This reverts the series of commits
67dbb4ea33281ab031a847807ce381
that changed the GART VM start offset. It fixed some machines, but
seems to continually interact badly with some X versions.
Quoth Ben Herrenschmidt:
"So I think at this point, the best is that we keep the old bogus code
that at least is consistent with the bug in the server. I'm working on a
big patch to X that reworks the memory map stuff completely and fixes
those issues on the server side, I'll do a DRM patch matching this X fix
as well so that the memory map is only ever set in one place and with
what I hope is a correct algorithm..."
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
As reported by Jules Villard <jvillard@ens-lyon.fr> and some others, the
recent GART aperture start reconfiguration causes problems on some
setups.
What I _think_ might be happening is that the X server is also trying to
muck around with the card memory map and is forcing it back into a wrong
setting that also happens to no longer match what the DRM wants to do
and blows up. There are bugs all over the place in that code (and still
some bugs in the DRM as well anyway).
This patch attempts to avoid that by using the largest of the 2 values,
which I think will cause it to behave as it used to for you and will
still fix the problem with machines that have an aperture size smaller
than the video memory.
Acked-by: Jules Villard <jvillard@ens-lyon.fr>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fix a typo which breaks radeon drm compilation with gcc 2.95.3.
The offending line was added back in 2.6.11-rc3, but was harmless
back then. A recent addition nearby changed it into a compilation
breaker: commit 281ab031a8.
The doubled semi-colon ends up being an empty instruction, and the
variable declaration thus ends up being in the middle of "code".
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Mark M. Hoffman <mhoffman@lightlink.com>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This finally fixes the radeon memory mapping bug that was incorrectly
fixed by the previous patch. This time, we use the actual vram size as
the size to calculate how far to move the AGP aperture from the
framebuffer in card's memory space.
If there are still issues with this patch, they are due to bugs in the X
driver that I'm working on fixing too.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Mark M. Hoffman <mhoffman@lightlink.com>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ben noticed that on certain cards we've landed the AGP space on top of
the second aperture instead of after it.. Which messes things up a lot
on those machines.
This just moves the gart further out, a more correct fix is in the works
from Ben for after 2.6.15.
Signed-off-by: Dave Airlie <airlied@linux.ie>
CC: Ben Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This fixes a NULL pointer reference in DRM. The SiS driver tries to
allocate a big chunk of memory, but the return value is never checked.
Reported in Novell bugzilla #132271:
https://bugzilla.novell.com/show_bug.cgi?id=132271
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
A fix for a locking bug which is triggered when a client tries to lock with
flag DMA_QUIESCENT (typically the X server), but gets interrupted by a signal.
The locking IOCTL should then return an error, but if DMA_QUIESCENT succeeds
it returns 0, and the client falsely thinks it has the lock. In addition
The client waits for DMA_QUISCENT and possibly DMA_READY without having the lock.
From: Thomas Hellstrom
Signed-off-by: Dave Airlie <airlied@linux.ie>
We memset the structure across opens except for the flags. The correct
fix is more intrusive but this should fix a problem with bad iounmaps
seen on AGP radeons acting like PCI ones.
Signed-off-by: Dave Airlie <airlied@linux.ie>
The DRM only uses drm_alloc_pages for non-SG PCI cards using DRM.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Dave Airlie <airlied@linux.ie>
PCI->PCI bridge, then bus->self is allowed to be NULL. Certainly that's
the case on my Pegasos, and it makes the MGA DRM driver oops...
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Add PCI DMA blitengine to VIA DRM
Add portability code for porting VIA to FreeBSD.
Sync via_drm.h with 3d driver
From: Thomas Hellstrom <unichrome@shipmail.org>, Eric Anholt <anholt@freebsd.org>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Alan Hourihane wants to set MTRR in the DDX only as otherwise
we get problems with the shared memory chipset.
From: Alan Hourihane <alanh@fairlite.demon.co.uk>
Signed-off-by: Dave Airlie <airlied@linux.ie>
This simplifies the sysfs code for the drm and add a dri_library_name
attribute which can be used by a userspace app to figure out which
library to load.
