5.2.0-rc4-next-20190613
dmesg
nouveau :01:00.0: DRM: GPU lockup - switching to software fbcon
nouveau :01:00.0: fifo: SCHED_ERROR 0a [CTXSW_TIMEOUT]
nouveau :01:00.0: fifo: runlist 0: scheduled for recovery
nouveau :01:00.0: fifo: channel 5: killed
nouveau :01:00.0: fifo
On 6/13/19 2:43 AM, Christoph Hellwig wrote:
> The migrate_vma helper is only used by noveau to migrate device private
> pages around. Other HMM_MIRROR users like amdgpu or infiniband don't
> need it.
>
> Signed-off-by: Christoph Hellwig
> ---
> drivers/gpu/drm/nouveau/Kconfig | 1 +
>
On 6/13/19 2:43 AM, Christoph Hellwig wrote:
> noveau is currently using this through an odd hmm wrapper, and I plan
"nouveau"
> to switch it to the real thing later in this series.
>
> Signed-off-by: Christoph Hellwig
> ---
Reviewed-by: John Hubbard
thanks,
--
John Hubbard
NVIDIA
>
On 6/13/19 2:43 AM, Christoph Hellwig wrote:
> ->mapping isn't even used by HMM users, and the field at the same offset
> in the zone_device part of the union is declared as pad. (Which btw is
> rather confusing, as DAX uses ->pgmap and ->mapping from two different
> sides of the union, but DAX
On 6/13/19 5:43 PM, Ira Weiny wrote:
> On Thu, Jun 13, 2019 at 07:58:29PM +, Jason Gunthorpe wrote:
>> On Thu, Jun 13, 2019 at 12:53:02PM -0700, Ralph Campbell wrote:
>>>
...
>> Hum, so the only thing this config does is short circuit here:
>>
>> static inline bool is_device_public_page(const
On 6/13/19 2:43 AM, Christoph Hellwig wrote:
> This function has never been used since it was first added to the kernel
> more than a year and a half ago, and if we ever grow a consumer of the
> MEMORY_DEVICE_PUBLIC infrastructure it can easily use devm_memremap_pages
> directly now that we've
On 6/13/19 5:11 PM, Ralph Campbell wrote:
> In nouveau_dmem_pages_alloc(), the drm->dmem->mutex is unlocked before
> calling nouveau_dmem_chunk_alloc().
> Reacquire the lock before continuing to the next page.
>
> Signed-off-by: Ralph Campbell
> ---
>
> I found this while testing Jason
On Thu, Jun 13, 2019 at 07:58:29PM +, Jason Gunthorpe wrote:
> On Thu, Jun 13, 2019 at 12:53:02PM -0700, Ralph Campbell wrote:
> >
> > On 6/13/19 12:44 PM, Jason Gunthorpe wrote:
> > > On Thu, Jun 13, 2019 at 11:43:21AM +0200, Christoph Hellwig wrote:
> > > > The code hasn't been used since
On Thu, Jun 13, 2019 at 08:40:46PM +, Jason Gunthorpe wrote:
> On Thu, Jun 13, 2019 at 11:27:39AM -0700, Dan Williams wrote:
> > On Thu, Jun 13, 2019 at 2:43 AM Christoph Hellwig wrote:
> > >
> > > Hi Dan, Jérôme and Jason,
> > >
> > > below is a series that cleans up the dev_pagemap
On Thu, Jun 13, 2019 at 05:11:21PM -0700, Ralph Campbell wrote:
> In nouveau_dmem_pages_alloc(), the drm->dmem->mutex is unlocked before
> calling nouveau_dmem_chunk_alloc().
> Reacquire the lock before continuing to the next page.
>
> Signed-off-by: Ralph Campbell
> ---
>
> I found this while
On Thu, Jun 13, 2019 at 11:43:16AM +0200, Christoph Hellwig wrote:
> The functionality is identical to the one currently open coded in
> device-dax.
>
> Signed-off-by: Christoph Hellwig
> ---
> drivers/dax/dax-private.h | 4 ---
> drivers/dax/device.c | 52
In nouveau_dmem_pages_alloc(), the drm->dmem->mutex is unlocked before
calling nouveau_dmem_chunk_alloc().
