[PATCH 3/6] drm/radeon: Writeback endian fixes

On Wed, Jul 13, 2011 at 5:41 PM, Benjamin Herrenschmidt wrote: > On Wed, 2011-07-13 at 10:42 -0400, Alex Deucher wrote: >> On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt >> wrote: >> > The writeback ring pointer and IH ring pointer are read using le32_to_cpu >> > so we do not want the

[PATCH 4/6] drm/radeon: Add a rmb() in IH processing

On Wed, Jul 13, 2011 at 5:42 PM, Benjamin Herrenschmidt wrote: > On Wed, 2011-07-13 at 10:48 -0400, Alex Deucher wrote: >> On Wed, Jul 13, 2011 at 10:43 AM, Alex Deucher >> wrote: >> > On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt >> > wrote: >> >> We should have a read memory

[PATCH 2/6] drm/radeon: ATOM Endian fix for atombios_crtc_program_pll()

On Wed, Jul 13, 2011 at 5:40 PM, Benjamin Herrenschmidt wrote: > On Wed, 2011-07-13 at 10:38 -0400, Alex Deucher wrote: > >> > ? ? ? ? ? ? ? ?case 6: >> > - ? ? ? ? ? ? ? ? ? ? ? args.v6.ulCrtcPclkFreq.ucCRTC = crtc_id; >> > - ? ? ? ? ? ? ? ? ? ? ? args.v6.ulCrtcPclkFreq.ulPixelClock = >> >

[PATCH v2 1/2] radeon: Don't generate new fence for page flip.

2011/7/13 Michel D?nzer : > From: Michel D?nzer > > Use the fence of the new frontbuffer, if any. > > Generating a new fence could cause us to wait for completely unrelated > rendering to finish before performing the flip. > > Signed-off-by: Michel D?nzer Reviewed-by: Alex Deucher > --- > >

[PATCH] drm: platform multi-device support

Include the device id in the bus-id to give userspace a way to open the correct "cardN" when there are multiple device instances. Signed-off-by: Rob Clark --- drivers/gpu/drm/drm_platform.c |5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) diff --git

[PATCH 0/10] fs shader debugging

On Wed, Jul 13, 2011 at 01:51:42PM -0700, Ben Widawsky wrote: > > > Version 2 of the patch series is pretty much the same as version 1. 2 of the > patches have already been picked up by the kernel and mesa so they are > gone. > > The only major change is in mesa where I no longer load a binary

[PATCH 01/10] intel: shared header for shader debugging

On Wed, 13 Jul 2011 13:51:43 -0700, Ben Widawsky wrote: > > Signed-off-by: Ben Widawsky > --- > intel/Makefile.am |3 ++- > intel/intel_debug.h | 44 > 2 files changed, 46 insertions(+), 1 deletions(-) > > diff --git a/intel/Makefile.am

[PULL] drm-intel-next

art -- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 198 bytes Desc: Digital signature URL: <http://lists.freedesktop.org/archives/dri-devel/attachments/20110713/0ec2fe6c/attachment-0001.pgp>

[PATCH] i915: slab shrinker have to return -1 if it cant shrink any objects

On Wed, Jul 13, 2011 at 09:40:31AM +0100, Chris Wilson wrote: > On Wed, 13 Jul 2011 17:19:22 +0900, KOSAKI Motohiro jp.fujitsu.com> wrote: > > (2011/07/13 16:41), Chris Wilson wrote: > > > On Wed, 13 Jul 2011 09:19:24 +0900, KOSAKI Motohiro > > jp.fujitsu.com> wrote: > > >> (2011/07/12 19:06),

[PATCH] drm/i915: Add quirk to disable SSC on Lenovo U160 LVDS

On Wed, 13 Jul 2011 13:09:28 -0700, "Keith Packard" wrote: Non-text part: multipart/signed > > We've tried several times to make this machine 'just work', but every > patch that does causes many other machines to fail. This adds a quirk > which special cases this hardware and forces ssc to be >

