Re: [PATCH v2] mtd: spi-nor: Use DMA-safe buffer for JEDEC ID in spi_nor_read_id()

[Boris Brezillon <boris.brezillon@xxxxxxxxxxx>]

On Mon, 17 Sep 2018 15:51:24 +0300
Jarkko Nikula <jarkko.nikula@xxxxxxxxxxxxxxx> wrote:

> On 09/17/2018 03:28 PM, Boris Brezillon wrote:
> > On Mon, 17 Sep 2018 12:04:43 +0300
> > Jarkko Nikula <jarkko.nikula@xxxxxxxxxxxxxxx> wrote:
> >   
> >> On 09/17/2018 11:46 AM, Boris Brezillon wrote:  
> >>> On Mon, 17 Sep 2018 11:41:31 +0300
> >>> Jarkko Nikula <jarkko.nikula@xxxxxxxxxxxxxxx> wrote:
> >>>      
> >>>> On 09/17/2018 11:28 AM, Boris Brezillon wrote:  
> >>>>>
> >>>>> Hm, I just had a look at all ->read/write_reg() call sites and it seems
> >>>>> we have the same problem in other places. Maybe it's just better to fix
> >>>>> the problem in m25p80 ->read/write_reg() implems for now  
> > 
> > Do you plan to send a new fix doing that, or should I? I'd really like
> > to have the fix queued to mtd/master early this week, so that it
> > spends a bit of time in -next.
> >   
> I guess better if you could as I haven't started anything yet. I'm not 
> so familiar with the code to see what would be best way to fix, 
> reverting m25p80_read_reg/m25p80_write_reg back to spi_write(), the 
> whole commit or do some kmemdup() there.

kmemdup is better (or any safe alternative validated by Kees), we
definitely don't want to go back to spi_write().


______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/