Re: [PATCH RFC v2 00/12] socket sendmsg MSG_ZEROCOPY
From: Eric Dumazet <hidden>
Date: 2017-02-28 23:22:19
Also in:
linux-api
Possibly related (same subject, not in this thread)
- 2017-03-02 · Re: [PATCH RFC v2 00/12] socket sendmsg MSG_ZEROCOPY · Andy Lutomirski <luto@amacapital.net>
- 2017-03-01 · Re: [PATCH RFC v2 00/12] socket sendmsg MSG_ZEROCOPY · Eric Dumazet <hidden>
- 2017-03-01 · Re: [PATCH RFC v2 00/12] socket sendmsg MSG_ZEROCOPY · David Miller <davem@davemloft.net>
- 2017-03-01 · Re: [PATCH RFC v2 00/12] socket sendmsg MSG_ZEROCOPY · David Miller <davem@davemloft.net>
- 2017-03-01 · Re: [PATCH RFC v2 00/12] socket sendmsg MSG_ZEROCOPY · Tom Herbert <hidden>
On Tue, 2017-02-28 at 14:52 -0800, Andy Lutomirski wrote:
The user pages are a gift to the kernel. The application may not modify this memory ever, otherwise the page cache and on-disk data may differ. This is just not okay IMO.
TCP works just fine in this case. TX checksum will be computed by the NIC after/while data is copied. If really the application changes the data, that will not cause any problems, other than user side consistency. This is why we require a copy (for all buffers that came from zero-copy) if network stack hits a device that can not offload TX checksum. Even pwrite() does not guarantee consistency if multiple threads are using it on overlapping regions.