[PATCH] i915: slab shrinker have to return -1 if it cant shrink any objects

On Wed, Jul 13, 2011 at 05:19:22PM +0900, KOSAKI Motohiro wrote: > (2011/07/13 16:41), Chris Wilson wrote: > >> (snip) > >> while (total_scan >= SHRINK_BATCH) { > >> long this_scan = SHRINK_BATCH; > >> int shrink_ret; > >>

drm_lock()->block_all_signals() is broken

On 07/12, Dave Airlie wrote: > > On Tue, Jul 12, 2011 at 7:15 PM, Oleg Nesterov wrote: > > Hello. > > > > I tried many times to ask about the supposed behaviour of > > block_all_signals() in drm, but it seems nobody can answer. > > Its not hard to explain, basically there exists hardware that you

[PULL] drm-intel-next

. Name: not available Type: application/pgp-signature Size: 198 bytes Desc: Digital signature URL: <http://lists.freedesktop.org/archives/dri-devel/attachments/20110713/1490f1ec/attachment.pgp>

[PATCH] i915: slab shrinker have to return -1 if it cant shrink any objects

(2011/07/13 16:41), Chris Wilson wrote: > On Wed, 13 Jul 2011 09:19:24 +0900, KOSAKI Motohiro jp.fujitsu.com> wrote: >> (2011/07/12 19:06), Chris Wilson wrote: >>> On Tue, 12 Jul 2011 18:36:50 +0900, KOSAKI Motohiro >> jp.fujitsu.com> wrote: Hi, sorry for the delay. > On

[PATCH 2/2] radeon: Don't clobber error return value in page flipping cleanup paths.

From: Michel D?nzer Signed-off-by: Michel D?nzer Reviewed-by: Alex Deucher --- drivers/gpu/drm/radeon/radeon_display.c |9 ++--- 1 files changed, 2 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_display.c

[PATCH v2 1/2] radeon: Don't generate new fence for page flip.

From: Michel D?nzer Use the fence of the new frontbuffer, if any. Generating a new fence could cause us to wait for completely unrelated rendering to finish before performing the flip. Signed-off-by: Michel D?nzer --- v2: * Rebase against drm-core-next. * Leave

[PATCH 6/6] drm/radeon: Fix the definition of RADEON_BUF_SWAP_32BIT

(Note that this is duplicated under various other names such as R600_BUF_SWAP_32BIT etc...). At least now all the definitions agree. Signed-off-by: Benjamin Herrenschmidt --- (resent adding dri-devel to the CC list to hit patchwork) drivers/gpu/drm/radeon/radeon_reg.h |2 +- 1 files

[PATCH 5/6] drm/radeon: Do an MMIO read on interrupts when not uisng MSIs

When not using MSIs, there is no guarantee that DMA from the device has been fully flushed to point where it's visible to the CPU when taking an interrupt. To get this guarantee, we need to perform an MMIO read from the device, which will flush all outstanding DMAs from bridges between the device

[PATCH 4/6] drm/radeon: Add a rmb() in IH processing

We should have a read memory barrier between reading the WPTR from memory and reading ring entries based on that value (ie, we need to ensure both loads are done in order by the CPU). It could be argued that the MMIO reads in r600_ack_irq() might be enough to get that barrier but I prefer keeping

[PATCH 3/6] drm/radeon: Writeback endian fixes

The writeback ring pointer and IH ring pointer are read using le32_to_cpu so we do not want the chip to byteswap them on big-endian. We still want to byteswap the ring itself and the IBs, so we don't touch that but we remove setting of the byteswap bits in CP_RB_RPTR_ADDR and IH_CNTL. In

[PATCH 2/6] drm/radeon: ATOM Endian fix for atombios_crtc_program_pll()

v6 of the structure was programmed incorrectly: args.v6.ulCrtcPclkFreq.ulPixelClock = cpu_to_le32(clock / 10); ulPixelClock is a 24-bit bitfield. This statement would thus do a 32-bit swap of (clock / 10) and drop the top 8 bits which are ... the LSB. Not what we want. Instead use masks &

[PATCH 1/6] drm/radeon: Remove a bunch of useless _iomem casts

Just defining rdev->rmmio properly in the first place should do the trick. In some cases, the cast were also complete dups as the original variable was already of the right type. Signed-off-by: Benjamin Herrenschmidt --- (resent adding dri-devel to the CC list to hit patchwork)

[Bug 39202] KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 --- Comment #5 from Siganderson 2011-07-13 15:27:24 PDT --- Created an attachment (id=49063) --> (https://bugs.freedesktop.org/attachment.cgi?id=49063) dmesg -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email ---