Reacquire the lock before continuing to the next page.
Signed-off-by: Ralph Campbell
---
I found this while testing Jason Gunthorpe's hmm tree but this is
independant of those changes. I
On 6/13/19 2:43 AM, Christoph Hellwig wrote:
This replaces the hacky ->fault callback, which is currently directly
called from common code through a hmm specific data structure as an
exercise in layering violations.
Signed-off-by: Christoph Hellwig
---
include/linux/hmm.h | 6 --
On Thu, Jun 13, 2019 at 11:21:01PM +0200, Christoph Hellwig wrote:
> On Thu, Jun 13, 2019 at 08:40:46PM +, Jason Gunthorpe wrote:
> > > Perhaps we should pull those out and resend them through hmm.git?
> >
> > It could be done - but how bad is the conflict resolution?
>
> Trivial. All but
On 6/13/19 2:43 AM, Christoph Hellwig wrote:
> This code is a trivial wrapper around device model helpers, which
> should have been integrated into the driver device model usage from
> the start. Assuming it actually had users, which it never had since
> the code was added more than 1 1/2 years
On Thu, Jun 13, 2019 at 08:40:46PM +, Jason Gunthorpe wrote:
> > Perhaps we should pull those out and resend them through hmm.git?
>
> It could be done - but how bad is the conflict resolution?
Trivial. All but one patch just apply using git-am, and the other one
just has a few lines of
On Thu, 13 Jun 2019 14:24:20 -0600 Logan Gunthorpe wrote:
>
>
> On 2019-06-13 2:21 p.m., Dan Williams wrote:
> > On Thu, Jun 13, 2019 at 1:18 PM Logan Gunthorpe wrote:
> >>
> >>
> >>
> >> On 2019-06-13 12:27 p.m., Dan Williams wrote:
> >>> On Thu, Jun 13, 2019 at 2:43 AM Christoph Hellwig
On Thu, Jun 13, 2019 at 11:27:39AM -0700, Dan Williams wrote:
> On Thu, Jun 13, 2019 at 2:43 AM Christoph Hellwig wrote:
> >
> > Hi Dan, Jérôme and Jason,
> >
> > below is a series that cleans up the dev_pagemap interface so that
> > it is more easily usable, which removes the need to wrap it in
On 2019-06-13 2:21 p.m., Dan Williams wrote:
> On Thu, Jun 13, 2019 at 1:18 PM Logan Gunthorpe wrote:
>>
>>
>>
>> On 2019-06-13 12:27 p.m., Dan Williams wrote:
>>> On Thu, Jun 13, 2019 at 2:43 AM Christoph Hellwig wrote:
Hi Dan, Jérôme and Jason,
below is a series that
On Thu, Jun 13, 2019 at 1:18 PM Logan Gunthorpe wrote:
>
>
>
> On 2019-06-13 12:27 p.m., Dan Williams wrote:
> > On Thu, Jun 13, 2019 at 2:43 AM Christoph Hellwig wrote:
> >>
> >> Hi Dan, Jérôme and Jason,
> >>
> >> below is a series that cleans up the dev_pagemap interface so that
> >> it is
On 2019-06-13 12:27 p.m., Dan Williams wrote:
> On Thu, Jun 13, 2019 at 2:43 AM Christoph Hellwig wrote:
>>
>> Hi Dan, Jérôme and Jason,
>>
>> below is a series that cleans up the dev_pagemap interface so that
>> it is more easily usable, which removes the need to wrap it in hmm
>> and thus
On Thu, Jun 13, 2019 at 1:12 PM Logan Gunthorpe wrote:
>
>
>
> On 2019-06-13 3:43 a.m., Christoph Hellwig wrote:
> > Passing the actual typed structure leads to more understandable code
> > vs the actual references.
>
> Ha, ok, I originally suggested this to Dan when he introduced the
>
On 2019-06-13 3:43 a.m., Christoph Hellwig wrote:
> The dev_pagemap is a growing too many callbacks. Move them into a
> separate ops structure so that they are not duplicated for multiple
> instances, and an attacker can't easily overwrite them.
