copy_from_iter_nocache() is only "nocache" for iovecs. Enhance it to also
use a nocache copy for bvecs. This improves performance by 2-3X when
splice()ing to a file in a DAX-mounted, pmem-backed file system.
Cc: Ross Zwisler <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: "H. Peter Anvin" <[email protected]>
Cc: <[email protected]>
Cc: Al Viro <[email protected]>
Cc: Dan Williams <[email protected]>
Signed-off-by: Brian Boylston <[email protected]>
Reviewed-by: Toshi Kani <[email protected]>
Reported-by: Oliver Moreno <[email protected]>
---
arch/x86/include/asm/pmem.h | 6 +++---
lib/iov_iter.c | 11 +++++++++--
2 files changed, 12 insertions(+), 5 deletions(-)
diff --git a/arch/x86/include/asm/pmem.h b/arch/x86/include/asm/pmem.h
index 643eba4..d071f45c 100644
--- a/arch/x86/include/asm/pmem.h
+++ b/arch/x86/include/asm/pmem.h
@@ -73,12 +73,12 @@ static inline void arch_wb_cache_pmem(void *addr, size_t size)
}
/*
- * copy_from_iter_nocache() on x86 only uses non-temporal stores for iovec
- * iterators, so for other types (bvec & kvec) we must do a cache write-back.
+ * copy_from_iter_nocache() on x86 uses non-temporal stores for iovec and
+ * bvec iterators, but for kvec we must do a cache write-back.
*/
static inline bool __iter_needs_pmem_wb(struct iov_iter *i)
{
- return iter_is_iovec(i) == false;
+ return (i->type & ITER_KVEC) == ITER_KVEC;
}
/**
diff --git a/lib/iov_iter.c b/lib/iov_iter.c
index 7e3138c..df4cb00 100644
--- a/lib/iov_iter.c
+++ b/lib/iov_iter.c
@@ -342,6 +342,13 @@ static void memcpy_from_page(char *to, struct page *page, size_t offset, size_t
kunmap_atomic(from);
}
+static void memcpy_from_page_nocache(char *to, struct page *page, size_t offset, size_t len)
+{
+ char *from = kmap_atomic(page);
+ __copy_from_user_inatomic_nocache(to, from, len);
+ kunmap_atomic(from);
+}
+
static void memcpy_to_page(struct page *page, size_t offset, const char *from, size_t len)
{
char *to = kmap_atomic(page);
@@ -392,8 +399,8 @@ size_t copy_from_iter_nocache(void *addr, size_t bytes, struct iov_iter *i)
iterate_and_advance(i, bytes, v,
__copy_from_user_nocache((to += v.iov_len) - v.iov_len,
v.iov_base, v.iov_len),
- memcpy_from_page((to += v.bv_len) - v.bv_len, v.bv_page,
- v.bv_offset, v.bv_len),
+ memcpy_from_page_nocache((to += v.bv_len) - v.bv_len,
+ v.bv_page, v.bv_offset, v.bv_len),
memcpy((to += v.iov_len) - v.iov_len, v.iov_base, v.iov_len)
)
--
1.8.3.1
On Fri, Oct 07, 2016 at 10:55:11AM -0500, Brian Boylston wrote:
> copy_from_iter_nocache() is only "nocache" for iovecs. Enhance it to also
> use a nocache copy for bvecs. This improves performance by 2-3X when
> splice()ing to a file in a DAX-mounted, pmem-backed file system.
> +static void memcpy_from_page_nocache(char *to, struct page *page, size_t offset, size_t len)
> +{
> + char *from = kmap_atomic(page);
> + __copy_from_user_inatomic_nocache(to, from, len);
> + kunmap_atomic(from);
> +}
At the very least, it will blow up on any architecture with split
userland and kernel MMU contexts. You *can't* feed a kernel pointer
to things like that and expect it to work. At the very least, you
need to add memcpy_nocache() and have it default to memcpy(), with
non-dummy version on x86. And use _that_, rather than messing with
__copy_from_user_inatomic_nocache()
On Fri, 2016-10-07 at 18:08 +0100, Al Viro wrote:
> On Fri, Oct 07, 2016 at 10:55:11AM -0500, Brian Boylston wrote:
> >
> > copy_from_iter_nocache() is only "nocache" for iovecs. Enhance it
> > to also use a nocache copy for bvecs. This improves performance by
> > 2-3X when splice()ing to a file in a DAX-mounted, pmem-backed file
> > system.
>
> >
> > +static void memcpy_from_page_nocache(char *to, struct page *page,
> > size_t offset, size_t len)
> > +{
> > + char *from = kmap_atomic(page);
> > + __copy_from_user_inatomic_nocache(to, from, len);
> > + kunmap_atomic(from);
> > +}
>
> At the very least, it will blow up on any architecture with split
> userland and kernel MMU contexts. You *can't* feed a kernel pointer
> to things like that and expect it to work. At the very least, you
> need to add memcpy_nocache() and have it default to memcpy(), with
> non-dummy version on x86. And use _that_, rather than messing with
> __copy_from_user_inatomic_nocache()
Good point. I think we can add memcpy_nocache() which calls
__copy_from_user_inatomic_nocache() on x86 and defauts to memcpy() on
other architectures.
Thanks,
-Toshi
Kani, Toshimitsu wrote on 2016-10-10:
> On Fri, 2016-10-07 at 18:08 +0100, Al Viro wrote:
>> On Fri, Oct 07, 2016 at 10:55:11AM -0500, Brian Boylston wrote:
>>>
>>> copy_from_iter_nocache() is only "nocache" for iovecs. Enhance it
>>> to also use a nocache copy for bvecs. This improves performance by
>>> 2-3X when splice()ing to a file in a DAX-mounted, pmem-backed file
>>> system.
>>
>>>
>>> +static void memcpy_from_page_nocache(char *to, struct page *page,
>>> size_t offset, size_t len)
>>> +{
>>> + char *from = kmap_atomic(page);
>>> + __copy_from_user_inatomic_nocache(to, from, len);
>>> + kunmap_atomic(from);
>>> +}
>>
>> At the very least, it will blow up on any architecture with split
>> userland and kernel MMU contexts. You *can't* feed a kernel pointer
>> to things like that and expect it to work. At the very least, you
>> need to add memcpy_nocache() and have it default to memcpy(), with
>> non-dummy version on x86. And use _that_, rather than messing with
>> __copy_from_user_inatomic_nocache()
>
> Good point. I think we can add memcpy_nocache() which calls
> __copy_from_user_inatomic_nocache() on x86 and defauts to memcpy() on
> other architectures.
Thanks, Al and Toshi, for the feedback. I'll re-work and come back.
Brian