Re: [PATCH 09/10] arch: introduce memremap()

2015-07-22 Thread Dan Williams
On Wed, Jul 22, 2015 at 5:55 PM, Luis R. Rodriguez wrote: > On Wed, Jul 22, 2015 at 04:15:41PM -0700, Dan Williams wrote: > Anyway so why is WB and WT allowed if you are going to remove WC and not add > UC? > Because all existing usages of WB and WT mappings are clearly free of I/O side effect c

Re: [PATCH 09/10] arch: introduce memremap()

2015-07-22 Thread Luis R. Rodriguez
On Wed, Jul 22, 2015 at 04:15:41PM -0700, Dan Williams wrote: > On Wed, Jul 22, 2015 at 3:55 PM, Luis R. Rodriguez wrote: > > On Tue, Jul 21, 2015 at 09:04:22PM -0700, Dan Williams wrote: > >> On Tue, Jul 21, 2015 at 4:58 PM, Luis R. Rodriguez wrote: > >> > On Sun, Jul 19, 2015 at 08:18:23PM -040

Re: [PATCH 09/10] arch: introduce memremap()

2015-07-22 Thread Dan Williams
On Wed, Jul 22, 2015 at 3:55 PM, Luis R. Rodriguez wrote: > On Tue, Jul 21, 2015 at 09:04:22PM -0700, Dan Williams wrote: >> On Tue, Jul 21, 2015 at 4:58 PM, Luis R. Rodriguez wrote: >> > On Sun, Jul 19, 2015 at 08:18:23PM -0400, Dan Williams wrote: >> >> diff --git a/include/linux/io.h b/include

Re: [PATCH 09/10] arch: introduce memremap()

2015-07-22 Thread Luis R. Rodriguez
On Tue, Jul 21, 2015 at 09:04:22PM -0700, Dan Williams wrote: > On Tue, Jul 21, 2015 at 4:58 PM, Luis R. Rodriguez wrote: > > On Sun, Jul 19, 2015 at 08:18:23PM -0400, Dan Williams wrote: > >> diff --git a/include/linux/io.h b/include/linux/io.h > >> index 080a4fbf2ba4..2983b6e63970 100644 > >> --

Re: [PATCH 09/10] arch: introduce memremap()

2015-07-21 Thread Dan Williams
On Tue, Jul 21, 2015 at 4:58 PM, Luis R. Rodriguez wrote: > On Sun, Jul 19, 2015 at 08:18:23PM -0400, Dan Williams wrote: >> diff --git a/include/linux/io.h b/include/linux/io.h >> index 080a4fbf2ba4..2983b6e63970 100644 >> --- a/include/linux/io.h >> +++ b/include/linux/io.h >> @@ -192,4 +192,15

Re: [PATCH 09/10] arch: introduce memremap()

2015-07-21 Thread Luis R. Rodriguez
On Sun, Jul 19, 2015 at 08:18:23PM -0400, Dan Williams wrote: > diff --git a/include/linux/io.h b/include/linux/io.h > index 080a4fbf2ba4..2983b6e63970 100644 > --- a/include/linux/io.h > +++ b/include/linux/io.h > @@ -192,4 +192,15 @@ static inline int arch_phys_wc_index(int handle) > #endif > #

Re: [PATCH 09/10] arch: introduce memremap()

2015-07-20 Thread Mark Rutland
On Mon, Jul 20, 2015 at 03:39:44PM +0100, Dan Williams wrote: > On Mon, Jul 20, 2015 at 5:00 AM, Mark Rutland wrote: > > Hi, > > > > On Mon, Jul 20, 2015 at 01:18:23AM +0100, Dan Williams wrote: > >> Existing users of ioremap_cache() are mapping memory that is known in > >> advance to not have i/o

Re: [PATCH 09/10] arch: introduce memremap()

2015-07-20 Thread Dan Williams
On Mon, Jul 20, 2015 at 5:00 AM, Mark Rutland wrote: > Hi, > > On Mon, Jul 20, 2015 at 01:18:23AM +0100, Dan Williams wrote: >> Existing users of ioremap_cache() are mapping memory that is known in >> advance to not have i/o side effects. These users are forced to cast >> away the __iomem annotat

Re: [PATCH 09/10] arch: introduce memremap()

2015-07-20 Thread Mark Rutland
Hi, On Mon, Jul 20, 2015 at 01:18:23AM +0100, Dan Williams wrote: > Existing users of ioremap_cache() are mapping memory that is known in > advance to not have i/o side effects. These users are forced to cast > away the __iomem annotation, or otherwise neglect to fix the sparse > errors thrown wh

[PATCH 09/10] arch: introduce memremap()

2015-07-19 Thread Dan Williams
Existing users of ioremap_cache() are mapping memory that is known in advance to not have i/o side effects. These users are forced to cast away the __iomem annotation, or otherwise neglect to fix the sparse errors thrown when dereferencing pointers to this memory. Provide memremap() as a non __io