>
> Signed-off-by: Christoph Hellwig
> ---
>
On Thu, Jun 13, 2019 at 12:35 PM Jason Gunthorpe wrote:
>
> On Thu, Jun 13, 2019 at 11:43:12AM +0200, Christoph Hellwig wrote:
> > Just check if there is a ->page_free operation set and take care of the
> > static key enable, as well as the put using device managed resources.
> > diff --git
On 2019-06-13 3:43 a.m., Christoph Hellwig wrote:
> Passing the actual typed structure leads to more understandable code
> vs the actual references.
Ha, ok, I originally suggested this to Dan when he introduced the
callback[1].
Reviewed-by: Logan Gunthorpe
Logan
[1]
On Thu, Jun 13, 2019 at 11:43:25AM +0200, Christoph Hellwig wrote:
> The migrate_vma helper is only used by noveau to migrate device private
> pages around. Other HMM_MIRROR users like amdgpu or infiniband don't
> need it.
>
> Signed-off-by: Christoph Hellwig
> ---
>
[+cc Rafael, Martin, zigarrre]
On Thu, Jun 13, 2019 at 02:35:14PM +0800, Daniel Drake wrote:
> From: Lukas Wunner
>
> The integrated HDA controller on Nvidia GPUs can be hidden with a bit in
> the GPU's config space. Information about this scheme was provided by
> NVIDIA on their forums.
>
>
On Thu, Jun 13, 2019 at 11:43:24AM +0200, Christoph Hellwig wrote:
> All the mm/hmm.c code is better keyed off HMM_MIRROR. Also let nouveau
> depend on it instead of the mix of a dummy dependency symbol plus the
> actually selected one. Drop various odd dependencies, as the code is
> pretty
On Thu, Jun 13, 2019 at 12:53:02PM -0700, Ralph Campbell wrote:
>
> On 6/13/19 12:44 PM, Jason Gunthorpe wrote:
> > On Thu, Jun 13, 2019 at 11:43:21AM +0200, Christoph Hellwig wrote:
> > > The code hasn't been used since it was added to the tree, and doesn't
> > > appear to actually be usable.
On Thu, Jun 13, 2019 at 11:43:23AM +0200, Christoph Hellwig wrote:
> The ZONE_DEVICE support doesn't depend on anything HMM related, just on
> various bits of arch support as indicated by the architecture. Also
> don't select the option from nouveau as it isn't present in many setups,
> and
On 6/13/19 12:44 PM, Jason Gunthorpe wrote:
On Thu, Jun 13, 2019 at 11:43:21AM +0200, Christoph Hellwig wrote:
The code hasn't been used since it was added to the tree, and doesn't
appear to actually be usable. Mark it as BROKEN until either a user
comes along or we finally give up on it.
On Thu, Jun 13, 2019 at 11:43:21AM +0200, Christoph Hellwig wrote:
> The code hasn't been used since it was added to the tree, and doesn't
> appear to actually be usable. Mark it as BROKEN until either a user
> comes along or we finally give up on it.
>
> Signed-off-by: Christoph Hellwig
>
On Thu, Jun 13, 2019 at 9:15 AM Ilia Mirkin wrote:
>
> On Thu, Jun 13, 2019 at 2:35 AM Daniel Drake wrote:
> >
> > From: Lukas Wunner
> >
> > The integrated HDA controller on Nvidia GPUs can be hidden with a bit in
> > the GPU's config space. Information about this scheme was provided by
> >
On Thu, Jun 13, 2019 at 11:43:20AM +0200, Christoph Hellwig wrote:
> There isn't really much value add in the hmm_devmem_add wrapper. Just
> factor out a little helper to find the resource, and otherwise let the
> driver implement the dev_pagemap_ops directly.
Was this commit message written
On Thu, Jun 13, 2019 at 11:43:17AM +0200, Christoph Hellwig wrote:
> hmm_vma_alloc_locked_page is scheduled to go away, use the proper
> mm function directly.
