| 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
| ||
|
Message-ID: <CAGETcx9iA_irfpO2yJSPszeNrfwfYAV0KkZ+AyB7gcDo0v8p1g@mail.gmail.com>
Date: Thu, 9 Jan 2020 10:52:46 -0800
From: Saravana Kannan <saravanak@...gle.com>
To: Ard Biesheuvel <ard.biesheuvel@...aro.org>
Cc: Ard Biesheuvel <ardb@...nel.org>,
linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
Will Deacon <will@...nel.org>,
Bjorn Helgaas <bhelgaas@...gle.com>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Android Kernel Team <kernel-team@...roid.com>,
linux-efi <linux-efi@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] efi: arm: defer probe of PCIe backed efifb on DT systems
On Thu, Jan 9, 2020 at 6:06 AM Ard Biesheuvel <ard.biesheuvel@...aro.org> wrote:
>
> On Tue, 24 Dec 2019 at 05:41, Saravana Kannan <saravanak@...gle.com> wrote:
> >
> > From: Ard Biesheuvel <ardb@...nel.org>
> >
> > The new of_devlink support breaks PCIe probing on ARM platforms booting
> > via UEFI if the firmware exposes a EFI framebuffer that is backed by a
> > PCI device. The reason is that the probing order gets reversed,
> > resulting in a resource conflict on the framebuffer memory window when
> > the PCIe probes last, causing it to give up entirely.
> >
> > Given that we rely on PCI quirks to deal with EFI framebuffers that get
> > moved around in memory, we cannot simply drop the memory reservation, so
> > instead, let's use the device link infrastructure to register this
> > dependency, and force the probing to occur in the expected order.
> >
> > Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
> > Signed-off-by: Ard Biesheuvel <ardb@...nel.org>
> > Co-developed-by: Saravana Kannan <saravanak@...gle.com>
> > Signed-off-by: Saravana Kannan <saravanak@...gle.com>
> > ---
> >
> > Hi Ard,
> >
> > I compile tested it and I think it should work. If you can actually run
> > and test it, that'd be nice.
> >
> > You can also optimize find_pci_overlap_node() by caching the result if
> > you think that's necessary.
> >
> > Right now this code will run always just like your code did. But once I
> > rename of_devlink to fw_devlink, this code won't be run if fw_devlink is
> > disabled.
> >
> > v1 -> v2:
> > - Rewrote the device linking part to not depend on initcall ordering
> >
> > drivers/firmware/efi/arm-init.c | 106 ++++++++++++++++++++++++++++++--
> > 1 file changed, 102 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/firmware/efi/arm-init.c b/drivers/firmware/efi/arm-init.c
> > index 904fa09e6a6b..8b789ff83af0 100644
> > --- a/drivers/firmware/efi/arm-init.c
> > +++ b/drivers/firmware/efi/arm-init.c
> > @@ -10,10 +10,12 @@
> > #define pr_fmt(fmt) "efi: " fmt
> >
> > #include <linux/efi.h>
> > +#include <linux/fwnode.h>
> > #include <linux/init.h>
> > #include <linux/memblock.h>
> > #include <linux/mm_types.h>
> > #include <linux/of.h>
> > +#include <linux/of_address.h>
> > #include <linux/of_fdt.h>
> > #include <linux/platform_device.h>
> > #include <linux/screen_info.h>
> > @@ -276,15 +278,111 @@ void __init efi_init(void)
> > efi_memmap_unmap();
> > }
> >
> > +static bool efifb_overlaps_pci_range(const struct of_pci_range *range)
> > +{
> > + u64 fb_base = screen_info.lfb_base;
> > +
> > + if (screen_info.capabilities & VIDEO_CAPABILITY_64BIT_BASE)
> > + fb_base |= (u64)(unsigned long)screen_info.ext_lfb_base << 32;
> > +
> > + return fb_base >= range->cpu_addr &&
> > + fb_base < (range->cpu_addr + range->size);
> > +}
> > +
> > +static struct device_node *find_pci_overlap_node(void)
> > +{
> > + struct device_node *np;
> > +
> > + for_each_node_by_type(np, "pci") {
> > + struct of_pci_range_parser parser;
> > + struct of_pci_range range;
> > + int err;
> > +
> > + err = of_pci_range_parser_init(&parser, np);
> > + if (err) {
> > + pr_warn("of_pci_range_parser_init() failed: %d\n", err);
> > + continue;
> > + }
> > +
> > + for_each_of_pci_range(&parser, &range)
> > + if (efifb_overlaps_pci_range(&range))
> > + return np;
> > + }
> > + return NULL;
> > +}
> > +
> > +/*
> > + * If the efifb framebuffer is backed by a PCI graphics controller, we have
> > + * to ensure that this relation is expressed using a device link when
> > + * running in DT mode, or the probe order may be reversed, resulting in a
> > + * resource reservation conflict on the memory window that the efifb
> > + * framebuffer steals from the PCIe host bridge.
