similar to: [patch] pin/unpin must flush tlb

Index: linux-2.6.13/arch/i386/kernel/ldt.c =================================================================== --- linux-2.6.13.orig/arch/i386/kernel/ldt.c 2005-08-08 15:27:26.000000000 -0700 +++ linux-2.6.13/arch/i386/kernel/ldt.c 2005-08-08 21:13:19.000000000 -0700 @@ -18,6 +18,7 @@ #include <asm/system.h> #include <asm/ldt.h> #include <asm/desc.h> +#include

Index: linux-2.6.13/arch/i386/kernel/ldt.c =================================================================== --- linux-2.6.13.orig/arch/i386/kernel/ldt.c 2005-08-08 15:27:26.000000000 -0700 +++ linux-2.6.13/arch/i386/kernel/ldt.c 2005-08-08 21:13:19.000000000 -0700 @@ -18,6 +18,7 @@ #include <asm/system.h> #include <asm/ldt.h> #include <asm/desc.h> +#include

Add Xen accessors for LDT updates. Todo: Bring back return values from write_ldt_entry. Signed-off-by: Zachary Amsden <zach@vmware.com> Index: linux-2.6.13/arch/i386/kernel/ldt.c =================================================================== --- linux-2.6.13.orig/arch/i386/kernel/ldt.c 2005-08-08 21:13:56.000000000 -0700 +++ linux-2.6.13/arch/i386/kernel/ldt.c 2005-08-08

Add Xen accessors for LDT updates. Todo: Bring back return values from write_ldt_entry. Signed-off-by: Zachary Amsden <zach@vmware.com> Index: linux-2.6.13/arch/i386/kernel/ldt.c =================================================================== --- linux-2.6.13.orig/arch/i386/kernel/ldt.c 2005-08-08 21:13:56.000000000 -0700 +++ linux-2.6.13/arch/i386/kernel/ldt.c 2005-08-08

The new interfaces drm_gem_vram_{pin/unpin}_reserved() are variants of the GEM VRAM pin/unpin functions that do not reserve the BO during validation. The mgag200 driver requires this behavior for its cursor handling. The patch also converts the driver to use the new interfaces. Signed-off-by: Thomas Zimmermann <tzimmermann at suse.de> --- drivers/gpu/drm/drm_gem_vram_helper.c | 75

On Mon, May 20, 2019 at 06:19:00PM +0200, Daniel Vetter wrote: > On Thu, May 16, 2019 at 06:27:45PM +0200, Thomas Zimmermann wrote: > > The new interfaces drm_gem_vram_{pin/unpin}_reserved() are variants of the > > GEM VRAM pin/unpin functions that do not reserve the BO during validation. > > The mgag200 driver requires this behavior for its cursor handling. The > >

Hi, > It turns out that the bochs and vbox drivers automatically reserved and > unreserved the BO from within their pin and unpin functions. The other > drivers; ast, hibmc and mgag200; performed reservation explicitly. With the > GEM VRAM conversion, automatic BO reservation within pin and unpin functions > accidentally got lost. So for bochs and vbox, ttm_bo_validate() worked on

Am 20.05.19 um 18:26 schrieb Daniel Vetter: > [CAUTION: External Email] > > On Mon, May 20, 2019 at 06:19:00PM +0200, Daniel Vetter wrote: >> On Thu, May 16, 2019 at 06:27:45PM +0200, Thomas Zimmermann wrote: >>> The new interfaces drm_gem_vram_{pin/unpin}_reserved() are variants of the >>> GEM VRAM pin/unpin functions that do not reserve the BO during validation.

