netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH net-next v4] net/core: fix wrong return value in __splice_segment
@ 2025-08-14  1:01 Pengtao He
  2025-08-14 13:38 ` Willem de Bruijn
  0 siblings, 1 reply; 3+ messages in thread
From: Pengtao He @ 2025-08-14  1:01 UTC (permalink / raw)
  To: David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni,
	Simon Horman, Willem de Bruijn, Mina Almasry, Jason Xing,
	Michal Luczaj, Eric Biggers, Alexander Lobakin
  Cc: netdev, linux-kernel, Pengtao He

If *len is equal to 0 at the beginning of __splice_segment
it returns true directly. But when decreasing *len from
a positive number to 0 in __splice_segment, it returns false.
The caller needs to call __splice_segment again.

Recheck *len if it changes, return true in time.
Reduce unnecessary calls to __splice_segment.

Signed-off-by: Pengtao He <hept.hept.hept@gmail.com>
---
v4:
Correct the commit message.
v3:
Reduce once condition evaluation.
v2:
Correct the commit message and target tree.
v1:
https://lore.kernel.org/netdev/20250723063119.24059-1-hept.hept.hept@gmail.com/
---
 net/core/skbuff.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index ee0274417948..23b776cd9879 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -3112,7 +3112,9 @@ static bool __splice_segment(struct page *page, unsigned int poff,
 		poff += flen;
 		plen -= flen;
 		*len -= flen;
-	} while (*len && plen);
+		if (!*len)
+			return true;
+	} while (plen);
 
 	return false;
 }
-- 
2.49.0


^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH net-next v4] net/core: fix wrong return value in __splice_segment
  2025-08-14  1:01 [PATCH net-next v4] net/core: fix wrong return value in __splice_segment Pengtao He
@ 2025-08-14 13:38 ` Willem de Bruijn
  2025-08-19  2:07   ` Pengtao He
  0 siblings, 1 reply; 3+ messages in thread
From: Willem de Bruijn @ 2025-08-14 13:38 UTC (permalink / raw)
  To: Pengtao He, David S. Miller, Eric Dumazet, Jakub Kicinski,
	Paolo Abeni, Simon Horman, Willem de Bruijn, Mina Almasry,
	Jason Xing, Michal Luczaj, Eric Biggers, Alexander Lobakin
  Cc: netdev, linux-kernel, Pengtao He

Pengtao He wrote:
> If *len is equal to 0 at the beginning of __splice_segment
> it returns true directly. But when decreasing *len from
> a positive number to 0 in __splice_segment, it returns false.
> The caller needs to call __splice_segment again.
> 
> Recheck *len if it changes, return true in time.
> Reduce unnecessary calls to __splice_segment.

Fix is a strong term. The existing behavior is correct, it just takes
an extra pass through the loop in caller __skb_splice_bits. As also
indicated by this patch targeting net-next.

I would suggest something like "net: avoid one loop iteration in __skb_splice_bits"


> Signed-off-by: Pengtao He <hept.hept.hept@gmail.com>
> ---
> v4:
> Correct the commit message.
> v3:
> Reduce once condition evaluation.
> v2:
> Correct the commit message and target tree.
> v1:
> https://lore.kernel.org/netdev/20250723063119.24059-1-hept.hept.hept@gmail.com/
> ---
>  net/core/skbuff.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/net/core/skbuff.c b/net/core/skbuff.c
> index ee0274417948..23b776cd9879 100644
> --- a/net/core/skbuff.c
> +++ b/net/core/skbuff.c
> @@ -3112,7 +3112,9 @@ static bool __splice_segment(struct page *page, unsigned int poff,
>  		poff += flen;
>  		plen -= flen;
>  		*len -= flen;
> -	} while (*len && plen);
> +		if (!*len)
> +			return true;
> +	} while (plen);
>  
>  	return false;
>  }
> -- 
> 2.49.0
> 



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH net-next v4] net/core: fix wrong return value in __splice_segment
  2025-08-14 13:38 ` Willem de Bruijn
@ 2025-08-19  2:07   ` Pengtao He
  0 siblings, 0 replies; 3+ messages in thread
From: Pengtao He @ 2025-08-19  2:07 UTC (permalink / raw)
  To: willemdebruijn.kernel
  Cc: aleksander.lobakin, almasrymina, davem, ebiggers, edumazet,
	hept.hept.hept, horms, kerneljasonxing, kuba, linux-kernel, mhal,
	netdev, pabeni, willemb

> Pengtao He wrote:
> > If *len is equal to 0 at the beginning of __splice_segment
> > it returns true directly. But when decreasing *len from
> > a positive number to 0 in __splice_segment, it returns false.
> > The caller needs to call __splice_segment again.
> > 
> > Recheck *len if it changes, return true in time.
> > Reduce unnecessary calls to __splice_segment.
> 
> Fix is a strong term. The existing behavior is correct, it just takes
> an extra pass through the loop in caller __skb_splice_bits. As also
> indicated by this patch targeting net-next.
> 
> I would suggest something like "net: avoid one loop iteration in __skb_splice_bits"

Thanks for the detailed and good suggestion. I will correct it.

> 
> 
> > Signed-off-by: Pengtao He <hept.hept.hept@gmail.com>
> > ---
> > v4:
> > Correct the commit message.
> > v3:
> > Reduce once condition evaluation.
> > v2:
> > Correct the commit message and target tree.
> > v1:
> > https://lore.kernel.org/netdev/20250723063119.24059-1-hept.hept.hept@gmail.com/
> > ---
> >  net/core/skbuff.c | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> > 
> > diff --git a/net/core/skbuff.c b/net/core/skbuff.c
> > index ee0274417948..23b776cd9879 100644
> > --- a/net/core/skbuff.c
> > +++ b/net/core/skbuff.c
> > @@ -3112,7 +3112,9 @@ static bool __splice_segment(struct page *page, unsigned int poff,
> >  		poff += flen;
> >  		plen -= flen;
> >  		*len -= flen;
> > -	} while (*len && plen);
> > +		if (!*len)
> > +			return true;
> > +	} while (plen);
> >  
> >  	return false;
> >  }
> > -- 
> > 2.49.0
> > 

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2025-08-19  2:07 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-08-14  1:01 [PATCH net-next v4] net/core: fix wrong return value in __splice_segment Pengtao He
2025-08-14 13:38 ` Willem de Bruijn
2025-08-19  2:07   ` Pengtao He

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).