[Bug 39202] KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 --- Comment #4 from Siganderson 2011-07-13 15:26:57 PDT --- Created an attachment (id=49062) --> (https://bugs.freedesktop.org/attachment.cgi?id=49062) xorg conf -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email

[Bug 39202] KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 --- Comment #3 from Siganderson 2011-07-13 15:26:34 PDT --- Created an attachment (id=49061) --> (https://bugs.freedesktop.org/attachment.cgi?id=49061) xorg log -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email

[PATCH 01/10] intel: shared header for shader debugging

Signed-off-by: Ben Widawsky --- intel/Makefile.am |3 ++- intel/intel_debug.h | 44 2 files changed, 46 insertions(+), 1 deletions(-) diff --git a/intel/Makefile.am b/intel/Makefile.am index b6a9014..7a44aaf 100644 --- a/intel/Makefile.am

[PATCH 0/10] fs shader debugging

Version 2 of the patch series is pretty much the same as version 1. 2 of the patches have already been picked up by the kernel and mesa so they are gone. The only major change is in mesa where I no longer load a binary blob from an environment variable, but instead compile the bytes directly in

[Bug 39202] KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 --- Comment #2 from Alex Deucher 2011-07-13 13:22:55 PDT --- Please attach your xorg log and config and dmesg output. -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because:

[PATCH] drm/i915: Add quirk to disable SSC on Lenovo U160 LVDS

ntel_init_quirks(struct drm_device *dev) -- 1.7.5.4 -- keith.packard at intel.com -- next part -- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <http://lists.freedesktop.org/archives/dri-devel/attachments/20110713/3bfbe3e4/attachment.pgp>

[Bug 35988] Power management not working on ATI M92 LP [Mobility Radeon HD 4300]

https://bugs.freedesktop.org/show_bug.cgi?id=35988 --- Comment #9 from leonardo 2011-07-13 13:09:19 PDT --- I have the same card on DELL Studio 15". I wouldn't say that power management does not work. I'm using "profile" method with "low" setting. After about ~20 mins of uptime this is the

[Bug 39202] KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 --- Comment #1 from Siganderson 2011-07-13 12:25:57 PDT --- video card: ATI Radeon HD 5670 512 MB Ram -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are the

[Bug 39202] New: KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 Summary: KDE desktop effects with 3.0 rc6 kernel Product: Mesa Version: git Platform: All OS/Version: Linux (All) Status: NEW Severity: normal Priority: medium

[PATCH] drm/i915/bios: Avoid temporary allocation whilst searching for downclock

available URL: <http://lists.freedesktop.org/archives/dri-devel/attachments/20110713/2d5de019/attachment.pgp>

[PULL] drm-intel-next

t;http://lists.freedesktop.org/archives/dri-devel/attachments/20110713/640c0dae/attachment.pgp>

[PATCH 5/6] drm/radeon: Do an MMIO read on interrupts when not uisng MSIs

On Wed, Jul 13, 2011 at 10:44 AM, Alex Deucher wrote: > On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt > wrote: >> When not using MSIs, there is no guarantee that DMA from the device >> has been fully flushed to point where it's visible to the CPU when >> taking an interrupt. To get

[PATCH 4/6] drm/radeon: Add a rmb() in IH processing

On Wed, Jul 13, 2011 at 10:43 AM, Alex Deucher wrote: > On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt > wrote: >> We should have a read memory barrier between reading the WPTR from >> memory and reading ring entries based on that value (ie, we need to >> ensure both loads are done in

[PATCH 6/6] drm/radeon: Fix the definition of RADEON_BUF_SWAP_32BIT

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt wrote: > (Note that this is duplicated under various other names such > as R600_BUF_SWAP_32BIT etc...). At least now all the definitions > agree. > > Signed-off-by: Benjamin Herrenschmidt Reviewed-by: Alex Deucher > --- > > (resent

[PATCH 5/6] drm/radeon: Do an MMIO read on interrupts when not uisng MSIs

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt wrote: > When not using MSIs, there is no guarantee that DMA from the device > has been fully flushed to point where it's visible to the CPU when > taking an interrupt. To get this guarantee, we need to perform an > MMIO read from the

