| Message ID | 1522420199-23548-1-git-send-email-okaya@codeaurora.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 3/30/2018 10:29 AM, Sinan Kaya wrote: > In the abscence of a write barrier or when using a strongly ordered > architecture, writeX() should at least have a compiler barrier in > it to prevent commpiler from clobbering the execution order. Same is true for readX(). I'll wait for review feedback on this before posting another change for readX().
On Fri, Mar 30, 2018 at 10:29:58AM -0400, Sinan Kaya wrote: > The default implementation of mapping writeX() to __raw_writeX() is wrong. > writeX() has stronger ordering semantics. Compiler is allowed to reorder > __raw_writeX(). > > In the abscence of a write barrier or when using a strongly ordered > architecture, writeX() should at least have a compiler barrier in > it to prevent commpiler from clobbering the execution order. You want the barrier _before_ the call to __raw_writel() - you need to ensure that writes to memory are emitted by the compiler _before_ the write to the hardware - the write to the hardware may start DMA, and it may be reading data that the program thinks it previously wrote. Similarly, for readl(), you need the barrier after __raw_readl() to ensure that other reads in the program aren't scheduled before a potential DMA status register read.
diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h index b4531e3..fbbf2bb 100644 --- a/include/asm-generic/io.h +++ b/include/asm-generic/io.h @@ -153,6 +153,7 @@ static inline void writeb(u8 value, volatile void __iomem *addr) static inline void writew(u16 value, volatile void __iomem *addr) { __raw_writew(cpu_to_le16(value), addr); + barrier(); } #endif @@ -161,6 +162,7 @@ static inline void writew(u16 value, volatile void __iomem *addr) static inline void writel(u32 value, volatile void __iomem *addr) { __raw_writel(__cpu_to_le32(value), addr); + barrier(); } #endif @@ -170,6 +172,7 @@ static inline void writel(u32 value, volatile void __iomem *addr) static inline void writeq(u64 value, volatile void __iomem *addr) { __raw_writeq(__cpu_to_le64(value), addr); + barrier(); } #endif #endif /* CONFIG_64BIT */
The default implementation of mapping writeX() to __raw_writeX() is wrong. writeX() has stronger ordering semantics. Compiler is allowed to reorder __raw_writeX(). In the abscence of a write barrier or when using a strongly ordered architecture, writeX() should at least have a compiler barrier in it to prevent commpiler from clobbering the execution order. Signed-off-by: Sinan Kaya <okaya@codeaurora.org> --- include/asm-generic/io.h | 3 +++ 1 file changed, 3 insertions(+)