| Age | Commit message (Collapse) | Author |
|---|
|
ok mglocker@
|
|
|
|
ok deraadt@
|
|
at attach time for devices which set bDefaultFrameIndex larger than
the existing frame descriptors.
Spotted, tested, and OK yuo@
|
|
running before we play with things.
Lots more cleanup needed, but now you can shut X without it crashing.
|
|
|
|
ok oga@
|
|
|
|
|
|
able to give them to me.
|
|
shaves a few bytes, but that's just gravy.
|
|
|
|
broke radeondrm. Fixing it is ugly, so another change will have to be made
later.
I /hate/ the drm_buf api, it will die as soon as it can.
|
|
worked on with djm
|
|
9k backend allocator is available.
vic is the first driver to properly support "jumbo" frames. this is the
model every other driver should follow.
|
|
are fairly small. tested by a few, and kettenis says he will test it sooner
if it is commited...
|
|
Some device could not handle the device descriptor request
with short bLength (less than the device descriptor size(18 byte)).
To handle such kind of devices, add complemental request loop to
current logic.
(memo: and I also found a device which can not response to any device
descriptor request. It always returns only 8 byte-length head of the
descriptor for 1st query. This is the reason why the device descriptor
loop of this usb subsystem start with 8-byte length query.)
|
|
this causes vic to start up with 4 mbufs on the ring, and the allocator
will let it increase the number of rx mbufs as usage demands.
testing shows that i need only 20 to 30 mbufs on the rx ring to
cope with full speed io. we used to always put 100 on the ring, so
we're now saving 140k of kernerl virtual address space (70 *
MCLBYTES).
this relies on the previous commit that copes with failures of mbuf
allocation in the rx path.
similar changes must be made to all the other drivers. i'll happily review
diffs people send in. SUBTLE HINTS ARE COMING YOUR WAY.
|
|
fill a slot when the mbuf allocator gives us an mbuf cluster.
|
|
|
|
|
|
and detach devices on a gpiobus at runtime. The offset and mask locators
in kernel config files can still be used for static configuration, so this
does not break existing custome kernel configs.
- Have gpioow(4) unmap the pins it used from the gpiobus during detach
Changes to gpioctl(8) will be a separate committ.
ok uwe
|
|
This lets early arc4random() calls to return at least non-uniqe values
on different machines. The dmesg will have different MAC addresses,
etc. Also ignore the entropy count in arc4_stir, and just take what's
there. sizeof(struct timeval) should be sizof(struct timespec)
This was forgotten when we switched to nanotime.
With help from mickey.
OK djm@, deraadt@
|
|
|
|
Someone must have fumbled with this line by accident while doing the
original port of the driver.
|
|
|
|
Add SDMMC_ASSERT_LOCKED(), replacing many locking calls, and sprinkling
the assertion to make sure that future changes don't catch us out
"please commit" uwe@
|
|
dma_bufs; allow the driver to provide the size and not have the private
data pointer. only radeon only needs this so far, but intel with GEM
also needs it.
Postclose was only used for freeing said private data, so this allows me
to remove the postclose callback and rename preclose to close.
|
|
ask the driver how large they need the structure we allocate to be, and
use inheritance like we do for struct device. Simplifies things a little
bit and saves us a pointer.
|
|
|
|
|
|
|
|
|
|
setup the drm_attach_args with the needed information.
require interface version 1.1 libdrm has been requesting it for ages),
which means that we can set the busid string at attach time. (generated in
drm_attach_pci), also pass in the interrupt line and bst.
Now we don't need dev->pa at all, so remove it, finally.
|
|
other way round. More dev->pa reduction
|
|
not zero it during the cleanup ioctl.
|
|
|
|
the ioctl handler.
|
|
on X startup.
found by ckuethe@
|
|
"I'm all for it." marco@ "Yeah" deraadt@
|
|
|
|
|
|
part of a server. there's no configuration yet, and several other drawbacks,
but it can be hammered into shape. i haven't moved the code forward in a year,
and marco wants it in the tree to hack on.
|
|
|
|
|
|
size.
kill (empty) drm_mem_{un,}init() while here.
|
|
While we don't have code for this (yet!), mean I will be able to update
libdrm, and consequently the xf86-video-intel driver to 2.5.x.
Add PARAM_HAS_GEM to the getparam ioctl, and return no support.
|
|
ok yuo@
|
|
dev->pa by doing the pci_intr_establish/disestablish dance in the driver
function, not in drm. This removes the need for
interrupt_{pre,post}install callbacks, instead just provide a
interrupt_install() callback.
|
|
just put a vblank_pipes argument to the driver structure which tells us
how many to set up
this means that intel doesn't lose all vblank count on vtswitch (it
disables the interrupt there), i've heard of a few uvm_faults where this
happens as well as things just going wrong.
This was suggested by Keith Packard who provided a different diff for
drm.git.
|