[PATCH 4/6] drm/radeon: Add a rmb() in IH processing

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt wrote: > We should have a read memory barrier between reading the WPTR from > memory and reading ring entries based on that value (ie, we need to > ensure both loads are done in order by the CPU). > > It could be argued that the MMIO reads

[PATCH 3/6] drm/radeon: Writeback endian fixes

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt wrote: > The writeback ring pointer and IH ring pointer are read using le32_to_cpu > so we do not want the chip to byteswap them on big-endian. > > We still want to byteswap the ring itself and the IBs, so we don't touch > that but we remove

[PATCH 2/6] drm/radeon: ATOM Endian fix for atombios_crtc_program_pll()

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt wrote: > v6 of the structure was programmed incorrectly: > > ?args.v6.ulCrtcPclkFreq.ulPixelClock = cpu_to_le32(clock / 10); > > ulPixelClock is a 24-bit bitfield. This statement would thus > do a 32-bit swap of (clock / 10) and drop the top

[PATCH 1/6] drm/radeon: Remove a bunch of useless _iomem casts

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt wrote: > Just defining rdev->rmmio properly in the first place should do > the trick. In some cases, the cast were also complete dups as > the original variable was already of the right type. > > Signed-off-by: Benjamin Herrenschmidt

[PATCH 1/2] radeon: Don't generate new fence for page flip.

On Mit, 2011-07-13 at 08:32 +0100, Dave Airlie wrote: > 2011/7/12 Michel D?nzer : > > From: Michel D?nzer > > > > Use the fence of the new frontbuffer, if any. > > What tree is this against, I can't apply it to drm-fixes or drm-core-next > here. I created it against 2.6.39.y. I'll respin

[PATCH] i915: slab shrinker have to return -1 if it cant shrink any objects

On Wed, 13 Jul 2011 17:19:22 +0900, KOSAKI Motohiro wrote: > (2011/07/13 16:41), Chris Wilson wrote: > > On Wed, 13 Jul 2011 09:19:24 +0900, KOSAKI Motohiro > jp.fujitsu.com> wrote: > >> (2011/07/12 19:06), Chris Wilson wrote: > >>> On Tue, 12 Jul 2011 18:36:50 +0900, KOSAKI Motohiro >>>

[PULL] drm-intel-next

-- next part -- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <http://lists.freedesktop.org/archives/dri-devel/attachments/20110713/ac763905/attachment.pgp>

[PATCH] i915: slab shrinker have to return -1 if it cant shrink any objects

(2011/07/12 19:06), Chris Wilson wrote: > On Tue, 12 Jul 2011 18:36:50 +0900, KOSAKI Motohiro jp.fujitsu.com> wrote: >> Hi, >> >> sorry for the delay. >> >>> On Wed, 29 Jun 2011 20:53:54 -0700, Keith Packard >>> wrote: On Fri, 24 Jun 2011 17:03:22 +0900, KOSAKI Motohiro >>> jp.fujitsu.com>

[PATCH] i915: slab shrinker have to return -1 if it cant shrink any objects

On Wed, 13 Jul 2011 09:19:24 +0900, KOSAKI Motohiro wrote: > (2011/07/12 19:06), Chris Wilson wrote: > > On Tue, 12 Jul 2011 18:36:50 +0900, KOSAKI Motohiro > jp.fujitsu.com> wrote: > >> Hi, > >> > >> sorry for the delay. > >> > >>> On Wed, 29 Jun 2011 20:53:54 -0700, Keith Packard > >>>

[PATCH 1/2] radeon: Don't generate new fence for page flip.

2011/7/12 Michel D?nzer : > From: Michel D?nzer > > Use the fence of the new frontbuffer, if any. What tree is this against, I can't apply it to drm-fixes or drm-core-next here. Dave.

[git pull] drm fixes

Hi Linus, Minor set of fixes from Alex including one hotplug regression and one intel AGP fix. Dave. The following changes since commit 620917de59eeb934b9f8cf35cc2d95c1ac8ed0fc: Linux 3.0-rc7 (2011-07-11 16:51:52 -0700) are available in the git repository at:

[Bug 39193] New: glCheckFramebufferStatusEXT segfaults in Gallium when checking status on a framebuffer bound to a texture that's bound to a pixmap

https://bugs.freedesktop.org/show_bug.cgi?id=39193 Summary: glCheckFramebufferStatusEXT segfaults in Gallium when checking status on a framebuffer bound to a texture that's bound to a pixmap Product: Mesa Version:

Re: [PATCH 1/2] radeon: Don't generate new fence for page flip.