From: Jon Smirl <jonsmirl@gmail.com>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Rename the driver hooks in the DRM to something a little more understandable:
preinit -> load
postinit -> (removed)
presetup -> firstopen
postsetup -> (removed)
open_helper -> open
prerelease -> preclose
free_filp_priv -> postclose
pretakedown -> lastclose
postcleanup -> unload
release -> reclaim_buffers_locked
version -> (removed)
postinit and version were replaced with generic code in the Linux DRM (drivers
now set their version numbers and description in the driver structure, like on
BSD). postsetup wasn't used at all. Fixes the savage hooks for
initializing and tearing down mappings at the right times. Testing involved at
least starting X, running glxgears, killing glxgears, exiting X, and repeating.
Tested on: FreeBSD (g200, g400, r200, r128)
Linux (r200, savage4)
From: Eric Anholt <anholt@freebsd.org>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Fix the PCIGART increment and add a cpu_to_le32 for ppc (untested)
Paulus was unsure if we need to cpu_to_le32 but the old code was definitely
wrong, so make it consistent and let the PPC guys figure it out later.
Signed-off-by: Dave Airlie <airlied@linux.ie>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Dave Jones <davej@codemonkey.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is the drivers/char/ part of the big kfree cleanup patch.
Remove pointless checks for NULL prior to calling kfree() in drivers/char/.
Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I recently picked up my older work to remove unnecessary #includes of
sched.h, starting from a patch by Dave Jones to not include sched.h
from module.h. This reduces the number of indirect includes of sched.h
by ~300. Another ~400 pointless direct includes can be removed after
this disentangling (patch to follow later).
However, quite a few indirect includes need to be fixed up for this.
In order to feed the patches through -mm with as little disturbance as
possible, I've split out the fixes I accumulated up to now (complete for
i386 and x86_64, more archs to follow later) and post them before the real
patch. This way this large part of the patch is kept simple with only
adding #includes, and all hunks are independent of each other. So if any
hunk rejects or gets in the way of other patches, just drop it. My scripts
will pick it up again in the next round.
Signed-off-by: Tim Schmielau <tim@physik3.uni-rostock.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I've seen similar failure on alpha.
Obviously, someone forgot to convert sg->handle stuff for
PCI gart case.
Signed-off-by: Dave Airlie <airlied@linux.ie>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The wrong state emission routines were being called for G550, and
consistent maps weren't correctly mapped...
Signed-off-by: Dave Airlie <airlied@linux.ie>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I've gotten a report on lkml, of a possible regression in the MGA DRM in
2.6.14-rc4 (since -rc1), I haven't been able to reproduce it here, but I've
figured out some possible issues in the mga code that were definitely
wrong, some of these are from DRM CVS, the main fix is the agp enable bit
on the old code path still used by everyone.....
Signed-off-by: Dave Airlie <airlied@linux.ie>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
In order to get some better debugging from people about certain hangs/crashes
we need to be able to turn AGP writeback off permanently...
Signed-off-by: Dave Airlie <airlied@linux.ie>
A bunch of create_proc_dir_entry() calls creating directories had crept
in since the last sweep; converted to proc_mkdir().
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I've been threatening this for a while, so no point hanging around.
This lindents the DRM code which was always really bad in tabbing department.
I've also fixed some misnamed files in comments and removed some trailing
whitespace.
Signed-off-by: Dave Airlie <airlied@linux.ie>
Add support for Radeon PCI Express cards (needs a new X.org DDX)
Also allows PCI GART table to be stored in VRAM for non PCIE cards
Signed-off-by: Dave Airlie <airlied@linux.ie>
Add support for GL_ATI_fragment_shader, new packets R200_EMIT_PP_AFS_0/1,
R200_EMIT_PP_TXCTLALL_0-5 (replaces R200_EMIT_PP_TXFILTER_0-5, 2 more regs)
and R200_EMIT_ATF_TFACTOR (replaces R200_EMIT_TFACTOR_0 (8 consts instead of 6)
From: Roland Scheidegger, David Airlie
Signed-off-by: David Airlie <airlied@linux.ie>
This patch contains the following small cleanups:
- make two needlessly global functions static
- drm_sysfs.c: every file should #include the header with the prototypes
of the global functions it is offering
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Dave Airlie <airlied@linux.ie>
On 32-bit PPC a 0 handle is valid for AGP space, the 32/64 lookup
doesn't handle 0 correctly.
