diff for duplicates of <20170320132132.GN17263@arm.com> diff --git a/a/1.txt b/N1/1.txt index 6555732..a1eb976 100644 --- a/a/1.txt +++ b/N1/1.txt @@ -1,10 +1,10 @@ On Thu, Mar 16, 2017 at 12:17:27PM +0000, David Woodhouse wrote: > On Fri, 2016-04-15 at 14:09 +0100, Will Deacon wrote: > > -> > > + if (write_combine) -> > > + vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot); -> > > + else -> > > + vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); +> > > +?????if (write_combine) +> > > +?????????????vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot); +> > > +?????else +> > > +?????????????vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); > > > > For consistency with ioremap, this should be pgprot_device. > @@ -15,9 +15,9 @@ created using pgprot_noncached is that the former allows for early acknowledgement of writes (e.g. at a bridge). See this recent series from Lorenzo that is also trying to clean this up: -http://lkml.kernel.org/r/20170227151436.18698-1-lorenzo.pieralisi@arm.com +http://lkml.kernel.org/r/20170227151436.18698-1-lorenzo.pieralisi at arm.com -> I note that VFIO is using pgprot_noncached() too, in vfio_pci_mmap() — +> I note that VFIO is using pgprot_noncached() too, in vfio_pci_mmap() ? > where it open-codes an entirely arch-agnostic version of > pci_mmap_page_range() all for itself. Should that be changed to > pgprot_device() too? @@ -41,13 +41,13 @@ it anyway" regions in /dev/mem. > 'resourceX_wc' variant which has write-combining support. > > How about we do the following (probably not in this order): -> • Kill pci_mmap_page_range() entirely. -> • Implement a generic version which has (arch-assisted) WC support -> but no knowledge of the horrid pci_resource_to_user() mapping. -> • Require pci_user_to_resource() to be provided by platforms with -> HAVE_ARCH_PCI_RESOURCE_TO_USER, and call that from *generic* code, -> for the legacy procfs interface, before invoking the generic -> replacement for pci_mmap_page_range(). +> ?? Kill pci_mmap_page_range() entirely. +> ?? Implement a generic version which has (arch-assisted) WC support +> ? ?but no knowledge of the horrid pci_resource_to_user() mapping. +> ?? Require pci_user_to_resource() to be provided by platforms with +> ? ?HAVE_ARCH_PCI_RESOURCE_TO_USER, and call that from *generic* code, +> ? ?for the legacy procfs interface, before invoking the generic +> ? ?replacement for pci_mmap_page_range(). > > (Yes, we still need to support mmap of I/O resources on... is it only > powerpc? And there are a few inconsistencies, like powerpc forcing WC diff --git a/a/content_digest b/N1/content_digest index bdda31a..6bbac7d 100644 --- a/a/content_digest +++ b/N1/content_digest @@ -1,30 +1,19 @@ "ref\01460581856-12380-1-git-send-email-jerin.jacob@caviumnetworks.com\0" "ref\020160415130953.GI22906@arm.com\0" "ref\01489666647.4195.238.camel@infradead.org\0" - "From\0Will Deacon <will.deacon@arm.com>\0" - "Subject\0Re: [PATCH v2] arm64: pci: add support for pci_mmap_page_range\0" + "From\0will.deacon@arm.com (Will Deacon)\0" + "Subject\0[PATCH v2] arm64: pci: add support for pci_mmap_page_range\0" "Date\0Mon, 20 Mar 2017 13:21:32 +0000\0" - "To\0David Woodhouse <dwmw2@infradead.org>\0" - "Cc\0Jerin Jacob <jerin.jacob@caviumnetworks.com>" - lorenzo.pieralisi@arm.com - david.daney@cavium.com - catalin.marinas@arm.com - Liviu.Dudau@arm.com - rrichter@cavium.com - hanjun.guo@linaro.org - bhelgaas@google.com - linux-arm-kernel@lists.infradead.org - linux-pci <linux-pci@vger.kernel.org> - " benh@kernel.crashing.org\0" + "To\0linux-arm-kernel@lists.infradead.org\0" "\00:1\0" "b\0" "On Thu, Mar 16, 2017 at 12:17:27PM +0000, David Woodhouse wrote:\n" "> On Fri, 2016-04-15 at 14:09 +0100, Will Deacon wrote:\n" "> > \n" - "> > > +\302\240\302\240\302\240\302\240\302\240if (write_combine)\n" - "> > > +\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);\n" - "> > > +\302\240\302\240\302\240\302\240\302\240else\n" - "> > > +\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);\n" + "> > > +?????if (write_combine)\n" + "> > > +?????????????vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);\n" + "> > > +?????else\n" + "> > > +?????????????vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);\n" "> > \n" "> > For consistency with ioremap, this should be pgprot_device.\n" "> \n" @@ -35,9 +24,9 @@ "acknowledgement of writes (e.g. at a bridge). See this recent series from\n" "Lorenzo that is also trying to clean this up:\n" "\n" - "http://lkml.kernel.org/r/20170227151436.18698-1-lorenzo.pieralisi@arm.com\n" + "http://lkml.kernel.org/r/20170227151436.18698-1-lorenzo.pieralisi at arm.com\n" "\n" - "> I note that VFIO is using pgprot_noncached() too, in vfio_pci_mmap() \342\200\224\n" + "> I note that VFIO is using pgprot_noncached() too, in vfio_pci_mmap() ?\n" "> where it open-codes an entirely arch-agnostic version of\n" "> pci_mmap_page_range() all for itself. Should that be changed to\n" "> pgprot_device() too?\n" @@ -61,13 +50,13 @@ "> 'resourceX_wc' variant which has write-combining support.\n" "> \n" "> How about we do the following (probably not in this order):\n" - "> \302\240\342\200\242 Kill pci_mmap_page_range() entirely.\n" - "> \302\240\342\200\242 Implement a generic version which has (arch-assisted) WC support\n" - "> \302\240 \302\240but no knowledge of the horrid pci_resource_to_user() mapping.\n" - "> \302\240\342\200\242 Require pci_user_to_resource() to be provided by platforms with\n" - "> \302\240 \302\240HAVE_ARCH_PCI_RESOURCE_TO_USER, and call that from *generic* code,\n" - "> \302\240 \302\240for the legacy procfs interface, before invoking the generic\n" - "> \302\240 \302\240replacement for pci_mmap_page_range().\n" + "> ?? Kill pci_mmap_page_range() entirely.\n" + "> ?? Implement a generic version which has (arch-assisted) WC support\n" + "> ? ?but no knowledge of the horrid pci_resource_to_user() mapping.\n" + "> ?? Require pci_user_to_resource() to be provided by platforms with\n" + "> ? ?HAVE_ARCH_PCI_RESOURCE_TO_USER, and call that from *generic* code,\n" + "> ? ?for the legacy procfs interface, before invoking the generic\n" + "> ? ?replacement for pci_mmap_page_range().\n" "> \n" "> (Yes, we still need to support mmap of I/O resources on... is it only\n" "> powerpc? And there are a few inconsistencies, like powerpc forcing WC\n" @@ -78,4 +67,4 @@ "\n" Will -7acb82621526f9d815a973b3266d04ad269399ff9cd0150de507a657ced90446 +72399b324c1246eab35a9621e2bbbbdf3cd494a96347e4bc2f28202fa57352b5
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.