The patched attached enables x86_64 xenlinux with "late pin, early unpin", which is already implemented for x86_32. Since we now only pin the root rather than any of the other levels, the overall performance became better especially with workloads that require heavy memory management operations. On 8-way x86_64 xenlinux (dom0) the kernel build was improved by about 10% (using make

virtio-gpu objects never move around, so effectively they are pinned all the time. This makes the the implementation pretty easy ;) Signed-off-by: Gerd Hoffmann <kraxel at redhat.com> --- drivers/gpu/drm/virtio/virtgpu_prime.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/virtio/virtgpu_prime.c b/drivers/gpu/drm/virtio/virtgpu_prime.c index

Hi, > I think would be good to have a lockdep_assert_held here for the ww_mutex. > > Also general thing: _reserved is kinda ttm lingo, for dma-buf reservations > we call the structure tracking the fences+lock the "reservation", but the > naming scheme used is _lock/_unlock. > > I think would be good to be consistent with that, and use _locked here. > Especially

[ This is ancient code. - dan ] fs/fs_context.c 168 { 169 int ret; 170 171 struct fs_parameter param = { 172 .key = key, 173 .type = fs_value_is_flag, 174 .size = v_size, 175 }; 176 177 if (value) { --> 178 param.string =

A kernel test bot reported a problem with the locktorture testcase that was triggered by the GEM VRAM helpers. ... [ 10.004734] RIP: 0010:ttm_bo_validate+0x41/0x141 [ttm] ... [ 10.015669] ? kvm_sched_clock_read+0x5/0xd [ 10.016157] ? get_lock_stats+0x11/0x3f [ 10.016607] drm_gem_vram_pin+0x77/0xa2 [drm_vram_helper] [ 10.017229]

A kernel test bot reported a problem with the locktorture testcase that was triggered by the GEM VRAM helpers. ... [ 10.004734] RIP: 0010:ttm_bo_validate+0x41/0x141 [ttm] ... [ 10.015669] ? kvm_sched_clock_read+0x5/0xd [ 10.016157] ? get_lock_stats+0x11/0x3f [ 10.016607] drm_gem_vram_pin+0x77/0xa2 [drm_vram_helper] [ 10.017229]

On Thu, May 16, 2019 at 06:27:45PM +0200, Thomas Zimmermann wrote: > The new interfaces drm_gem_vram_{pin/unpin}_reserved() are variants of the > GEM VRAM pin/unpin functions that do not reserve the BO during validation. > The mgag200 driver requires this behavior for its cursor handling. The > patch also converts the driver to use the new interfaces. > > Signed-off-by: Thomas

On Thu, May 16, 2019 at 06:27:45PM +0200, Thomas Zimmermann wrote: > The new interfaces drm_gem_vram_{pin/unpin}_reserved() are variants of the > GEM VRAM pin/unpin functions that do not reserve the BO during validation. > The mgag200 driver requires this behavior for its cursor handling. The > patch also converts the driver to use the new interfaces. > > Signed-off-by: Thomas

This will otherwise cause a lockdep splat if reservations were a real lock type, so warn when nouveau forgets to unpin a buffer, and fix up the ones I've hit. Signed-off-by: Maarten Lankhorst <maarten.lankhorst at canonical.com> --- diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c index cc79c79..acc6b08 100644 ---

This will otherwise cause a lockdep splat, so warn when nouveau forgets to unpin a buffer, and fix up the ones I've hit. Signed-off-by: Maarten Lankhorst <maarten.lankhorst at canonical.com> --- drivers/gpu/drm/nouveau/nouveau_abi16.c | 1 + drivers/gpu/drm/nouveau/nouveau_fbcon.c | 1 + drivers/gpu/drm/nouveau/nouveau_gem.c | 3 ++- drivers/gpu/drm/nouveau/nouveau_prime.c |

Fixes GART leak (as accounted by nouveau_drm.gem.gart_available). Signed-off-by: Marcin Slusarz <marcin.slusarz at gmail.com> --- Running glxinfo in a loop is enough to trigger it - after several thousand runs (depending on GART size), X server crashes and does not come back in a correct state (corruptions and crashes). With this patch applied, it's possible again to do full piglit

These objects leak VRAM - but only on module unload. Signed-off-by: Marcin Slusarz <marcin.slusarz at gmail.com> --- drivers/gpu/drm/nouveau/nv04_crtc.c | 6 +++++- drivers/gpu/drm/nouveau/nv10_fence.c | 7 ++++++- drivers/gpu/drm/nouveau/nv50_display.c | 21 ++++++++++++++++++--- drivers/gpu/drm/nouveau/nv50_fence.c | 5 ++++- drivers/gpu/drm/nouveau/nvc0_fence.c | 7 ++++++-