| commit | 02107148349f31eee7c0fb06fd7a880df73dbd20 | |
| author | Chuck Lever <cel@netapp.com> | |
| Tue, 3 Jan 2006 08:55:49 +0000 (3 09:55 +0100) | ||
| committer | Trond Myklebust <Trond.Myklebust@netapp.com> | |
| Fri, 6 Jan 2006 19:58:55 +0000 (6 14:58 -0500) | ||
| tree | 37bffd81e08b8e50394ce89a1aa7a3961f0ffbe7 | treesnapshot (tar.gz zip) |
| parent | 03c21733938aad0758f5f88e1cc7ede69fc3c910 | commitdiff |
SUNRPC: switchable buffer allocation
Add RPC client transport switch support for replacing buffer management
on a per-transport basis.
In the current IPv4 socket transport implementation, RPC buffers are
allocated as needed for each RPC message that is sent. Some transport
implementations may choose to use pre-allocated buffers for encoding,
sending, receiving, and unmarshalling RPC messages, however. For
transports capable of direct data placement, the buffers can be carved
out of a pre-registered area of memory rather than from a slab cache.
Test-plan:
Millions of fsx operations. Performance characterization with "sio" and
"iozone". Use oprofile and other tools to look for significant regression
in CPU utilization.
Signed-off-by: Chuck Lever <cel@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Add RPC client transport switch support for replacing buffer management
on a per-transport basis.
In the current IPv4 socket transport implementation, RPC buffers are
allocated as needed for each RPC message that is sent. Some transport
implementations may choose to use pre-allocated buffers for encoding,
sending, receiving, and unmarshalling RPC messages, however. For
transports capable of direct data placement, the buffers can be carved
out of a pre-registered area of memory rather than from a slab cache.
Test-plan:
Millions of fsx operations. Performance characterization with "sio" and
"iozone". Use oprofile and other tools to look for significant regression
in CPU utilization.
Signed-off-by: Chuck Lever <cel@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>