> > + */
> > +static int efifb_add_links(const struct fwnode_handle *fwnode,
> > + struct device *dev)
> > +{
> > + struct device_node *sup_np;
> > + struct device *sup_dev;
> > +
> > + sup_np = find_pci_overlap_node();
> > +
> > + /*
> > + * If there's no PCI graphics controller backing the efifb, we are
> > + * done here.
> > + */
> > + if (!sup_np)
> > + return 0;
> > +
> > + sup_dev = get_dev_from_fwnode(&sup_np->fwnode);
> > + of_node_put(sup_np);
> > +
> > + /*
> > + * Return -ENODEV if the PCI graphics controller device hasn't been
> > + * registered yet. This ensures that efifb isn't allowed to probe
> > + * and this function is retried again when new devices are
> > + * registered.
> > + */
> > + if (!sup_dev)
> > + return -ENODEV;
> > +
> > + /*
> > + * If this fails, retrying this function at a later point won't
> > + * change anything. So, don't return an error after this.
> > + */
> > + if (!device_link_add(dev, sup_dev, 0))
> > + dev_warn(dev, "device_link_add() failed\n");
> > +
> > + put_device(sup_dev);
> > +
> > + return 0;
> > +}
> > +
> > +static struct fwnode_operations efifb_fwnode_ops = {
>
> Please make this const
Ack
> > + .add_links = efifb_add_links,
> > +};
> > +
> > +static struct fwnode_handle efifb_fwnode = {
> > + .ops = &efifb_fwnode_ops,
> > +};
> > +
> > static int __init register_gop_device(void)
> > {
> > - void *pd;
> > + struct platform_device *pd;
> > + int err;
> >
> > if (screen_info.orig_video_isVGA != VIDEO_TYPE_EFI)
> > return 0;
> >
> > - pd = platform_device_register_data(NULL, "efi-framebuffer", 0,
> > - &screen_info, sizeof(screen_info));
> > - return PTR_ERR_OR_ZERO(pd);
> > + pd = platform_device_alloc("efi-framebuffer", 0);
> > + if (!pd)
> > + return -ENOMEM;
> > +
>
> Add
>
> if (IS_ENABLED(CONFIG_PCI))
>
> here
As in around the line below where I set the fwnode? Then it's going to
warn about unused variable. Maybe I should just do the if
(IS_ENABLED(CONFIG_PCI)) inside the add_links() function to short
circuit it.
Responding to your other email here. Adding ifdef CONFIG_PCI was my
first inclination, but then those 2 functions are defined if
CONFIG_OF_ADDRESS are defined and if not the stubs are always there.
There's no #ifdef CONFIG_PCI around any of them AFAICT. So I'm
confused about what's going on.
> > + pd->dev.fwnode = &efifb_fwnode;
> > +
> > + err = platform_device_add_data(pd, &screen_info, sizeof(screen_info));
> > + if (err)
> > + return err;
> > +
> > + return platform_device_add(pd);
> > }
> > subsys_initcall(register_gop_device);
> > --
> > 2.24.1.735.g03f4e72817-goog
> >
>
> With the changes above
Yeah, will make them all.
>
> Reviewed-by: Ard Biesheuvel <ardb@...nel.org>
>
Thanks!
-Saravana
Powered by blists - more mailing lists