Hi Viljami, On Thu, 2013-03-07 at 13:53 +0200, Viljami Ilola wrote: > I encountered this problem while trying to install OS. > > Kernel gave me messages like: > [12015.915146] SCSI OP 2ah with too big sectors 1347 exceeds backend > hw_max_sectors: 1024 > [12015.977744] SCSI OP 2ah with too big sectors 2048 exceeds backend > hw_max_sectors: 1024 > Thanks for reporting! > This seems to fix problem for me now, but I guess there is better solution. > --- target_core_file.c~ 2013-03-04 01:11:05.000000000 +0200 > +++ target_core_file.c 2013-03-07 12:48:52.058447827 +0200 > @@ -175,7 +175,8 @@ > } > > dev->dev_attrib.hw_block_size = FD_BLOCKSIZE; > - dev->dev_attrib.hw_max_sectors = FD_MAX_SECTORS; > + dev->dev_attrib.hw_max_sectors = fd_dev->fd_dev_size / FD_BLOCKSIZE; > } > > fd_dev->fd_block_size = dev->dev_attrib.hw_block_size; > > -- This value is actually limited by the total number of iovecs that vfs_writev() + vfs_readv() can accept (1024) in a single function call after the >= v3.5 kernel se_task removal. As you've seen with the block_size=512 default, this is currently limited to 512K payloads with FD_MAX_SECTORS=1024. So I think the proper resolution here is to go ahead and use a larger FD_MAX_SECTORS default (8192) for the block_size=512 FILEIO case, and then reduce hw_max_sectors as larger block_sizes are enabled via configfs on a per device backend basis I'll be sending out a patch to address this soon. Thanks! --nab -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