On Mit, 2011-07-13 at 08:32 +0100, Dave Airlie wrote: 2011/7/12 Michel Dänzer mic...@daenzer.net: From: Michel Dänzer michel.daen...@amd.com Use the fence of the new frontbuffer, if any. What tree is this against, I can't apply it to drm-fixes or drm-core-next here. I created it

[git pull] drm fixes

Hi Linus, Minor set of fixes from Alex including one hotplug regression and one intel AGP fix. Dave. The following changes since commit 620917de59eeb934b9f8cf35cc2d95c1ac8ed0fc: Linux 3.0-rc7 (2011-07-11 16:51:52 -0700) are available in the git repository at:

Re: [PATCH] i915: slab shrinker have to return -1 if it cant shrink any objects

On Wed, 13 Jul 2011 17:19:22 +0900, KOSAKI Motohiro kosaki.motoh...@jp.fujitsu.com wrote: (2011/07/13 16:41), Chris Wilson wrote: On Wed, 13 Jul 2011 09:19:24 +0900, KOSAKI Motohiro kosaki.motoh...@jp.fujitsu.com wrote: (2011/07/12 19:06), Chris Wilson wrote: On Tue, 12 Jul 2011

Re: [PATCH] i915: slab shrinker have to return -1 if it cant shrink any objects

On Wed, 13 Jul 2011 09:19:24 +0900, KOSAKI Motohiro kosaki.motoh...@jp.fujitsu.com wrote: (2011/07/12 19:06), Chris Wilson wrote: On Tue, 12 Jul 2011 18:36:50 +0900, KOSAKI Motohiro kosaki.motoh...@jp.fujitsu.com wrote: Hi, sorry for the delay. On Wed, 29 Jun 2011 20:53:54 -0700,

drm_lock()-block_all_signals() is broken

Hello. I tried many times to ask about the supposed behaviour of block_all_signals() in drm, but it seems nobody can answer. So I am going to send the patch which simply removes block_all_signals() and friends. There are numeruous problems with this interace, I can't even enumerate them. But I

Re: [PATCH] i915: slab shrinker have to return -1 if it cant shrink any objects

(2011/07/12 19:06), Chris Wilson wrote: On Tue, 12 Jul 2011 18:36:50 +0900, KOSAKI Motohiro kosaki.motoh...@jp.fujitsu.com wrote: Hi, sorry for the delay. On Wed, 29 Jun 2011 20:53:54 -0700, Keith Packard kei...@keithp.com wrote: On Fri, 24 Jun 2011 17:03:22 +0900, KOSAKI Motohiro

Re: [PATCH] i915: slab shrinker have to return -1 if it cant shrink any objects

(2011/07/13 16:41), Chris Wilson wrote: On Wed, 13 Jul 2011 09:19:24 +0900, KOSAKI Motohiro kosaki.motoh...@jp.fujitsu.com wrote: (2011/07/12 19:06), Chris Wilson wrote: On Tue, 12 Jul 2011 18:36:50 +0900, KOSAKI Motohiro kosaki.motoh...@jp.fujitsu.com wrote: Hi, sorry for the delay.

Re: [PATCH 1/6] drm/radeon: Remove a bunch of useless _iomem casts

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: Just defining rdev-rmmio properly in the first place should do the trick. In some cases, the cast were also complete dups as the original variable was already of the right type. Signed-off-by: Benjamin

Re: [PATCH 2/6] drm/radeon: ATOM Endian fix for atombios_crtc_program_pll()

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: v6 of the structure was programmed incorrectly:  args.v6.ulCrtcPclkFreq.ulPixelClock = cpu_to_le32(clock / 10); ulPixelClock is a 24-bit bitfield. This statement would thus do a 32-bit swap of (clock /

[Bug 39193] New: glCheckFramebufferStatusEXT segfaults in Gallium when checking status on a framebuffer bound to a texture that's bound to a pixmap

https://bugs.freedesktop.org/show_bug.cgi?id=39193 Summary: glCheckFramebufferStatusEXT segfaults in Gallium when checking status on a framebuffer bound to a texture that's bound to a pixmap Product: Mesa Version:

Re: [PATCH 1/2] radeon: Don't generate new fence for page flip.

