Re: [PATCH] scsi_lib.c: avoid max_hw_segments warning

2006-12-11 Thread James Bottomley
On Mon, 2006-12-11 at 18:18 +0100, Michael Schmitz wrote: > > Actually, isn't the correct fix to change SG_NONE to 1? > > > > We're deprecating the non-sg paths, however use_sg == 1 is equivalent to > > use_sg==0, you just set it up differently. > > I don't know how it's handled in the current cod

Re: [PATCH] scsi_lib.c: avoid max_hw_segments warning

2006-12-11 Thread Michael Schmitz
> > as per request by Geert Uytterhoeven: the patch below avoids a warning in > > blk_queue_max_hw_segments(), caused by a SCSI host adapter using SG_NONE. > > Please apply. > > Actually, isn't the correct fix to change SG_NONE to 1? > > We're deprecating the non-sg paths, however use_sg == 1 is eq

Re: [PATCH] scsi_lib.c: avoid max_hw_segments warning

2006-12-11 Thread James Bottomley
On Mon, 2006-12-11 at 17:09 +0100, Michael Schmitz wrote: > Hi, > > as per request by Geert Uytterhoeven: the patch below avoids a warning in > blk_queue_max_hw_segments(), caused by a SCSI host adapter using SG_NONE. > Please apply. Actually, isn't the correct fix to change SG_NONE to 1? We're

[PATCH] scsi_lib.c: avoid max_hw_segments warning

2006-12-11 Thread Michael Schmitz
Hi, as per request by Geert Uytterhoeven: the patch below avoids a warning in blk_queue_max_hw_segments(), caused by a SCSI host adapter using SG_NONE. Please apply. > > The second item concerns a warning given for each SCSI target scanned on > > the Falcon, where sg_tablesize is set to SG_NONE (