>
> Signed-off-by: Christoph Hellwig
> ---
> drivers/gpu/drm/nouveau/nouveau_dmem.c | 3 ++-
> 1 file changed, 2 insertions(+), 1
On Thu, Jun 13, 2019 at 11:43:14AM +0200, Christoph Hellwig wrote:
> struct dev_pagemap is always embedded into a containing structure, so
> there is no need to an additional private data field.
>
> Signed-off-by: Christoph Hellwig
> ---
> drivers/nvdimm/pmem.c| 2 +-
>
On Thu, Jun 13, 2019 at 11:43:12AM +0200, Christoph Hellwig wrote:
> Just check if there is a ->page_free operation set and take care of the
> static key enable, as well as the put using device managed resources.
> diff --git a/mm/hmm.c b/mm/hmm.c
> index c76a1b5defda..6dc769feb2e1 100644
> +++
On Thu, Jun 13, 2019 at 09:15:31AM -0400, Ilia Mirkin wrote:
> On Thu, Jun 13, 2019 at 2:35 AM Daniel Drake wrote:
> > Link: https://devtalk.nvidia.com/default/topic/1024022
> > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75985
> > Cc: Aaron Plattner
> > Cc: Peter Wu
> > Cc: Ilia
On Thu, Jun 13, 2019 at 11:43:11AM +0200, Christoph Hellwig wrote:
> Passing the actual typed structure leads to more understandable code
> vs the actual references.
>
> Signed-off-by: Christoph Hellwig
> drivers/dax/device.c | 7 +++
> drivers/nvdimm/pmem.c | 6
On Thu, Jun 13, 2019 at 11:43:10AM +0200, Christoph Hellwig wrote:
> The dev_pagemap is a growing too many callbacks. Move them into a
> separate ops structure so that they are not duplicated for multiple
> instances, and an attacker can't easily overwrite them.
Reviewed-by: Jason Gunthorpe
On Thu, Jun 13, 2019 at 11:43:09AM +0200, Christoph Hellwig wrote:
> Keep the physical address allocation that hmm_add_device does with the
> rest of the resource code, and allow future reuse of it without the hmm
> wrapper.
>
> Signed-off-by: Christoph Hellwig
> include/linux/ioport.h | 2 ++
On Thu, Jun 13, 2019 at 11:43:07AM +0200, Christoph Hellwig wrote:
> ->mapping isn't even used by HMM users, and the field at the same offset
> in the zone_device part of the union is declared as pad. (Which btw is
> rather confusing, as DAX uses ->pgmap and ->mapping from two different
> sides
On Thu, Jun 13, 2019 at 11:43:06AM +0200, Christoph Hellwig wrote:
> This function has never been used since it was first added to the kernel
> more than a year and a half ago, and if we ever grow a consumer of the
> MEMORY_DEVICE_PUBLIC infrastructure it can easily use devm_memremap_pages
>
On Thu, Jun 13, 2019 at 11:43:05AM +0200, Christoph Hellwig wrote:
> This code is a trivial wrapper around device model helpers, which
> should have been integrated into the driver device model usage from
> the start. Assuming it actually had users, which it never had since
> the code was added
On Thu, Jun 13, 2019 at 11:43:04AM +0200, Christoph Hellwig wrote:
> Signed-off-by: Christoph Hellwig
> ---
> mm/Kconfig | 10 --
> 1 file changed, 10 deletions(-)
So long as we are willing to run a hmm.git we can merge only complete
features going forward.
Thus we don't need the crazy
On Thu, Jun 13, 2019 at 2:43 AM Christoph Hellwig wrote:
>
> Hi Dan, Jérôme and Jason,
>
> below is a series that cleans up the dev_pagemap interface so that
> it is more easily usable, which removes the need to wrap it in hmm
> and thus allowing to kill a lot of code
>
> Diffstat:
>
> 22 files
As of kernel 5.1.9, on resume from suspend, my NV50 fan runs at full speed.
Not sure if it has to do with this new config option
(NOUVEAU_LEGACY_CTX_SUPPORT)?
This issue is not present using kernel 5.0.21.