2011/7/12 Michel Dänzer mic...@daenzer.net: From: Michel Dänzer michel.daen...@amd.com Use the fence of the new frontbuffer, if any. What tree is this against, I can't apply it to drm-fixes or drm-core-next here. Dave. ___ dri-devel mailing list

Re: [PATCH 3/6] drm/radeon: Writeback endian fixes

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: The writeback ring pointer and IH ring pointer are read using le32_to_cpu so we do not want the chip to byteswap them on big-endian. We still want to byteswap the ring itself and the IBs, so we don't touch

Re: [PATCH 4/6] drm/radeon: Add a rmb() in IH processing

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: We should have a read memory barrier between reading the WPTR from memory and reading ring entries based on that value (ie, we need to ensure both loads are done in order by the CPU). It could be argued

Re: [PATCH 5/6] drm/radeon: Do an MMIO read on interrupts when not uisng MSIs

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: When not using MSIs, there is no guarantee that DMA from the device has been fully flushed to point where it's visible to the CPU when taking an interrupt. To get this guarantee, we need to perform an MMIO

Re: [PATCH 6/6] drm/radeon: Fix the definition of RADEON_BUF_SWAP_32BIT

On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: (Note that this is duplicated under various other names such as R600_BUF_SWAP_32BIT etc...). At least now all the definitions agree. Signed-off-by: Benjamin Herrenschmidt b...@kernel.crashing.org

Re: [PATCH 4/6] drm/radeon: Add a rmb() in IH processing

On Wed, Jul 13, 2011 at 10:43 AM, Alex Deucher alexdeuc...@gmail.com wrote: On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: We should have a read memory barrier between reading the WPTR from memory and reading ring entries based on that value (ie, we

Re: [PATCH 5/6] drm/radeon: Do an MMIO read on interrupts when not uisng MSIs

On Wed, Jul 13, 2011 at 10:44 AM, Alex Deucher alexdeuc...@gmail.com wrote: On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: When not using MSIs, there is no guarantee that DMA from the device has been fully flushed to point where it's visible to the CPU

[PATCH 6/6] drm/radeon: Fix the definition of RADEON_BUF_SWAP_32BIT

(Note that this is duplicated under various other names such as R600_BUF_SWAP_32BIT etc...). At least now all the definitions agree. Signed-off-by: Benjamin Herrenschmidt b...@kernel.crashing.org --- (resent adding dri-devel to the CC list to hit patchwork) drivers/gpu/drm/radeon/radeon_reg.h

[PATCH 4/6] drm/radeon: Add a rmb() in IH processing

We should have a read memory barrier between reading the WPTR from memory and reading ring entries based on that value (ie, we need to ensure both loads are done in order by the CPU). It could be argued that the MMIO reads in r600_ack_irq() might be enough to get that barrier but I prefer keeping

[PATCH 5/6] drm/radeon: Do an MMIO read on interrupts when not uisng MSIs

When not using MSIs, there is no guarantee that DMA from the device has been fully flushed to point where it's visible to the CPU when taking an interrupt. To get this guarantee, we need to perform an MMIO read from the device, which will flush all outstanding DMAs from bridges between the device

[PATCH 1/6] drm/radeon: Remove a bunch of useless _iomem casts

Just defining rdev-rmmio properly in the first place should do the trick. In some cases, the cast were also complete dups as the original variable was already of the right type. Signed-off-by: Benjamin Herrenschmidt b...@kernel.crashing.org --- (resent adding dri-devel to the CC list to hit

[PATCH 3/6] drm/radeon: Writeback endian fixes

The writeback ring pointer and IH ring pointer are read using le32_to_cpu so we do not want the chip to byteswap them on big-endian. We still want to byteswap the ring itself and the IBs, so we don't touch that but we remove setting of the byteswap bits in CP_RB_RPTR_ADDR and IH_CNTL. In

[PATCH 2/2] radeon: Don't clobber error return value in page flipping cleanup paths.

