| 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: <CAL_JsqJV_8sCVB2fAi6kk19ZLbO+nKbk-kYsBNEbN+jR84LUgg@mail.gmail.com>
Date: Wed, 23 Sep 2020 10:41:45 -0600
From: Rob Herring <robh@...nel.org>
To: Jisheng Zhang <Jisheng.Zhang@...aptics.com>
Cc: Kishon Vijay Abraham I <kishon@...com>,
Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
Bjorn Helgaas <bhelgaas@...gle.com>,
Jingoo Han <jingoohan1@...il.com>,
Gustavo Pimentel <gustavo.pimentel@...opsys.com>,
Thierry Reding <treding@...dia.com>,
Vidya Sagar <vidyas@...dia.com>,
linux-omap <linux-omap@...r.kernel.org>,
PCI <linux-pci@...r.kernel.org>,
linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] PCI: dwc: Move allocate and map page for msi out of dw_pcie_msi_init()
On Wed, Sep 23, 2020 at 12:27 AM Jisheng Zhang
<Jisheng.Zhang@...aptics.com> wrote:
>
> Currently, dw_pcie_msi_init() allocates and maps page for msi, then
> program the PCIE_MSI_ADDR_LO and PCIE_MSI_ADDR_HI. The Root Complex
> may lose power during suspend-to-RAM, so when we resume, we want to
> redo the latter but not the former. If designware based driver (for
> example, pcie-tegra194.c) calls dw_pcie_msi_init() in resume path, the
> previous msi page will be leaked.
It's worse than this. I think there's also error paths too leaking the
page. Also, there's never a dma_unmap_page call which should happen
before freeing.
> Move the allocate and map msi page from dw_pcie_msi_init() to
> dw_pcie_host_init() to fix this problem.
>
> Fixes: 56e15a238d92 ("PCI: tegra: Add Tegra194 PCIe support")
> Signed-off-by: Jisheng Zhang <Jisheng.Zhang@...aptics.com>
> ---
> drivers/pci/controller/dwc/pci-dra7xx.c | 18 ++++++++++++-
> .../pci/controller/dwc/pcie-designware-host.c | 27 +++++++++----------
> 2 files changed, 30 insertions(+), 15 deletions(-)
>
> diff --git a/drivers/pci/controller/dwc/pci-dra7xx.c b/drivers/pci/controller/dwc/pci-dra7xx.c
> index dc387724cf08..4301cf844a4c 100644
> --- a/drivers/pci/controller/dwc/pci-dra7xx.c
> +++ b/drivers/pci/controller/dwc/pci-dra7xx.c
> @@ -490,7 +490,9 @@ static struct irq_chip dra7xx_pci_msi_bottom_irq_chip = {
> static int dra7xx_pcie_msi_host_init(struct pcie_port *pp)
> {
> struct dw_pcie *pci = to_dw_pcie_from_pp(pp);
> + struct device *dev = pci->dev;
> u32 ctrl, num_ctrls;
> + int ret;
>
> pp->msi_irq_chip = &dra7xx_pci_msi_bottom_irq_chip;
>
> @@ -506,7 +508,21 @@ static int dra7xx_pcie_msi_host_init(struct pcie_port *pp)
> ~0);
> }
>
> - return dw_pcie_allocate_domains(pp);
> + ret = dw_pcie_allocate_domains(pp);
> + if (ret)
> + return ret;
> +
> + pp->msi_page = alloc_page(GFP_KERNEL);
> + pp->msi_data = dma_map_page(dev, pp->msi_page, 0, PAGE_SIZE,
> + DMA_FROM_DEVICE);
> + ret = dma_mapping_error(dev, pp->msi_data);
> + if (ret) {
> + dev_err(dev, "Failed to map MSI data\n");
> + __free_page(pp->msi_page);
> + pp->msi_page = NULL;
> + dw_pcie_free_msi(pp);
> + }
I don't like having 2 copies of the same thing. Also, doesn't keystone
need this too?
The other thing is .msi_host_init() is abused by having an empty
function to disable MSI support. We should have a flag instead to
enable/disable MSI support and then we can key off of that in the
common code.
> + return ret;
> }
>
> static const struct dw_pcie_host_ops dra7xx_pcie_host_ops = {
> diff --git a/drivers/pci/controller/dwc/pcie-designware-host.c b/drivers/pci/controller/dwc/pcie-designware-host.c
> index 9dafecba347f..c23ba64f64fe 100644
> --- a/drivers/pci/controller/dwc/pcie-designware-host.c
> +++ b/drivers/pci/controller/dwc/pcie-designware-host.c
> @@ -294,20 +294,7 @@ void dw_pcie_free_msi(struct pcie_port *pp)
>
> void dw_pcie_msi_init(struct pcie_port *pp)
Might be good to rename this function with exactly what it does.
There's too many 'init' and 'setup' functions...
> {
> - struct dw_pcie *pci = to_dw_pcie_from_pp(pp);
> - struct device *dev = pci->dev;
> - u64 msi_target;
> -
> - pp->msi_page = alloc_page(GFP_KERNEL);
> - pp->msi_data = dma_map_page(dev, pp->msi_page, 0, PAGE_SIZE,
> - DMA_FROM_DEVICE);
> - if (dma_mapping_error(dev, pp->msi_data)) {
> - dev_err(dev, "Failed to map MSI data\n");
> - __free_page(pp->msi_page);
> - pp->msi_page = NULL;
> - return;
> - }
> - msi_target = (u64)pp->msi_data;
> + u64 msi_target = (u64)pp->msi_data;
>
> /* Program the msi_data */
> dw_pcie_wr_own_conf(pp, PCIE_MSI_ADDR_LO, 4,
> @@ -440,6 +427,18 @@ int dw_pcie_host_init(struct pcie_port *pp)
> irq_set_chained_handler_and_data(pp->msi_irq,
> dw_chained_msi_isr,
> pp);
> +
> + pp->msi_page = alloc_page(GFP_KERNEL);
> + pp->msi_data = dma_map_page(pci->dev, pp->msi_page,
> + 0, PAGE_SIZE,
> + DMA_FROM_DEVICE);
> + ret = dma_mapping_error(pci->dev, pp->msi_data);
> + if (ret) {
> + dev_err(pci->dev, "Failed to map MSI data\n");
> + __free_page(pp->msi_page);
> + pp->msi_page = NULL;
> + goto err_free_msi;
> + }
> } else {
> ret = pp->ops->msi_host_init(pp);
> if (ret < 0)
> --
> 2.28.0
>
Powered by blists - more mailing lists