Years ago I filed a similar issue:
60704 – [nouveau, git regression] - I2C PWM fan
On Thu, Jun 13, 2019 at 11:43:03AM +0200, Christoph Hellwig wrote:
> Hi Dan, Jérôme and Jason,
>
> below is a series that cleans up the dev_pagemap interface so that
> it is more easily usable, which removes the need to wrap it in hmm
> and thus allowing to kill a lot of code
Do you want some of
This is done with a sequence of blits, from level 0 -> level 1, then
level 1 -> level 2, etc. st/mesa has a helper which will call the
gallium driver's blit method with the appropriate parameters.
Internally, assuming there's nothing too funky going on, we'll use the
2d blit logic (class 0x902d).
Hi guys again. A homebrew developer (homebrew is custom software made for
the switch using openGL under nouveau) reported to me that 'glGenerateMipmap'
wasn't working on yuzu (Nintendo Switch emulator). I looked into it and I
noticed all the triangle data used by nouveau to render the mipmaps was
On Thu, Jun 13, 2019 at 2:35 AM Daniel Drake wrote:
>
> From: Lukas Wunner
>
> The integrated HDA controller on Nvidia GPUs can be hidden with a bit in
> the GPU's config space. Information about this scheme was provided by
> NVIDIA on their forums.
>
> Many laptops now ship with this device
When calling debugfs functions, there is no need to ever check the
return value. The function can work or not, but the code logic should
never do something different based on this.
Cc: Ben Skeggs
Cc: David Airlie
Cc: Daniel Vetter
Cc: dri-de...@lists.freedesktop.org
Cc:
This replaces the hacky ->fault callback, which is currently directly
called from common code through a hmm specific data structure as an
exercise in layering violations.
Signed-off-by: Christoph Hellwig
---
include/linux/hmm.h | 6 --
include/linux/memremap.h | 6 ++
Remove the clumsy hmm_devmem_page_{get,set}_drvdata helpers, and
instead just access the page directly. Also make the page data
a void pointer, and thus much easier to use.
Signed-off-by: Christoph Hellwig
---
drivers/gpu/drm/nouveau/nouveau_dmem.c | 18 +++--
include/linux/hmm.h
Just check if there is a ->page_free operation set and take care of the
static key enable, as well as the put using device managed resources.
Signed-off-by: Christoph Hellwig
---
drivers/nvdimm/pmem.c | 23 +++--
include/linux/mm.h| 10
kernel/memremap.c | 59
The code hasn't been used since it was added to the tree, and doesn't
appear to actually be usable. Mark it as BROKEN until either a user
comes along or we finally give up on it.
Signed-off-by: Christoph Hellwig
---
mm/Kconfig | 1 +
1 file changed, 1 insertion(+)
diff --git a/mm/Kconfig
There isn't really much value add in the hmm_devmem_add wrapper. Just
factor out a little helper to find the resource, and otherwise let the
driver implement the dev_pagemap_ops directly.
Signed-off-by: Christoph Hellwig
---
Documentation/vm/hmm.rst | 26
include/linux/hmm.h |
The migrate_vma helper is only used by noveau to migrate device private
pages around. Other HMM_MIRROR users like amdgpu or infiniband don't
need it.
Signed-off-by: Christoph Hellwig
---
drivers/gpu/drm/nouveau/Kconfig | 1 +
mm/Kconfig | 1 -
2 files changed, 1
All the mm/hmm.c code is better keyed off HMM_MIRROR. Also let nouveau
depend on it instead of the mix of a dummy dependency symbol plus the
actually selected one. Drop various odd dependencies, as the code is
pretty portable.
Signed-off-by: Christoph Hellwig
---
struct dev_pagemap is always embedded into a containing structure, so
there is no need to an additional private data field.
Signed-off-by: Christoph Hellwig
---
drivers/nvdimm/pmem.c| 2 +-
include/linux/memremap.h | 3 +--
kernel/memremap.c| 2 +-
mm/hmm.c | 9
The ZONE_DEVICE support doesn't depend on anything HMM related, just on
various bits of arch support as indicated by the architecture. Also
don't select the option from nouveau as it isn't present in many setups,
and depend on it instead.
