> -----Original Message-----
> From: Mark Brown <broonie@xxxxxxxxxx>
> Sent: 27 February 2023 23:53
> To: Krishna Yarlagadda <kyarlagadda@xxxxxxxxxx>
> Cc: linux-spi@xxxxxxxxxxxxxxx; linux-tegra@xxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx; thierry.reding@xxxxxxxxx; Jonathan Hunter
> <jonathanh@xxxxxxxxxx>; Sowjanya Komatineni
> <skomatineni@xxxxxxxxxx>; Laxman Dewangan <ldewangan@xxxxxxxxxx>
> Subject: Re: [PATCH] spi: tegra210-quad: Fix iterator outside loop
>
> On Mon, Feb 27, 2023 at 10:36:18AM +0000, Krishna Yarlagadda wrote:
>
> > > > > > - if (!xfer->cs_change) {
> > > > > > - tegra_qspi_transfer_end(spi);
> > > > > > - spi_transfer_delay_exec(xfer);
> > > > > > - }
> > > updating the length of the message.
> > > > > This looks like it'll do the wrong thing and do a change on every
> > > > > transfer if cs_change isn't set?
>
> > > > This condition is hit only in data phase which is end of message.
>
> > > Shouldn't this just be moved into the DATA_TRANSFER case statement?
> > Calling transfer_end after updating message length.
>
> Something seems to be mangled with your quoting/new text here so it's a
> bit unclear what you're saying here but if you're saying that this is
> due to needing to call tegra_qspi_transfer_end() after updating the
> length I'm not sure why - AFAICT that function doesn't reference the
> transfer length at all, it just writes out a command to configure the
> chip select? There's no issue with the message being finalised since
> that happens in the caller.
Yes. Length should not matter. Will move into DATA_TRANSFER case.
Also need to handle error case.
