Hi Christoph,
On Tue, 22 Feb 2022, Christoph Hellwig wrote:
> On Tue, Feb 22, 2022 at 10:09:42AM +1100, Finn Thain wrote:
> > Bring aha152x into line with 10 other drivers which assign
> > scsi_host_template.cmd_size = sizeof(struct scsi_pointer)
> > and avoid the "struct foo { struct bar; };" silliness.
> >
> > Remove a pointless scsi_pointer->have_data_in assignment.
>
> I think this going in the wrong direction. The scsi_pointer should go
> away entirelym and the fields actually used by the driver should move
> into the aha152x_cmd_priv structure instead.
>
> Same for all other drivers still using the scsi_pointer.
>
This patch is addressing an inconsistency in the patches already accepted
into 5.18/scsi-staging in Martin's repo.
A number of Bart's patches had the same effect as the patch you're
objecting to here. Hence,
$ git grep "cmd_size.*scsi_pointer"
drivers/scsi/a2091.c: .cmd_size = sizeof(struct scsi_pointer),
drivers/scsi/a3000.c: .cmd_size = sizeof(struct scsi_pointer),
drivers/scsi/fdomain.c: .cmd_size = sizeof(struct scsi_pointer),
drivers/scsi/gvp11.c: .cmd_size = sizeof(struct scsi_pointer),
drivers/scsi/imm.c: .cmd_size = sizeof(struct scsi_pointer),
drivers/scsi/mvme147.c: .cmd_size = sizeof(struct scsi_pointer),
drivers/scsi/pcmcia/nsp_cs.c: .cmd_size = sizeof(struct scsi_pointer),
drivers/scsi/pcmcia/sym53c500_cs.c: .cmd_size = sizeof(struct scsi_pointer),
drivers/scsi/ppa.c: .cmd_size = sizeof(struct scsi_pointer),
drivers/scsi/sgiwd93.c: .cmd_size = sizeof(struct scsi_pointer),
Since that series was very popular with reviewers, and being that this
patch is just more of the same, I have no idea as to how to proceed.
Are you asking me to rework Bart's series? Or are you asking Martin to
drop it, or both, or neither...