From: Michel Dänzer michel.daen...@amd.com Signed-off-by: Michel Dänzer michel.daen...@amd.com Reviewed-by: Alex Deucher alexander.deuc...@amd.com --- drivers/gpu/drm/radeon/radeon_display.c |9 ++--- 1 files changed, 2 insertions(+), 7 deletions(-) diff --git

[PATCH v2 1/2] radeon: Don't generate new fence for page flip.

From: Michel Dänzer michel.daen...@amd.com Use the fence of the new frontbuffer, if any. Generating a new fence could cause us to wait for completely unrelated rendering to finish before performing the flip. Signed-off-by: Michel Dänzer michel.daen...@amd.com --- v2: * Rebase against

Re: [PULL] drm-intel-next

On Wed, Jul 13, 2011 at 09:39:41AM -0700, Keith Packard wrote: Here's most of the patches I'm hoping to land after 3.0: Is this one intentionally not in or did it slip through? drm/i915: gracefully bail out when init_clock_gating-pointer is not set [1] [1]

[Bug 39202] New: KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 Summary: KDE desktop effects with 3.0 rc6 kernel Product: Mesa Version: git Platform: All OS/Version: Linux (All) Status: NEW Severity: normal Priority: medium

Re: [PULL] drm-intel-next

On Wed, 13 Jul 2011 19:22:14 +0200, Wolfram Sang w.s...@pengutronix.de wrote: Is this one intentionally not in or did it slip through? I thought I had replied to that -- it doesn't apply to either -fixes or -next at this point. I can try to fix it, but I'd prefer it if you'd figure out how

[Bug 39202] KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 --- Comment #1 from Siganderson dj_...@webmail.it 2011-07-13 12:25:57 PDT --- video card: ATI Radeon HD 5670 512 MB Ram -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because:

Re: [PATCH] drm/i915/bios: Avoid temporary allocation whilst searching for downclock

On Tue, 5 Jul 2011 09:55:34 +0100, Chris Wilson ch...@chris-wilson.co.uk wrote: Alan Cox reported a missing check on the kmalloc return value for the allocation of a temporary mode used for searching for the LVDS downlock frequency. This allocation is roughly 200 bytes, a little too large to

Re: [PULL] drm-intel-next

Sorry for not pinging you when I didn't hear back. No problem! Will fix it. Thanks, Wolfram -- Pengutronix e.K. | Wolfram Sang| Industrial Linux Solutions | http://www.pengutronix.de/ | signature.asc Description: Digital

[Bug 35988] Power management not working on ATI M92 LP [Mobility Radeon HD 4300]

https://bugs.freedesktop.org/show_bug.cgi?id=35988 --- Comment #9 from leonardo gene...@bsod.eu 2011-07-13 13:09:19 PDT --- I have the same card on DELL Studio 15. I wouldn't say that power management does not work. I'm using profile method with low setting. After about ~20 mins of uptime this is

[PATCH] drm/i915: Add quirk to disable SSC on Lenovo U160 LVDS

We've tried several times to make this machine 'just work', but every patch that does causes many other machines to fail. This adds a quirk which special cases this hardware and forces ssc to be disabled. There's no way to override this from the command line; that would be a significantly more

[PULL] drm-intel-next

Here's most of the patches I'm hoping to land after 3.0: * FBC cleanups from Chris Wilson. Fixes 'missing' CPU writes to the front buffer. We've enabled FBC by default, if we find regressions again, we'll turn it off before the release. * DP and HDMI support for formats other than 8bpc

[Bug 39202] KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 --- Comment #2 from Alex Deucher ag...@yahoo.com 2011-07-13 13:22:55 PDT --- Please attach your xorg log and config and dmesg output. -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail

[PATCH 01/10] intel: shared header for shader debugging

Signed-off-by: Ben Widawsky b...@bwidawsk.net --- intel/Makefile.am |3 ++- intel/intel_debug.h | 44 2 files changed, 46 insertions(+), 1 deletions(-) diff --git a/intel/Makefile.am b/intel/Makefile.am index b6a9014..7a44aaf 100644 ---

[PATCH 0/10] fs shader debugging

Version 2 of the patch series is pretty much the same as version 1. 2 of the patches have already been picked up by the kernel and mesa so they are gone. The only major change is in mesa where I no longer load a binary blob from an environment variable, but instead compile the bytes directly in

