| Age | Commit message (Collapse) | Author |
|---|
|
While I haven't tested the imaging extension, this matches what 915 does.
|
|
This removes the delayed texture upload optimization from 965, in exchange for
bringing us closer to PBO support. It also disables SGIS_generate_mipmap,
which didn't seem to be working before anyway, according to the lodbias demo.
|
|
|
|
965 gains fixed TTM typing of the buffer object buffers and unused PBO
functions, and 915 gains buffer size == 0 fixes from 965.
|
|
This adds (so far) unused PBO functions, and holding the lock while writing
to regions (which may be shared static screen regions).
|
|
|
|
|
|
|
|
The idling it was trying to ensure was covered by the
intel_miptree_image_map()->intel_region_map() that immediately followed it.
|
|
|
|
|
|
|
|
Each buffer object now has a relocation buffer pointer, which contains the
relocations for the buffer if there are any. At the point where we have to
create a new type of relocation entry, we can change the code over to allowing
multiple relocation lists, but trying to anticipate what that'll look like
now just increases complexity.
This is a 30% performance improvement on 965.
|
|
The user-space suballocator that was used avoided relocation computations by
using the general and surface state base registers and allocating those types
of buffers out of pools built on top of single buffer objects. It also
avoided calls into the buffer manager for these small state allocations, since
only one buffer object was being used.
However, the buffer allocation cost appears to be low, and with relocation
caching, computing relocations for buffers is essentially free. Additionally,
implementing the suballocator required a don't-fence-subdata flag to disable
waiting on buffer maps so that writing new data didn't block on rendering using
old data, and careful handling when mapping to update old data (which we need
to do for unavoidable relocations with FBOs). More importantly, when the
suballocator filled, it had no replacement algorithm and just threw out all
of the contents and forced them to be recomputed, which is a significant cost.
This is the first step, which just changes the buffer type, but doesn't yet
improve the hash table to not result in full recompute on overflow. Because
the buffers are all allocated out of the general buffer allocator, we can
no longer use the general/surface state bases to avoid relocations, and they
are set to 0 instead.
|
|
Now that the dri_bufmgr is stored in the context rather than the screen, all
access to one is single-threaded anyway.
|
|
|
|
bufmgr_fake doesn't care about it, but with ttm we would end up with the
buffer remaining referenced until application exit.
|
|
|
|
Changed the text for the default demos to be "auto" instead of "yes"
since that didn't make much sense in this context. Added an example
value for the --with-dri-drivers option as suggested by Brian.
|
|
Point bug reports to Bugzilla like bugs.html. Also, a warning from
autoconf w.r.t. AC_OUTPUT was cleaned up.
|
|
Brian suggested that there should be more consistency using xlib vs. x11
in the configure script. Changed the options and variables to suit.
|
|
Allow the user to disable use of assembly code through the --disable-asm
option. This is only relevant on the platforms where assembly is
normally enabled such as Linux and FreeBSD.
|
|
|
|
I broke this with cherry-pick resolving on
93c98a466947570e0589b662df49095b2f4bc43c.
|
|
Putting the bufmgr in the screen is not thread-safe since the emit_reloc
changes. It also led to a significant performance hit from pthread usage
for the attempted thread-safety (up to 12% of a cpu spent on refcounting
protection in single-threaded 965). The motivation had been to allow
multi-context bufmgr sharing in classic mode, but it wasn't worth the cost.
|
|
Report the compiler flags and macros that will be used in the build.
This just provides a quick way to see what configure has silently been
doing.
|
|
Allow the user to set the compiler debug flags and macros through the
option --enable-debug. This addes -DDEBUG to the macros and -g to the
CFLAGS and CXXFLAGS if gcc and g++ are in use.
|
|
Added a check that PKG_PROG_PKG_CONFIG is defined before calling it so
the user can see a proper error. Without it, the generated configure
script will just pass on to the next statement and bomb with a strange
shell syntax error.
|
|
|
|
This takes advantage of the DRM_BO_HINT_PRESUMED_OFFSET change and allows
the kernel to avoid mapping and re-writing buffers when relocations occur.
|
|
Rather than having the user run `make autoconf' and have the Makefile
setup the configs/current -> autoconf symlink, we can just do that in
configure. This allows the user to just run `make' to build.
|
|
Since install-sh is only used to satisfy an autoconf requirement, we
don't need the actual script. Instead, it's now just a symlink to
minstall, which provides a similar install wrapper script.
|
|
Make it clearer what the defaults for the AC_ARG_ENABLE options are and
how to change them.
|
|
Sync the default DRI_DIRS with the configs in master. Added some of the
comments from there, too.
|
|
Don't link glut with libXt since it's not needed.
