Bug 119351 - misleading information about buffer size limits in tcp(7)
Summary: misleading information about buffer size limits in tcp(7)
Status: NEW
Alias: None
Product: Documentation
Classification: Unclassified
Component: man-pages (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: documentation_man-pages@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-05-31 17:49 UTC by Nadav Har'El
Modified: 2016-05-31 17:55 UTC (History)
0 users

See Also:
Kernel Version:
Subsystem:
Regression: No
Bisected commit-id:


Attachments

Description Nadav Har'El 2016-05-31 17:49:59 UTC
The tcp(7) discussion on SO_SNDBUF and SO_RCVBUF contains the statement that: "The maximum sizes for socket buffers declared  via  the  SO_SNDBUF  and SO_RCVBUF    mechanisms    are   limited   by   the   values   in   the       /proc/sys/net/core/rmem_max  and   /proc/sys/net/core/wmem_max   files."

However, other sources seem to suggest that the files /proc/sys/net/ipv4/tcp_wmem and /proc/sys/net/ipv4/tcp_rmem (mentioned in the previous paragraph) actually contain the relevant limits (the third value in each file), while the core/ files are only relevant to other protocols. This is actually explained in tcp(7), down in the "tcp_rmem" section of the "/proc interfaces".
Comment 1 Nadav Har'El 2016-05-31 17:55:26 UTC
Hmm, reading the text again, the tcp_rmem section actually states that: "This value does not override  the  global net.core.rmem_max.". It's not clear what this actually means: Does it mean tcp_rmem maximum is just ignored? That it cannot *increase* core.rmem_max? Or what?

This whole topic is really unclear in tcp(7) :-(

Note You need to log in before you can comment on or make changes to this bug.