Re: [PATCH v3 01/14] x86, uaccess: introduce copy_from_iter_flushcache for pmem / cache-bypass operations

From: Dan Williams
Date: Sun Jun 18 2017 - 22:03:10 EST

On Sun, Jun 18, 2017 at 1:28 AM, Christoph Hellwig <hch@xxxxxx> wrote:
> On Fri, Jun 09, 2017 at 01:23:51PM -0700, Dan Williams wrote:
>> Implement a __copy_from_user_inatomic_flushcache, memcpy_page_flushcache, and
>> memcpy_flushcache, that guarantee that the destination buffer is not dirty in
>> the cpu cache on completion. The new copy_from_iter_flushcache and sub-routines
> Wouldn't writethrough be a better name?

I started with _writethrough, Ingo suggested _wt, and then Toshi
rightly pointed out that _wt might lead applications to assume that
the involved cache-lines are valid on return which may not be true. So
we settled on _flushcache in this thread [1].


>> will be used to replace the "pmem api" (include/linux/pmem.h +
>> arch/x86/include/asm/pmem.h). The availability of copy_from_iter_flushcache()
>> and memcpy_flushcache() are gated by the CONFIG_ARCH_HAS_UACCESS_FLUSHCACHE
>> config symbol, and fallback to copy_from_iter_nocache() and plain memcpy()
>> otherwise.
> What is UACCESS about memcpy_flushcache?

The uaccess part comes from the fact that the conversion provides all
the _flushcache versions of the copy_from_iter() operations
(__copy_from_user_flushcache, memcpy_page_flushcache,
memcpy_flushcache). It also stems from Al asking that
__copy_user_flushcache() live in arch/x86/lib/usercopy_64.c. That
said, I wouldn't object to a different name for the config symbol.