-off-by: Bernie Thompson <bernie@plugable.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Move to more consistent naming scheme
All udlfb functions start with udlfb_
All functions for udlfb's fbdev interface start with udlfb_ops_
All functinos for udlfb's usb interface start with udlfb_usb_
Signed-off-by: Bernie Thompson <bernie@plugable.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Add functions to pre-allocate and free usb bulk urbs for core render path.
Udlfb currently allocates a single urb, guarded by a mutex, that is a key
bottleneck. Because udlfb sends so much data, preallocation is most efficient.
Functions will be used by new rendering functions in later patches.
Signed-off-by: Bernie Thompson <bernie@plugable.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Reorganize the location of a few things to be closer to related code
Signed-off-by: Bernie Thompson <bernie@plugable.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Eliminate checkpatch.pl warnings and errors so later patches in series are clean
Signed-off-by: Bernie Thompson <bernie@plugable.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Add dynamic modeset support
udlfb uses EDID to find the monitor’s preferred mode
udlfb no longer has fixed mode tables – it’s able to set any mode
dynamically, from the standard VESA timing characteristics of the monitor.
Draws from probe and setmode code of both displaylink-mod 0.3 branch of
Roberto De Ioris, and Jaya Kumar's displaylinkfb.
Lays foundation for defio support and making backbuffer optional.
With additional changes to minimize diffs and clean for checkpatch.pl style.
Does not yet include new ioctls or refcount/mutex code from displaylink-mod.
Tested to work with existing xf-video-displaylink X server unmodified.
Signed-off-by: Bernie Thompson <bernie@plugable.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Fix printk format warning: use %td for ptrdiff:
drivers/staging/udlfb/udlfb.h:209: warning: format '%d' expects type 'int', but argument 3 has type 'long int'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Required for vmalloc_32 and vfree declarations on non-x86 platforms.
Signed-off-by: Amit Kucheria <amit.kucheria@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
There are others remaining due to the __iomem namespace of the
framebuffer data pointer.
Cc: Roberto De Ioris <roberto@unbit.it>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This cleans up a bunch of checkpatch.pl warnings in the udlfb.c file.
Cc: Roberto De Ioris <roberto@unbit.it>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This cleans up a bunch of checkpatch.pl warnings in the udlfb.h file.
Cc: Roberto De Ioris <roberto@unbit.it>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This adds the udlfb driver, a framebuffer driver for DisplayLink devices.
From: Roberto De Ioris <roberto@unbit.it>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Tejun Heo