Signed-off-by: Christoph Hellwig
---
Just use devm_memremap_pages instead of hmm_devmem_add pages to allow
killing that wrapper which doesn't provide a whole lot of benefits.
Signed-off-by: Christoph Hellwig
---
drivers/gpu/drm/nouveau/nouveau_dmem.c | 80 --
1 file changed, 38 insertions(+), 42
The functionality is identical to the one currently open coded in
device-dax.
Signed-off-by: Christoph Hellwig
---
drivers/dax/dax-private.h | 4 ---
drivers/dax/device.c | 52 +--
2 files changed, 1 insertion(+), 55 deletions(-)
diff --git
The only user of it has just been removed, and there wasn't really any need
to wrap a basic memory allocator to start with.
Signed-off-by: Christoph Hellwig
---
include/linux/hmm.h | 3 ---
mm/hmm.c| 14 --
2 files changed, 17 deletions(-)
diff --git
hmm_vma_alloc_locked_page is scheduled to go away, use the proper
mm function directly.
Signed-off-by: Christoph Hellwig
---
drivers/gpu/drm/nouveau/nouveau_dmem.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c
The dev_pagemap is a growing too many callbacks. Move them into a
separate ops structure so that they are not duplicated for multiple
instances, and an attacker can't easily overwrite them.
Signed-off-by: Christoph Hellwig
---
drivers/dax/device.c | 6 +-
Provide an internal refcounting logic if no ->ref field is provided
in the pagemap passed into devm_memremap_pages so that callers don't
have to reinvent it poorly.
Signed-off-by: Christoph Hellwig
---
include/linux/memremap.h | 4 +++
kernel/memremap.c | 60
This function has never been used since it was first added to the kernel
more than a year and a half ago, and if we ever grow a consumer of the
MEMORY_DEVICE_PUBLIC infrastructure it can easily use devm_memremap_pages
directly now that we've simplified the API for it.
Signed-off-by: Christoph
->mapping isn't even used by HMM users, and the field at the same offset
in the zone_device part of the union is declared as pad. (Which btw is
rather confusing, as DAX uses ->pgmap and ->mapping from two different
sides of the union, but DAX doesn't use hmm_devmem_free).
Signed-off-by:
noveau is currently using this through an odd hmm wrapper, and I plan
to switch it to the real thing later in this series.
Signed-off-by: Christoph Hellwig
---
mm/mempolicy.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index 01600d80ae01..f9023b5fba37
Keep the physical address allocation that hmm_add_device does with the
rest of the resource code, and allow future reuse of it without the hmm
wrapper.
Signed-off-by: Christoph Hellwig
---
include/linux/ioport.h | 2 ++
kernel/resource.c | 39 +++
Passing the actual typed structure leads to more understandable code
vs the actual references.
Signed-off-by: Christoph Hellwig
---
drivers/dax/device.c | 7 +++
drivers/nvdimm/pmem.c | 6 +++---
drivers/pci/p2pdma.c | 6 +++---
include/linux/memremap.h
This code is a trivial wrapper around device model helpers, which
should have been integrated into the driver device model usage from
the start. Assuming it actually had users, which it never had since
the code was added more than 1 1/2 years ago.
Signed-off-by: Christoph Hellwig
---
Signed-off-by: Christoph Hellwig
---
mm/Kconfig | 10 --
1 file changed, 10 deletions(-)
diff --git a/mm/Kconfig b/mm/Kconfig
index f0c76ba47695..0d2ba7e1f43e 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -675,16 +675,6 @@ config ARCH_HAS_HMM_MIRROR
depends on (X86_64 || PPC64)
Hi Dan, Jérôme and Jason,
below is a series that cleans up the dev_pagemap interface so that
it is more easily usable, which removes the need to wrap it in hmm
and thus allowing to kill a lot of code
Diffstat:
22 files changed, 245 insertions(+), 802 deletions(-)
Git tree:
From: Lukas Wunner
The integrated HDA controller on Nvidia GPUs can be hidden with a bit in
the GPU's config space. Information about this scheme was provided by
NVIDIA on their forums.
Many laptops now ship with this device hidden, meaning that Linux users
of affected platforms (where the HDMI
76 matches
Mail list logo