From: Ben Herrenschmidt <benh@kernel.crashing.org> and Paul Mackerras <paulus@samba.org>
Signed-off-by: Dave Airlie <airlied@linux.ie>
I found why my G5 was crashing when using the linux-2.6 version of the
DRM + git-drm.patch from 2.6.13-rc6-mm1, but not with the CVS DRM.
The reason was that dev->agp->cant_use_aperture wasn't getting set,
and the reason for that was that <linux/version.h> no longer gets
included and the #if LINUX_VERSION_CODE < 0x020408 in drm_agpsupport.c
was going the wrong way. With this patch (and a few others) a 32-bit
server works correctly, as does DRI.
From: Paul Mackerras <paulus@samba.org>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Remove a bogus check on whether an area is memory (we need a better interface)
also change pgprot flags for powerpc
don't check on x86-64 either
From: Paul Mackerras <paulus@samba.org>
Signed-off-by: Dave Airlie <airlied@linux.ie>
This adds initial r300 3D support to the radeon DRM.
From: Nicolai Haehnle, Vladimir Dergachev, and others.
Signed-off-by: David Airlie <airlied@linux.ie>
The gamma driver has been broken for quite a while, it doesn't build,
we don't have a userspace, mine is in Ireland etc...
Signed-off-by: Dave Airlie <airlied@linux.ie>
This converts the drm_handle_t to unsigned int.
This is currently safe to do as we don't pass these across the kernel/user
boundary, but userspace does use these, but no-one builds userspace against
the kernel headers at present so it is okay to switch over the kernel copy
of drm.h at this point. (The CVS tree will switch over soon in sync with
some Mesa changes)
From: Egbert Eich <eich@suse.de>
Signed-off-by: Dave Airlie <airlied@linux.ie>
I basically combined Paul's patches with additions that I had made
for PCI scatter gather.
I also tried more carefully to avoid problems with the same token
assigned multiple times while trying to use the base address in the
token if possible to gain as much backward compatibility as possible
for broken DRI clients.
From: Paul Mackerras <paulus@samba.org> and Egbert Eich <eich@suse.de>
Signed-off-by: Dave Airlie <airlied@linux.ie>
This fixes the information copied back to userspace by the get reserved
contexts ioctl.
From: Egbert Eich <eich@suse.de>
Signed-off-by: Dave Airlie <airlied@linux.ie>
This is a patch from DRM CVS that cleans up some code that was in CVS
that I never moved to the kernel, this patch produces the result of the
cleanups and puts it into the kernel drm.
From: Eric Anholt <anholt@freebsd.org>, Jon Smirl, Dave Airlie
Signed-off-by: Dave Airlie <airlied@linux.ie>
This patch adds serveral new ioctls and a new query to get_param query to
support PCI MGA cards.
Two ioctls were added to implement interrupt based waiting. With this change,
the client-side driver no longer needs to map the primary DMA region or the
MMIO region. Previously, end-of-frame waiting was done by busy waiting in the
client-side driver until one of the MMIO registers (the current DMA pointer)
matched a pointer to the end of primary DMA space. By using interrupts, the
busy waiting and the extra mappings are removed.
A third ioctl was added to bootstrap DMA. This ioctl, which is used by the
X-server, moves a *LOT* of code from the X-server into the kernel. This allows
the kernel to do whatever needs to be done to setup DMA buffers. The entire
process and the locations of the buffers are hidden from user-mode.
Additionally, a get_param query was added to differentiate between G4x0 cards
and G550 cards. A gap was left in the numbering sequence so that, if needed,
G450 cards could be distinguished from G400 cards. According to Ville
Syrjälä, the G4x0 cards and the G550 cards handle anisotropic filtering
differently. This seems the most compatible way to let the client-side driver
know which card it's own. Doing this very small change now eliminates the
need to bump the DRM minor version twice.
http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2
(airlied - this may not work at this point, I think the follow on buffer
cleanup patches will be needed)
From: Ian Romanick <idr@us.ibm.com>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Added device_is_agp callback to drm_driver. This function is called by the
platform-specific drm_device_is_agp function. Added implementation of this
function the the Linux-specific portion of the MGA driver to detect PCI G450
cards. Added code to the Linux-specific portion of the generic DRM layer to
not initialize AGP infrastructure if the card is not AGP (this matches what
already existed in BSD).
Fix up i810/i830 and i915 drivers to always return AGP as they don't always
report the capability.
