linux-kernel - Re: [PATCH] PCI: mvebu: Fix endianity when accessing pci emul bridge members

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 12 Aug 2022 15:58:41 +0200
From:   Pali Rohár <pali@...nel.org>
To:     Robin Murphy <robin.murphy@....com>
Cc:     Lorenzo Pieralisi <lpieralisi@...nel.org>,
        Thomas Petazzoni <thomas.petazzoni@...tlin.com>,
        Krzysztof Wilczyński <kw@...ux.com>,
        Bjorn Helgaas <bhelgaas@...gle.com>, linux-pci@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] PCI: mvebu: Fix endianity when accessing pci emul bridge
 members

On Friday 12 August 2022 11:32:59 Robin Murphy wrote:
> On 2022-08-12 10:40, Pali Rohár wrote:
> > PCI emul bridge members iolimitupper, iobaseupper, memlimit and membase are
> > of type __le16, so correctly access these members via le16_to_cpu() macros.
> > 
> > Fixes: 4ded69473adb ("PCI: mvebu: Propagate errors when updating PCI_IO_BASE and PCI_MEM_BASE registers")
> > Reported-by: kernel test robot <lkp@...el.com>
> > Signed-off-by: Pali Rohár <pali@...nel.org>
> > ---
> >   drivers/pci/controller/pci-mvebu.c | 12 ++++++------
> >   1 file changed, 6 insertions(+), 6 deletions(-)
> > 
> > diff --git a/drivers/pci/controller/pci-mvebu.c b/drivers/pci/controller/pci-mvebu.c
> > index c1ffdb06c971..00ea0836b81a 100644
> > --- a/drivers/pci/controller/pci-mvebu.c
> > +++ b/drivers/pci/controller/pci-mvebu.c
> > @@ -523,7 +523,7 @@ static int mvebu_pcie_handle_iobase_change(struct mvebu_pcie_port *port)
> >   	/* Are the new iobase/iolimit values invalid? */
> >   	if (conf->iolimit < conf->iobase ||
> > -	    conf->iolimitupper < conf->iobaseupper)
> > +	    le16_to_cpu(conf->iolimitupper) < le16_to_cpu(conf->iobaseupper))
> >   		return mvebu_pcie_set_window(port, port->io_target, port->io_attr,
> >   					     &desired, &port->iowin);
> > @@ -535,10 +535,10 @@ static int mvebu_pcie_handle_iobase_change(struct mvebu_pcie_port *port)
> >   	 * is the CPU address.
> >   	 */
> >   	desired.remap = ((conf->iobase & 0xF0) << 8) |
> > -			(conf->iobaseupper << 16);
> > +			le16_to_cpu(conf->iobaseupper << 16);
> 
> This will always give 0, even when natively LE.

You are right, I overlooked it and I put closing parenthesis at wrong
place. Bit shifting should be applied after le to cpu conversion. I will
fix it in V2.

> >   	desired.base = port->pcie->io.start + desired.remap;
> >   	desired.size = ((0xFFF | ((conf->iolimit & 0xF0) << 8) |
> > -			 (conf->iolimitupper << 16)) -
> > +			 le16_to_cpu(conf->iolimitupper << 16)) -
> 
> Similarly here.
> 
> Robin.
> 
> >   			desired.remap) +
> >   		       1;
> > @@ -552,7 +552,7 @@ static int mvebu_pcie_handle_membase_change(struct mvebu_pcie_port *port)
> >   	struct pci_bridge_emul_conf *conf = &port->bridge.conf;
> >   	/* Are the new membase/memlimit values invalid? */
> > -	if (conf->memlimit < conf->membase)
> > +	if (le16_to_cpu(conf->memlimit) < le16_to_cpu(conf->membase))
> >   		return mvebu_pcie_set_window(port, port->mem_target, port->mem_attr,
> >   					     &desired, &port->memwin);
> > @@ -562,8 +562,8 @@ static int mvebu_pcie_handle_membase_change(struct mvebu_pcie_port *port)
> >   	 * window to setup, according to the PCI-to-PCI bridge
> >   	 * specifications.
> >   	 */
> > -	desired.base = ((conf->membase & 0xFFF0) << 16);
> > -	desired.size = (((conf->memlimit & 0xFFF0) << 16) | 0xFFFFF) -
> > +	desired.base = ((le16_to_cpu(conf->membase) & 0xFFF0) << 16);
> > +	desired.size = (((le16_to_cpu(conf->memlimit) & 0xFFF0) << 16) | 0xFFFFF) -
> >   		       desired.base + 1;
> >   	return mvebu_pcie_set_window(port, port->mem_target, port->mem_attr, &desired,