On Wed, Feb 21, 2024 at 12:49:46PM +0100, Markus Armbruster wrote:
> Date: Wed, 21 Feb 2024 12:49:46 +0100
> From: Markus Armbruster <arm...@redhat.com>
> Subject: Re: [PATCH 5/6] hw/pci-bridge/cxl_upstream: Fix missing
> ERRP_GUARD() in cxl_usp_realize()
>
> Zhao Liu <zhao1....@linux.intel.com> writes:
>
> > From: Zhao Liu <zhao1....@intel.com>
> >
> > As the comment in qapi/error, dereferencing @errp requires
> > ERRP_GUARD():
> >
> > * = Why, when and how to use ERRP_GUARD() =
> > *
> > * Without ERRP_GUARD(), use of the @errp parameter is restricted:
> > * - It must not be dereferenced, because it may be null.
> > * - It should not be passed to error_prepend() or
> > * error_append_hint(), because that doesn't work with &error_fatal.
> > * ERRP_GUARD() lifts these restrictions.
> > *
> > * To use ERRP_GUARD(), add it right at the beginning of the function.
> > * @errp can then be used without worrying about the argument being
> > * NULL or &error_fatal.
> > *
> > * Using it when it's not needed is safe, but please avoid cluttering
> > * the source with useless code.
> >
> > Currently, since cxl_usp_realize() - as a PCIDeviceClass.realize()
> > method - doesn't get the NULL errp parameter, it doesn't trigger the
> > dereference issue.
> >
> > To follow the requirement of errp, add missing ERRP_GUARD() in
> > cxl_usp_realize()().
> >
> > Suggested-by: Markus Armbruster <arm...@redhat.com>
> > Signed-off-by: Zhao Liu <zhao1....@intel.com>
> > ---
> > Suggested by credit:
> > Markus: Referred his explanation about ERRP_GUARD().
> > ---
> > hw/pci-bridge/cxl_upstream.c | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/hw/pci-bridge/cxl_upstream.c b/hw/pci-bridge/cxl_upstream.c
> > index e87eb4017713..03d123cca0ef 100644
> > --- a/hw/pci-bridge/cxl_upstream.c
> > +++ b/hw/pci-bridge/cxl_upstream.c
> > @@ -289,6 +289,7 @@ static void free_default_cdat_table(CDATSubHeader
> > **cdat_table, int num,
> >
> > static void cxl_usp_realize(PCIDevice *d, Error **errp)
> > {
> > + ERRP_GUARD();
> > PCIEPort *p = PCIE_PORT(d);
> > CXLUpstreamPort *usp = CXL_USP(d);
> > CXLComponentState *cxl_cstate = &usp->cxl_cstate;
>
> The dereference is
>
> cxl_doe_cdat_init(cxl_cstate, errp);
> if (*errp) {
> goto err_cap;
> }
>
> As noted in review of PATCH 3, we check *errp, because
> cxl_doe_cdat_init() returns void. Could be improved to return bool,
> along with its callees ct3_load_cdat() and ct3_build_cdat(), but that's
> a slightly more ambitious cleanup, so
>
> Reviewed-by: Markus Armbruster <arm...@redhat.com>
>
Thanks! It's a good idea and I can continue to consider such the cleanup
in the follow up.
Will also add the dereference description in commit message to make
review easier.
Regards,
Zhao
