On Tue, 14 Nov 2023 at 12:21, Avri Altman <Avri.Altman@xxxxxxx> wrote:
>
> > > > }
> > > > @@ -1032,6 +1044,28 @@ static inline void
> > > > mmc_blk_reset_success(struct
> > > mmc_blk_data *md, int type)
> > > > md->reset_done &= ~type;
> > > > }
> > > >
> > > > +/* close-ended ffu */
> > > > +static void mmc_blk_check_ce_ffu(struct mmc_queue_req *mq_rq) {
> > > > + struct mmc_blk_ioc_data **idata = mq_rq->drv_op_data;
> > > > +
> > > > + if (mq_rq->ioc_count != 4)
> > > > + return;
> > > > +
> > > > + if (idata[0]->ic.opcode != MMC_SWITCH)
> > > > + return;
> > > > +
> > > > + if (MMC_EXTRACT_INDEX_FROM_ARG(idata[0]->ic.arg) !=
> > > > + EXT_CSD_MODE_CONFIG)
> > > > + return;
> > > > +
> > > > + if (idata[1]->ic.opcode == MMC_SET_BLOCK_COUNT &&
> > > > + idata[2]->ic.opcode == MMC_WRITE_MULTIPLE_BLOCK) {
> > > > + idata[1]->flags |= MMC_BLK_IOC_DROP;
> > > > + idata[2]->flags |= MMC_BLK_IOC_SBC;
> > > > + }
> > >
> > > Could this be more generic e.g. simply
> > >
> > > for (i = 1; i < mq_rq->ioc_count; i++)
> > > if (idata[i - 1]->ic.opcode == MMC_SET_BLOCK_COUNT &&
> > > mmc_op_multi(idata[i + 1]->ic.opcode)) {
> > I guess you meant (idata[i]
> >
> > > idata[i - 1]->flags |= MMC_BLK_IOC_DROP;
> > > idata[i]->flags |= MMC_BLK_IOC_SBC;
> > > }
> > >
> > > with no need to check for 4 cmds, MMC_SWITCH or
> > EXT_CSD_MODE_CONFIG
> > Ditto
> Or, I can follow your proposal.
> Will send a v3.
Please do! I think it's better to solve this in a generic way, rather
than making it specific for FFU.
Kind regards
Uffe
