regmap bug fix

Marek Behun marek.behun at nic.cz
Thu Mar 25 13:28:54 CET 2021


On Thu, 25 Mar 2021 15:41:55 +1300
Simon Glass <sjg at chromium.org> wrote:

> eHi Marek,
> 
> On Thu, 25 Mar 2021 at 13:46, Marek Behun <marek.behun at nic.cz> wrote:
> >
> > On Thu, 25 Mar 2021 13:38:13 +1300
> > Simon Glass <sjg at chromium.org> wrote:
> >  
> > > Hi Marek,
> > >
> > > On Wed, 17 Mar 2021 at 04:19, Marek Behun <marek.behun at nic.cz> wrote:  
> > > >
> > > > Simon, Heiko, Bin,
> > > >
> > > > Pratyush discovered that the solution implemented by the patch
> > > >   regmap: fix a serious pointer casting bug
> > > > is wrong. The cpu_to_le32() / le32_to_cpu() shifts data to the correct
> > > > position, but on big endian machines it also reverses byte order.
> > > > Somehow this went right through my head when I thought this up.
> > > >
> > > > I have sent a new version, with subject
> > > >   [PATCH u-boot v3.1 01/39] regmap: fix a serious pointer casting bug
> > > >
> > > > The new solution utilizes an union { u8; u16; u32; u64; }, since all
> > > > members of an union start at the same address.
> > > >
> > > > Could you please review this? Thanks.  
> > >
> > > You already have my review tag on that.
> > >
> > > I dropped it from u-boot-dm/next. The problem was that it v1 was in my
> > > queue, but not later versions.
> > >
> > > Regards,
> > > Simon  
> >
> > I don't have your review for version v3.1 :) to which you just replied.
> >
> > I had your review tag for v3, and sent v3.1 as a reply to
> > Pratyush'scomplains on v3.  
> 
> Fractional versions? In that case I'm going to hold out for version 3.125.

:) I sent it as a reply for v3, and did not want to call it v4
because for v4 I am going to send the whole series again. Sorry.

Anyway I probably will send v4 next week, there are still some problems
and I am working on something different now.

Marek


More information about the U-Boot mailing list