On Wed, May 02, 2018 at 05:46:17AM -0700, Christoph Hellwig wrote:
> Any more comments? Especially from the x86, mips and powerpc arch
> maintainers? I'd like to merge this in a few days as various other
> patches depend on it.
I've pulled it in to make forward progress. Any additional
On Tue, 2018-05-08 at 15:07 -0500, Gary R Hook wrote:
> On 05/08/2018 01:48 PM, Joe Perches wrote:
> > On Tue, 2018-05-08 at 12:08 -0500, Hook, Gary wrote:
> > > On 5/7/2018 6:47 PM, kbuild test robot wrote:
> > > >
> > > > All error/warnings (new ones prefixed by >>):
> > > >
> > > > In
On 05/08/2018 03:42 PM, Joe Perches wrote:
On Tue, 2018-05-08 at 15:07 -0500, Gary R Hook wrote:
On 05/08/2018 01:48 PM, Joe Perches wrote:
On Tue, 2018-05-08 at 12:08 -0500, Hook, Gary wrote:
On 5/7/2018 6:47 PM, kbuild test robot wrote:
All error/warnings (new ones prefixed by >>):
On 05/08/2018 01:48 PM, Joe Perches wrote:
On Tue, 2018-05-08 at 12:08 -0500, Hook, Gary wrote:
On 5/7/2018 6:47 PM, kbuild test robot wrote:
All error/warnings (new ones prefixed by >>):
In file included from include/linux/intel-iommu.h:32:0,
from
/github.com/0day-ci/linux/commits/Gary-R-Hook/iommu-Enable-debugfs-exposure-of-IOMMU-driver-internals/20180508-062918
> > base: https://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu.git next
> > config: x86_64-randconfig-x016-201818 (attached as .config)
> > compiler: gcc
Currently GART writes one page entry at a time. More optimal would be to
aggregate the writes and flush BUS buffer in the end, this gives map/unmap
10-40% (depending on size of mapping) performance boost compared to a
flushing after each entry update.
Signed-off-by: Dmitry Osipenko
Introduce iotlb_sync_map() callback that is invoked in the end of
iommu_map(). This new callback allows IOMMU drivers to avoid syncing
on mapping of each contiguous chunk and sync only when whole mapping
is completed, optimizing performance of the mapping operation.
Signed-off-by: Dmitry Osipenko
GART aperture is shared by all devices, hence there is a single IOMMU
domain and group shared by these devices. Allocation of a group per
device only wastes resources and allowance of having more than one domain
is simply wrong because IOMMU mappings made by the users of "different"
domains will
GART can't handle all devices, ignore devices that aren't related to GART.
Device tree must explicitly assign GART IOMMU to the devices.
Signed-off-by: Dmitry Osipenko
---
drivers/iommu/tegra-gart.c | 33 -
1 file changed, 32 insertions(+), 1
Properly clean up allocated resources on driver probe failure.
Signed-off-by: Dmitry Osipenko
---
drivers/iommu/tegra-gart.c | 13 ++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/drivers/iommu/tegra-gart.c b/drivers/iommu/tegra-gart.c
index
GART driver is built-in, hence it can't be unloaded. This patch merely
removes the dead code.
Signed-off-by: Dmitry Osipenko
---
drivers/iommu/tegra-gart.c | 25 +++--
1 file changed, 3 insertions(+), 22 deletions(-)
diff --git a/drivers/iommu/tegra-gart.c
In order to report clients name and access direction on GART page fault,
MC driver needs to access GART registers. Add facility that provides
access to the GART.
Signed-off-by: Dmitry Osipenko
---
drivers/memory/tegra/mc.c | 26 +++---
Hello,
This series addresses multiple shortcomings of the GART driver:
1. Thierry noticed that Memory Controller driver uses registers that belong
to GART in [0] and for now MC driver only reports the fact of GART's page
fault. The first two patches of the series are addressing this
Remove unneeded 'includes' and sort them in alphabet order. Also remove
pr_fmt since there is no pr_xxx() and it doesn't affect dev_xxx().
Signed-off-by: Dmitry Osipenko
---
drivers/iommu/tegra-gart.c | 17 +
1 file changed, 5 insertions(+), 12 deletions(-)
GART contains registers needed by the Memory Controller driver. Provide
access to the MC driver by utilizing its GART-integration facility.
Signed-off-by: Dmitry Osipenko
---
drivers/iommu/tegra-gart.c | 23 +++
1 file changed, 23 insertions(+)
diff --git
]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system]
url:
https://github.com/0day-ci/linux/commits/Gary-R-Hook/iommu-Enable-debugfs-exposure-of-IOMMU-driver-internals/20180508-062918
base: https://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu.git
On Tue, May 08, 2018 at 11:05:20AM +0100, Robin Murphy wrote:
> On 24/04/18 15:02, Christoph Hellwig wrote:
>> Only used by the AMD GART driver, which must be built in.
>
> FWIW debug_dma_dump_mappings() is also called by the Intel VT-d driver, but
> the same reasoning still applies.
I'll update
Hi Jacob,
Looks mostly good to me, I just have a couple more comments
On 04/05/18 19:07, Jacob Pan wrote:
> Now the passdown invalidation granularities look like:
> (sorted by coarseness), will send out in v5 patchset soon if no issues.
>
> /**
> * enum iommu_inv_granularity - Generic
Hi Christoph,
On 2018-04-24 16:02, Christoph Hellwig wrote:
> Most mainstream architectures are using 65536 entries, so lets stick to
> that. If someone is really desperate to override it that can still be
> done through , but I'd rather see a really good
> rationale for that.
>
> dma_debug_init
On 25/04/18 06:58, Christoph Hellwig wrote:
This looks interesting. I suspect it is going to blow up in
quite a few places, so maybe at least for now it might make sense
to have a separate config option?
True, it's nice to verify this for 'traditional' dma_map_sg() usage, but
places where
On 24/04/18 15:02, Christoph Hellwig wrote:
Only used by the AMD GART driver, which must be built in.
FWIW debug_dma_dump_mappings() is also called by the Intel VT-d driver,
but the same reasoning still applies. This does rather beg the question
of whether it's right to have bits of
On 27/04/18 16:53, Christoph Hellwig wrote:
There is no arch specific code required for dma-debug, so there is no
need to opt into the support either.
Makes sense, and a purely negative diffstat is always pleasing :)
Reviewed-by: Robin Murphy
Signed-off-by: Christoph
On 24/04/18 15:02, Christoph Hellwig wrote:
Just keep a single variable with a descriptive name instead of two
with confusing names.
Reviewed-by: Robin Murphy
Signed-off-by: Christoph Hellwig
---
lib/dma-debug.c | 20
1 file
On 24/04/18 15:02, Christoph Hellwig wrote:
Most mainstream architectures are using 65536 entries, so lets stick to
that. If someone is really desperate to override it that can still be
done through , but I'd rather see a really good
rationale for that.
dma_debug_init is now called as a
24 matches
Mail list logo