Fix the MGA to not report AGP for a card that has an AGP chip behind a PCI
bridge.
From: Ian Romanick, Dave Airlie, Alan Hourihane
Signed-off-by: Dave Airlie <airlied@linux.ie>
This patch splits some ioctl functions so that they can be called
in-kernel by a DRM driver. The driver will use them later.
From: Ian Romanick <idr@us.ibm.com>
Signed-off-by: Dave Airlie <airlied@linux.ie>
The DRM needs to change the drm_pci interface for FreeBSD compatiblity,
this patch introduces the drm_dma_handle_t and uses it in the Linux code.
From: Tonnerre Lombard, Eric Anholt, and Sergey Vlasov
Signed-off-by: David Airlie <airlied@linux.ie>
drm_addbufs. This makes the code more like the BSD code, and makes the
drm_addbufs_* functions callable in-kernel.
From: Ian Romanick <idr@us.ibm.com>
Signed-off-by: Dave Airlie <airlied@linux.ie>
The patch makes drmAddBufs/drmMapBufs can handle buffers in video memory
The attached patch adds a new buffer type DRM_FB_BUFFER. It works like
AGP memory but uses video memory.
From: Austin Yuan <austinyuan@viatech.com.cn>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Added a new DRM map type _DRM_CONSISTENT for consistent PCI memory. It
uses drm_pci_alloc/free for allocating/freeing the memory.
From: Felix Kuhling <fxkuehl@gmx.de>
Signed-off-by: David Airlie <airlied@linux.ie>
This adds compatiblity ioctls for mga/r128 and i915 DRM drivers.
From: Paul Mackerras, David Airlie, Alan Hourihane, Egbert Eich.
Signed-off-by: David Airlie <airlied@linux.ie>
This patch contains the following cleanups:
- make needlessly global functions static
- remove the following unused global functions:
- drm_fops.c: drm_read
- i915_dma.c: i915_do_cleanup_pageflip
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Add DRM device driver for VIA Unichrome chipsets
From: Unichrome Project http://unichrome.sf.net, Erdi Chen, Thomas Hellstrom Signed-off-by: Dave Airlie <airlied@linux.ie>
After the previous fix in 2.6.12, this patch should properly fix the
radeon IRQ handling code.
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Dave Airlie <airlied@linux.ie>
The patch is against a 2.6.11 kernel tree. I am running this with a
32-bit X server (compiled up from X.org CVS as of a couple of weeks
ago) and 32-bit DRI libraries and clients. All the userland stuff is
identical to what I am using under a 32-bit kernel on my G4 powerbook
(which is a 32-bit machine of course). I haven't tried compiling up a
64-bit X server or clients yet.
In the compatibility routines I have assumed that the kernel can
safely access user addresses after set_fs(KERNEL_DS). That is, where
an ioctl argument structure contains pointers to other structures, and
those other structures are already compatible between the 32-bit and
64-bit ABIs (i.e. they only contain things like chars, shorts or
ints), I just check the address with access_ok() and then pass it
through to the 64-bit ioctl code. I believe this approach may not
work on sparc64, but it does work on ppc64 and x86_64 at least.
One tricky area which may need to be revisited is the question of how
to handle the handles which we pass back to userspace to identify
mappings. These handles are generated in the ADDMAP ioctl and then
passed in as the offset value to mmap. However, offset values for
mmap seem to be generated in other ways as well, particularly for AGP
mappings.
The approach I have ended up with is to generate a fake 32-bit handle
only for _DRM_SHM mappings. The handles for other mappings (AGP, REG,
FB) are physical addresses which are already limited to 32 bits, and
generating fake handles for them created all sorts of problems in the
mmap/nopage code.
This patch has been updated to use the new compatibility ioctls.
From: Paul Mackerras <paulus@samba.org>
Signed-off-by: Dave Airlie <airlied@linux.ie>
This removes a bogus hack from the radeon IRQ handler.
There is a better fix from myself and benh in DRM CVS but I'll wait
until 2.6.13-rc so it gets more testing.
Signed-off-by: Dave Airlie <airlied@linux.ie>
drm: fix r128_state.c switch statements.. in drivers/char/drm/r128_state.c
(linux-2.6.12-rc2), some breaks are missing in the switch statement. See
trivial fix below.
Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
