On Mon, Mar 27, 2017 at 08:41:10PM -0500, Bjorn Helgaas wrote:
> On Mon, Mar 27, 2017 at 10:49:30AM +0100, Lorenzo Pieralisi wrote:
> > The PCI specifications (Rev 3.0, 3.2.5 "Transaction Ordering and
> > Posting") mandate non-posted configuration transactions. As further
> > highlighted in the PCIe specifications (4.0 - Rev0.3, "Ordering
> > Considerations for the Enhanced Configuration Access Mechanism"),
> > through ECAM and ECAM-derivative configuration mechanism, the memory
> > mapped transactions from the host CPU into Configuration Requests on the
> > PCI express fabric may create ordering problems for software because
> > writes to memory address are typically posted transactions (unless the
> > architecture can enforce through virtual address mapping non-posted
> > write transactions behaviour) but writes to Configuration Space are not
> > posted on the PCI express fabric.
> >
> > Current DT and ACPI host bridge controllers map PCI configuration space
> > (ECAM and ECAM-derivative) into the virtual address space through
> > ioremap() calls, that are non-cacheable device accesses on most
> > architectures, but may provide "bufferable" or "posted" write semantics
> > in architecture like eg ARM/ARM64 that allow ioremap'ed regions writes
> > to be buffered in the bus connecting the host CPU to the PCI fabric;
> > this behaviour, as underlined in the PCIe specifications, may trigger
> > transactions ordering rules and must be prevented.
> >
> > Introduce a new generic and explicit API to create a memory
> > mapping for ECAM and ECAM-derivative config space area that
> > defaults to ioremap_nocache() (which should provide a sane default
> > behaviour) but still allowing architectures on which ioremap_nocache()
> > results in posted write transactions to override the function
> > call with an arch specific implementation that complies with
> > the PCI specifications for configuration transactions.
> >
> > Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@xxxxxxx>
> > Cc: Arnd Bergmann <arnd@xxxxxxxx>
> > Cc: Will Deacon <will.deacon@xxxxxxx>
> > Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> > Cc: Bjorn Helgaas <bhelgaas@xxxxxxxxxx>
> > Cc: Richard Henderson <rth@xxxxxxxxxxx>
> > Cc: Russell King <linux@xxxxxxxxxxxxxxx>
> > Cc: Tony Luck <tony.luck@xxxxxxxxx>
> > Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
> > Cc: Ingo Molnar <mingo@xxxxxxxxxx>
> > Cc: Haavard Skinnemoen <hskinnemoen@xxxxxxxxx>
> > ---
> > arch/alpha/include/asm/io.h | 1 +
> > arch/avr32/include/asm/io.h | 1 +
> > arch/frv/include/asm/io.h | 1 +
> > arch/ia64/include/asm/io.h | 1 +
> > arch/x86/include/asm/io.h | 1 +
> > include/asm-generic/io.h | 4 ++++
> > 6 files changed, 9 insertions(+)
> >
> > diff --git a/arch/alpha/include/asm/io.h b/arch/alpha/include/asm/io.h
> > index ff40491..27379ea 100644
> > --- a/arch/alpha/include/asm/io.h
> > +++ b/arch/alpha/include/asm/io.h
> > @@ -300,6 +300,7 @@ static inline void __iomem * ioremap_nocache(unsigned long offset,
> > }
> >
> > #define ioremap_uc ioremap_nocache
> > +#define ioremap_nopost ioremap_nocache
> >
> > static inline void iounmap(volatile void __iomem *addr)
> > {
> > diff --git a/arch/avr32/include/asm/io.h b/arch/avr32/include/asm/io.h
> > index f855646..3f1ced8 100644
> > --- a/arch/avr32/include/asm/io.h
> > +++ b/arch/avr32/include/asm/io.h
> > @@ -298,6 +298,7 @@ extern void __iounmap(void __iomem *addr);
> > #define ioremap_wc ioremap_nocache
> > #define ioremap_wt ioremap_nocache
> > #define ioremap_uc ioremap_nocache
> > +#define ioremap_nopost ioremap_nocache
> >
> > #define cached(addr) P1SEGADDR(addr)
> > #define uncached(addr) P2SEGADDR(addr)
> > diff --git a/arch/frv/include/asm/io.h b/arch/frv/include/asm/io.h
> > index 8062fc7..302fb8c 100644
> > --- a/arch/frv/include/asm/io.h
> > +++ b/arch/frv/include/asm/io.h
> > @@ -290,6 +290,7 @@ static inline void __iomem *ioremap_fullcache(unsigned long physaddr, unsigned l
> >
> > #define ioremap_wc ioremap_nocache
> > #define ioremap_uc ioremap_nocache
> > +#define ioremap_nopost ioremap_nocache
> >
> > extern void iounmap(void volatile __iomem *addr);
> >
> > diff --git a/arch/ia64/include/asm/io.h b/arch/ia64/include/asm/io.h
> > index 5de673a..70a4985 100644
> > --- a/arch/ia64/include/asm/io.h
> > +++ b/arch/ia64/include/asm/io.h
> > @@ -434,6 +434,7 @@ static inline void __iomem * ioremap_cache (unsigned long phys_addr, unsigned lo
> > }
> > #define ioremap_cache ioremap_cache
> > #define ioremap_uc ioremap_nocache
> > +#define ioremap_nopost ioremap_nocache
> >
> >
> > /*
> > diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h
> > index 7afb0e2..50b292f 100644
> > --- a/arch/x86/include/asm/io.h
> > +++ b/arch/x86/include/asm/io.h
> > @@ -171,6 +171,7 @@ static inline unsigned int isa_virt_to_bus(volatile void *address)
> > extern void __iomem *ioremap_nocache(resource_size_t offset, unsigned long size);
> > extern void __iomem *ioremap_uc(resource_size_t offset, unsigned long size);
> > #define ioremap_uc ioremap_uc
> > +#define ioremap_nopost ioremap_nocache
>
> These are all the same as the default from asm-generic.h. Do we really
> need these definitions in alpha, avr32, frv, ia64, x86?
Those arches do not include asm-generic.h (I suppose for a good reason)
but a definition is needed anyway if we want code using ioremap_nopost()
to be unconditional. This is one way of doing it, there are others not
sure they are any better, I am open to suggestions.
Thanks,
Lorenzo
> > extern void __iomem *ioremap_cache(resource_size_t offset, unsigned long size);
> > extern void __iomem *ioremap_prot(resource_size_t offset, unsigned long size, unsigned long prot_val);
> > diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h
> > index 7ef015e..0e81938 100644
> > --- a/include/asm-generic/io.h
> > +++ b/include/asm-generic/io.h
> > @@ -915,6 +915,10 @@ extern void ioport_unmap(void __iomem *p);
> > #endif /* CONFIG_GENERIC_IOMAP */
> > #endif /* CONFIG_HAS_IOPORT_MAP */
> >
> > +#ifndef ioremap_nopost
> > +#define ioremap_nopost ioremap_nocache
> > +#endif
> > +
> > #ifndef xlate_dev_kmem_ptr
> > #define xlate_dev_kmem_ptr xlate_dev_kmem_ptr
> > static inline void *xlate_dev_kmem_ptr(void *addr)
> > --
> > 2.10.0
> >
