2022-07-28 03:22:24

by Yongji Xie

Subject: [PATCH v4 2/5] vduse: Use memcpy_{to,from}_page() in do_bounce()

kmap_atomic() is being deprecated in favor of kmap_local_page().

The use of kmap_atomic() in do_bounce() is all thread local therefore
kmap_local_page() is a sufficient replacement.

Convert to kmap_local_page() but, instead of open coding it,
use the helpers memcpy_to_page() and memcpy_from_page().

Signed-off-by: Xie Yongji <[email protected]>
Acked-by: Jason Wang <[email protected]>
---
drivers/vdpa/vdpa_user/iova_domain.c | 9 ++++-----
1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/vdpa/vdpa_user/iova_domain.c b/drivers/vdpa/vdpa_user/iova_domain.c
index bca1f0b8850c..50d7c08d5450 100644
--- a/drivers/vdpa/vdpa_user/iova_domain.c
+++ b/drivers/vdpa/vdpa_user/iova_domain.c
@@ -138,18 +138,17 @@ static void do_bounce(phys_addr_t orig, void *addr, size_t size,
{
unsigned long pfn = PFN_DOWN(orig);
unsigned int offset = offset_in_page(orig);
- char *buffer;
+ struct page *page;
unsigned int sz = 0;

while (size) {
sz = min_t(size_t, PAGE_SIZE - offset, size);

- buffer = kmap_atomic(pfn_to_page(pfn));
+ page = pfn_to_page(pfn);
if (dir == DMA_TO_DEVICE)
- memcpy(addr, buffer + offset, sz);
+ memcpy_from_page(addr, page, offset, sz);
else
- memcpy(buffer + offset, addr, sz);
- kunmap_atomic(buffer);
+ memcpy_to_page(page, offset, addr, sz);

size -= sz;
pfn++;
--
2.20.1

2022-07-29 02:39:53

by Muchun Song

Subject: Re: [PATCH v4 2/5] vduse: Use memcpy_{to,from}_page() in do_bounce()

On Thu, Jul 28, 2022 at 11:20 AM Xie Yongji <[email protected]> wrote:
>
> kmap_atomic() is being deprecated in favor of kmap_local_page().
>
> The use of kmap_atomic() in do_bounce() is all thread local therefore
> kmap_local_page() is a sufficient replacement.
>
> Convert to kmap_local_page() but, instead of open coding it,
> use the helpers memcpy_to_page() and memcpy_from_page().
>
> Signed-off-by: Xie Yongji <[email protected]>
> Acked-by: Jason Wang <[email protected]>

Reviewed-by: Muchun Song <[email protected]>