Re: [PATCH net] tcp: remove empty skb from write queue in error cases

2019-09-11 Thread Eric Dumazet
On Wed, Sep 11, 2019 at 7:36 PM Christoph Paasch wrote: > > Hello, > > On Mon, Aug 26, 2019 at 11:04 AM Eric Dumazet wrote: > > > > Vladimir Rutsky reported stuck TCP sessions after memory pressure > > events. Edge Trigger epoll() user would never receive an EPOLLOUT > > notification allowing the

Re: [PATCH net] tcp: remove empty skb from write queue in error cases

2019-09-11 Thread Christoph Paasch
Hello, On Mon, Aug 26, 2019 at 11:04 AM Eric Dumazet wrote: > > Vladimir Rutsky reported stuck TCP sessions after memory pressure > events. Edge Trigger epoll() user would never receive an EPOLLOUT > notification allowing them to retry a sendmsg(). > > Jason tested the case of sk_stream_alloc_skb

Re: [PATCH net] tcp: remove empty skb from write queue in error cases

2019-08-27 Thread David Miller
From: Eric Dumazet Date: Mon, 26 Aug 2019 09:19:15 -0700 > Vladimir Rutsky reported stuck TCP sessions after memory pressure > events. Edge Trigger epoll() user would never receive an EPOLLOUT > notification allowing them to retry a sendmsg(). > > Jason tested the case of sk_stream_alloc_skb() r

Re: [PATCH net] tcp: remove empty skb from write queue in error cases

2019-08-26 Thread Eric Dumazet
On 8/26/19 9:56 PM, Jason Baron wrote: > > > On 8/26/19 12:19 PM, Eric Dumazet wrote: >> Vladimir Rutsky reported stuck TCP sessions after memory pressure >> events. Edge Trigger epoll() user would never receive an EPOLLOUT >> notification allowing them to retry a sendmsg(). >> >> Jason tested

Re: [PATCH net] tcp: remove empty skb from write queue in error cases

2019-08-26 Thread Jason Baron
On 8/26/19 12:19 PM, Eric Dumazet wrote: > Vladimir Rutsky reported stuck TCP sessions after memory pressure > events. Edge Trigger epoll() user would never receive an EPOLLOUT > notification allowing them to retry a sendmsg(). > > Jason tested the case of sk_stream_alloc_skb() returning NULL,

Re: [PATCH net] tcp: remove empty skb from write queue in error cases

2019-08-26 Thread Neal Cardwell
On Mon, Aug 26, 2019 at 12:19 PM Eric Dumazet wrote: > > Vladimir Rutsky reported stuck TCP sessions after memory pressure > events. Edge Trigger epoll() user would never receive an EPOLLOUT > notification allowing them to retry a sendmsg(). > > Jason tested the case of sk_stream_alloc_skb() retur

Re: [PATCH net] tcp: remove empty skb from write queue in error cases

2019-08-26 Thread Soheil Hassas Yeganeh
On Mon, Aug 26, 2019 at 12:19 PM Eric Dumazet wrote: > > Vladimir Rutsky reported stuck TCP sessions after memory pressure > events. Edge Trigger epoll() user would never receive an EPOLLOUT > notification allowing them to retry a sendmsg(). > > Jason tested the case of sk_stream_alloc_skb() retur

[PATCH net] tcp: remove empty skb from write queue in error cases

2019-08-26 Thread Eric Dumazet
Vladimir Rutsky reported stuck TCP sessions after memory pressure events. Edge Trigger epoll() user would never receive an EPOLLOUT notification allowing them to retry a sendmsg(). Jason tested the case of sk_stream_alloc_skb() returning NULL, but there are other paths that could lead both sendmsg