diff options
author | Alan Stern <stern@rowland.harvard.edu> | 2010-02-26 11:49:39 -0500 |
---|---|---|
committer | Greg Kroah-Hartman <gregkh@suse.de> | 2010-03-02 14:55:07 -0800 |
commit | 49d6271b85a3e18062eaf4d6f8d899abe00a7725 (patch) | |
tree | 83b92caacf276bdd054a03320fd6f48cbe96c562 /crypto/crypto_null.c | |
parent | efe7daf2231a6beb59f0f641461d19fa62fb83ab (diff) | |
download | linux-49d6271b85a3e18062eaf4d6f8d899abe00a7725.tar.gz linux-49d6271b85a3e18062eaf4d6f8d899abe00a7725.tar.bz2 linux-49d6271b85a3e18062eaf4d6f8d899abe00a7725.zip |
usb-storage: use max_hw_sectors instead of max_sectors
This patch (as1347) makes some adjustments to the way usb-storage
handles the request-queue parameters.
USB host controllers are able to handle arbitrarily long
scatter-gather lists, since they are limited only by main memory and
not by the controller hardware. Hence the sg_tablesize field in the
host template can be increased to the maximum value.
Drivers like usb-storage aren't supposed to touch the queue's
max_sectors parameter; instead they are supposed to use the
max_hw_sectors value. Accordingly, the patch replaces calls of
queue_max_sectors() with calls of queue_max_hw_sectors(). Oddly
enough, the blk_queue_max_sectors() routine is nevertheless still
appropriate.
The existing code imposes a limit of SCSI_DEFAULT_MAX_SECTORS (1024)
on the values accepted by the max_sectors attribute file. There's no
reason not to accept larger values, so the limit is removed. (It
would be nice to change the file's name to max_hw_sectors, but the old
name is already a well-established API.)
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
CC: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Diffstat (limited to 'crypto/crypto_null.c')
0 files changed, 0 insertions, 0 deletions