Re: [bug report] EDAC/{skx_common,i10nm}: Refactor show_retry_rd_err_log()

[Dan Carpenter <dan.carpenter@linaro.org>]

On Wed, Apr 23, 2025 at 04:26:21PM +0000, Luck, Tony wrote:
> > >     362                         /* Clear RRL status if RRL in Linux control mode. */
> > >     363                         if (retry_rd_err_log == 2 && !j && (log & status_mask))
> > > --> 364                                 write_imc_reg(imc, ch, offset, width, log &
> > > ~status_mask);
> > >
> > > This will clear the high 32 bits of log.
> >
> > It's OK to clear the high 32 bits of 'log', as we only write the low 32 bits of 'log' to
> > the 1st RRL register, which is a u32 type.
> >
> > To improve code sanity, it might be worthwhile to create a patch that changes
> > 'status_mask' to a u64 type. @Luck, Tony, should I create a fix patch on top of the
> > current patch series to fix this warning?
> 
> Qiuxu,
> 
> Yes. Write a patch for this and give a Reported-by: credit to Dan.
> 

Thanks.  But mention what you said that it's a false positive.

It's hard to silence a warning like this in Smatch because you'd need to
read into the write_imc_reg() to see that only 4 byte widths are
supported.

regards,
dan carpenter