Re: [PATCH 088/190] Revert "mmc_spi: add a status check for spi_sync_locked"

[Greg Kroah-Hartman]

On Wed, Apr 28, 2021 at 11:41:25AM +0300, Laurent Pinchart wrote:
> Hi Greg,
> 
> On Wed, Apr 28, 2021 at 09:18:03AM +0200, Greg Kroah-Hartman wrote:
> > On Thu, Apr 22, 2021 at 10:08:45AM +0200, Ulf Hansson wrote:
> > > On Wed, 21 Apr 2021 at 15:19, Laurent Pinchart wrote:
> > > > On Wed, Apr 21, 2021 at 02:59:23PM +0200, Greg Kroah-Hartman wrote:
> > > > > This reverts commit 611025983b7976df0183390a63a2166411d177f1.
> > > > >
> > > > > Commits from @umn.edu addresses have been found to be submitted in "bad
> > > > > faith" to try to test the kernel community's ability to review "known
> > > > > malicious" changes.  The result of these submissions can be found in a
> > > > > paper published at the 42nd IEEE Symposium on Security and Privacy
> > > > > entitled, "Open Source Insecurity: Stealthily Introducing
> > > > > Vulnerabilities via Hypocrite Commits" written by Qiushi Wu (University
> > > > > of Minnesota) and Kangjie Lu (University of Minnesota).
> > > > >
> > > > > Because of this, all submissions from this group must be reverted from
> > > > > the kernel tree and will need to be re-reviewed again to determine if
> > > > > they actually are a valid fix.  Until that work is complete, remove this
> > > > > change to ensure that no problems are being introduced into the
> > > > > codebase.
> > > > >
> > > > > Cc: Kangjie Lu <kjlu@xxxxxxx>
> > > > > Cc: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
> > > > > Cc: Ulf Hansson <ulf.hansson@xxxxxxxxxx>
> > > > > Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> > > >
> > > > Acked-by: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
> > > >
> > > > I don't spot an obvious issue with the original patch though.
> > > >
> > > > > ---
> > > > >  drivers/mmc/host/mmc_spi.c | 4 ----
> > > > >  1 file changed, 4 deletions(-)
> > > > >
> > > > > diff --git a/drivers/mmc/host/mmc_spi.c b/drivers/mmc/host/mmc_spi.c
> > > > > index 02f4fd26e76a..cc40b050e302 100644
> > > > > --- a/drivers/mmc/host/mmc_spi.c
> > > > > +++ b/drivers/mmc/host/mmc_spi.c
> > > > > @@ -800,10 +800,6 @@ mmc_spi_readblock(struct mmc_spi_host *host, struct spi_transfer *t,
> > > > >       }
> > > > >
> > > > >       status = spi_sync_locked(spi, &host->m);
> > > > > -     if (status < 0) {
> > > > > -             dev_dbg(&spi->dev, "read error %d\n", status);
> > > > > -             return status;
> > > > > -     }
> > > 
> > > Returning here means we never give back the ownership of the buffer to
> > > the CPU. Can that be considered as vulnerability?
> > 
> > It's a "resource leak", which is a bug.  If you want to declare that as
> > a "vulnerability" or not, I do not know.  Personally I do not think it
> > is...
> 
> How is that a resource leak ? The dma_sync_single_for_device() calls
> above this block don't take the buffer ownership away from the CPU in a
> way that leaks it.

Ick, this is really twisty.

The calls to dma_sync_single_for_device() call down to
dma_direct_sync_single_for_device() for when you can directly talk to
the hardware, or to the platform sync_single_for_device() callback for
arches that can not.

Those are messy, but the worst they all seem to do is invalidate or
flush some caches, no type of resource allocation that I could determine
here.

So I'll trust you, and drop the revert and mark this one as "good
patch" :)

Thanks again for the review, much appreciated.

greg k-h