| Message ID | 20190620154952.31286-1-volodymyr_babchuk@epam.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | xen/arm: optee: fix compilation with GCC 4.8 | expand |
On Thu, 20 Jun 2019, Volodymyr Babchuk wrote: > GCC 4.8 is unable to see that variables guest_pg, guest_data and > xen_data will be always initialized before access, so we need to > initialize them earlier. > > Suggested-by: Stefano Stabellini <stefanos@xilinx.com> > Signed-off-by: Volodymyr Babchuk <volodymyr_babchuk@epam.com> I verified that it works correctly. Thanks for the patch! (Julien, I didn't commit it yet to give you a chance to give it a look too.) Reviewed-by: Stefano Stabellini <sstabellini@kernel.org> > --- > xen/arch/arm/tee/optee.c | 13 +++++++++++++ > 1 file changed, 13 insertions(+) > > diff --git a/xen/arch/arm/tee/optee.c b/xen/arch/arm/tee/optee.c > index 14381d6b2d..5526875e6f 100644 > --- a/xen/arch/arm/tee/optee.c > +++ b/xen/arch/arm/tee/optee.c > @@ -717,6 +717,19 @@ static int translate_noncontig(struct optee_domain *ctx, > gfn = gaddr_to_gfn(param->u.tmem.buf_ptr & > ~(OPTEE_MSG_NONCONTIG_PAGE_SIZE - 1)); > > + /* > + * We are initializing guest_pg, guest_data and xen_data with NULL > + * to make GCC 4.8 happy, as it can't infer that those variables > + * will be initialized with correct values in the loop below. > + * > + * This silences old GCC, but can lead to NULL dereference, in > + * case of programmer's mistake. To minimize chance of this, we > + * are initializing those variables there, instead of doing this > + * at beginning of the function. > + */ > + guest_pg = NULL; > + xen_data = NULL; > + guest_data = NULL; > while ( pg_count ) > { > struct page_info *page; > -- > 2.21.0 >
On 21/06/2019 01:28, Stefano Stabellini wrote: > On Thu, 20 Jun 2019, Volodymyr Babchuk wrote: >> GCC 4.8 is unable to see that variables guest_pg, guest_data and >> xen_data will be always initialized before access, so we need to >> initialize them earlier. >> >> Suggested-by: Stefano Stabellini <stefanos@xilinx.com> >> Signed-off-by: Volodymyr Babchuk <volodymyr_babchuk@epam.com> > > I verified that it works correctly. Thanks for the patch! (Julien, I > didn't commit it yet to give you a chance to give it a look too.) > > Reviewed-by: Stefano Stabellini <sstabellini@kernel.org> Acked-by: Julien Grall <julien.grall@arm.com> I have committed it now. Cheers,
diff --git a/xen/arch/arm/tee/optee.c b/xen/arch/arm/tee/optee.c index 14381d6b2d..5526875e6f 100644 --- a/xen/arch/arm/tee/optee.c +++ b/xen/arch/arm/tee/optee.c @@ -717,6 +717,19 @@ static int translate_noncontig(struct optee_domain *ctx, gfn = gaddr_to_gfn(param->u.tmem.buf_ptr & ~(OPTEE_MSG_NONCONTIG_PAGE_SIZE - 1)); + /* + * We are initializing guest_pg, guest_data and xen_data with NULL + * to make GCC 4.8 happy, as it can't infer that those variables + * will be initialized with correct values in the loop below. + * + * This silences old GCC, but can lead to NULL dereference, in + * case of programmer's mistake. To minimize chance of this, we + * are initializing those variables there, instead of doing this + * at beginning of the function. + */ + guest_pg = NULL; + xen_data = NULL; + guest_data = NULL; while ( pg_count ) { struct page_info *page;
GCC 4.8 is unable to see that variables guest_pg, guest_data and xen_data will be always initialized before access, so we need to initialize them earlier. Suggested-by: Stefano Stabellini <stefanos@xilinx.com> Signed-off-by: Volodymyr Babchuk <volodymyr_babchuk@epam.com> --- xen/arch/arm/tee/optee.c | 13 +++++++++++++ 1 file changed, 13 insertions(+)