On Fri, Feb 14, 2020 at 06:36:50PM -0500, Sasha Levin wrote:
> On Fri, Feb 14, 2020 at 09:43:14AM -0800, Eric Biggers wrote:
> > On Fri, Feb 14, 2020 at 11:24:04AM -0500, Sasha Levin wrote:
> > > From: "Gustavo A. R. Silva" <gustavo@xxxxxxxxxxxxxx>
> > >
> > > [ Upstream commit 987f028b8637cfa7658aa456ae73f8f21a7a7f6f ]
> > >
> > > Old code in the kernel uses 1-byte and 0-byte arrays to indicate the
> > > presence of a "variable length array":
> > >
> > > struct something {
> > > int length;
> > > u8 data[1];
> > > };
> > >
> > > struct something *instance;
> > >
> > > instance = kmalloc(sizeof(*instance) + size, GFP_KERNEL);
> > > instance->length = size;
> > > memcpy(instance->data, source, size);
> > >
> > > There is also 0-byte arrays. Both cases pose confusion for things like
> > > sizeof(), CONFIG_FORTIFY_SOURCE, etc.[1] Instead, the preferred mechanism
> > > to declare variable-length types such as the one above is a flexible array
> > > member[2] which need to be the last member of a structure and empty-sized:
> > >
> > > struct something {
> > > int stuff;
> > > u8 data[];
> > > };
> > >
> > > Also, by making use of the mechanism above, we will get a compiler warning
> > > in case the flexible array does not occur last in the structure, which
> > > will help us prevent some kind of undefined behavior bugs from being
> > > unadvertenly introduced[3] to the codebase from now on.
> > >
> > > [1] https://github.com/KSPP/linux/issues/21
> > > [2] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
> > > [3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")
> > >
> > > Signed-off-by: Gustavo A. R. Silva <gustavo@xxxxxxxxxxxxxx>
> > > Link: https://lore.kernel.org/r/20200120235326.GA29231@xxxxxxxxxxxxxx
> > > Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> > > Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
> > > ---
> > > drivers/char/hpet.c | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/char/hpet.c b/drivers/char/hpet.c
> > > index 5b38d7a8202a1..38c2ae93ce492 100644
> > > --- a/drivers/char/hpet.c
> > > +++ b/drivers/char/hpet.c
> > > @@ -112,7 +112,7 @@ struct hpets {
> > > unsigned long hp_delta;
> > > unsigned int hp_ntimer;
> > > unsigned int hp_which;
> > > - struct hpet_dev hp_dev[1];
> > > + struct hpet_dev hp_dev[];
> > > };
> > >
> >
> > Umm, why are you backporting this without the commit that fixes it? Does your
>
> mhm, for some reason it failed to apply to 4.19 and older. I can look at
> that.
>
> > AUTOSEL process really still not pay attention to Fixes tags? They are there
> > for a reason.
>
> Yes, it looks at the Fixes tag, thank you for the explanation.
If there is commit A which is fixed by commit B, and commit A applies to the
stable branch but commit B does not, then commit A shouldn't be backported
without manually fixing commit B first -- since otherwise a known bug would be
backported.
You really need to be handling this as part of your AUTOSEL process. If it
happened here, it has happened elsewhere too, given the hundreds/thousands of
commits you're selecting for stable.
- Eric