Re: [PATCH 01/10] intel: shared header for shader debugging

On Wed, 13 Jul 2011 13:51:43 -0700, Ben Widawsky b...@bwidawsk.net wrote: Signed-off-by: Ben Widawsky b...@bwidawsk.net --- intel/Makefile.am |3 ++- intel/intel_debug.h | 44 2 files changed, 46 insertions(+), 1 deletions(-) diff

Re: [PATCH 2/6] drm/radeon: ATOM Endian fix for atombios_crtc_program_pll()

On Wed, 2011-07-13 at 10:38 -0400, Alex Deucher wrote: case 6: - args.v6.ulCrtcPclkFreq.ucCRTC = crtc_id; - args.v6.ulCrtcPclkFreq.ulPixelClock = cpu_to_le32(clock / 10); + args.v6.ulDispEngClkFreq =

Re: [PATCH 3/6] drm/radeon: Writeback endian fixes

On Wed, 2011-07-13 at 10:42 -0400, Alex Deucher wrote: On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: The writeback ring pointer and IH ring pointer are read using le32_to_cpu so we do not want the chip to byteswap them on big-endian. We still

Re: [PATCH 4/6] drm/radeon: Add a rmb() in IH processing

On Wed, 2011-07-13 at 10:48 -0400, Alex Deucher wrote: On Wed, Jul 13, 2011 at 10:43 AM, Alex Deucher alexdeuc...@gmail.com wrote: On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: We should have a read memory barrier between reading the WPTR from

Re: [PATCH 0/10] fs shader debugging

On Wed, Jul 13, 2011 at 01:51:42PM -0700, Ben Widawsky wrote: Version 2 of the patch series is pretty much the same as version 1. 2 of the patches have already been picked up by the kernel and mesa so they are gone. The only major change is in mesa where I no longer load a binary blob

[Bug 39202] KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 --- Comment #3 from Siganderson dj_...@webmail.it 2011-07-13 15:26:34 PDT --- Created an attachment (id=49061) -- (https://bugs.freedesktop.org/attachment.cgi?id=49061) xorg log -- Configure bugmail:

[Bug 39202] KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 --- Comment #4 from Siganderson dj_...@webmail.it 2011-07-13 15:26:57 PDT --- Created an attachment (id=49062) -- (https://bugs.freedesktop.org/attachment.cgi?id=49062) xorg conf -- Configure bugmail:

[Bug 39202] KDE desktop effects with 3.0 rc6 kernel

https://bugs.freedesktop.org/show_bug.cgi?id=39202 --- Comment #5 from Siganderson dj_...@webmail.it 2011-07-13 15:27:24 PDT --- Created an attachment (id=49063) -- (https://bugs.freedesktop.org/attachment.cgi?id=49063) dmesg -- Configure bugmail:

Re: [PATCH v2 1/2] radeon: Don't generate new fence for page flip.

2011/7/13 Michel Dänzer mic...@daenzer.net: From: Michel Dänzer michel.daen...@amd.com Use the fence of the new frontbuffer, if any. Generating a new fence could cause us to wait for completely unrelated rendering to finish before performing the flip. Signed-off-by: Michel Dänzer

Re: [PATCH 2/6] drm/radeon: ATOM Endian fix for atombios_crtc_program_pll()

On Wed, Jul 13, 2011 at 5:40 PM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: On Wed, 2011-07-13 at 10:38 -0400, Alex Deucher wrote:                case 6: -                       args.v6.ulCrtcPclkFreq.ucCRTC = crtc_id; -                       args.v6.ulCrtcPclkFreq.ulPixelClock

Re: [PATCH 3/6] drm/radeon: Writeback endian fixes

On Wed, Jul 13, 2011 at 5:41 PM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: On Wed, 2011-07-13 at 10:42 -0400, Alex Deucher wrote: On Wed, Jul 13, 2011 at 2:28 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: The writeback ring pointer and IH ring pointer are read using

[PATCH] drm: platform multi-device support

Include the device id in the bus-id to give userspace a way to open the correct cardN when there are multiple device instances. Signed-off-by: Rob Clark r...@ti.com --- drivers/gpu/drm/drm_platform.c |5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) diff --git