(see commit ce98779571eee3f51d9f571fecf8deb83dd60f78)
|
|
|
|
|
|
|
|
This is currently believed to work but be a significant performance loss.
Performance recovery should be soon to follow.
The dri_bo_fake_disable_backing_store() call was added to allow backing store
disable like bufmgr_fake.c did, which is a significant performance win (though
it's missing the no-fence-subdata part).
This commit is a squash merge of the 965-ttm branch, which had some history
I wanted to avoid pulling due to noisiness and brokenness at many points
for git-bisecting.
|
|
|
|
|
|
Report some of the common settings back to the user after configure
has completed.
|
|
The user can request specific DRI drivers to build rather than the
default of "all that build on this platform". This allows the list of
drivers to be easily slimmed down.
This is controlled through the option --with-dri-drivers. For example:
./configure --with-driver=dri --with-dri-drivers="i965,nouveau"
Unfortunately, using this setting means the DRI drivers aren't filtered
by platform anymore and you might try to build something that doesn't
work.
|
|
Allow the user to specify that they want static libraries through the
--{enable,disable}-{static,shared} switches like libtool. The mesa build
only allows for one at a time, so static will be chosen if someone has
passed --enable-static or --disable-shared.
This also allows the mklib options to be set at build time. This allows
-static to be set for mklib, but any platform specific settings are
allowed by setting MKLIB_OPTIONS for configure.
Handling of the program libraries through the APP_LIB_DEPS variable is
pretty ugly, but it seems to work.
|
|
The user can request specific demos directories to build in. For
example:
./configure --with-demos="demos,xdemos"
The drawback is that we don't check for the necessary libararies in
that case, only that the directory in progs/ exists.
|
|
Allow the user to specify channel bits of 16 or 32 to enable OSMesa16 or
OSMesa32 instead of the default OSMesa. This option is controlled
through the option --with-osmesa-bits=BITS and is only honored when the
driver is osmesa.
The osdemos are not enabled in the 16 or 32 bit case because the
Makefile is currently hardcoded to link to -lOSMesa.
|
|
Added autoconf support for using OSMesa as the driver instead linking
it to libGL. This is enabled through --with-driver=osmesa.
To differentiate these cases, another option --enable-x11-osmesa is used
to enable or disable building OSMesa when the driver is x11.
|
|
Added autoconf bits to allow using DRI as the driver through the option
--with-dri-driver=DRIVER. The options are x11 (default) and dri. Three
DRI specific options for controlling the driver directory, direct
rendering and TLS are also added.
The DRI will probably not work for platforms besides linux and freebsd.
|
|
Add targets to the top Makefile so that configure and the necessary
scripts are added to the tarball for distribution.
Variables are used for autoconf and aclocal in case anyone needs to use
a specific version or pass any extra flags.
|
|
This adds the initial support for using autoconf configuration. Support
is available for shared Xlib driver builds. Later this will be extended
to dri and osmesa-only builds and possibly targetting non-X backends.
Support for static library builds will also be added.
The configure script fills in the autoconf config. This is then used by
running `make autoconf' after ./configure.
Testing has been done on Linux/GNU. The configure script tries to
faithfully reproduce the current configs/linux* and configs/freebsd*.
Other platforms can be handled later by adding similar statements and
feature tests.
Pkg-config is used to search for packages when possible. This makes the
build much more flexible and robust to the user's configuration. This
requires that the pkg-config autoconf macros pkg.m4 are included in
aclocal.m4. This requires autoconf and aclocal from autoconf and
automake, respectively.
|
