Re: [PATCH] pci: dwc: pcie_designware: Fix a sleep-in-atomic-context bug in dw_pcie_prog_outbound_atu

[Gustavo Pimentel]

Hi Jia,

On 02/09/2018 04:38, Jia-Ju Bai wrote:
> The driver may sleep with holding a spinlock and in an interupt handler.
> 
> The function call paths (from bottom to top) in Linux-4.16 are:
> 
> [FUNC] usleep_range
> drivers/pci/dwc/pcie-designware.c, 181: 
> 	usleep_range in dw_pcie_prog_outbound_atu
> drivers/pci/dwc/pcie-designware-host.c, 479: 
> 	dw_pcie_prog_outbound_atu in dw_pcie_rd_other_conf
> drivers/pci/dwc/pcie-designware-host.c, 561: 
> 	dw_pcie_rd_other_conf in dw_pcie_rd_conf
> drivers/pci/access.c, 66: 
> 	[FUNC_PTR]dw_pcie_rd_conf in pci_bus_read_config_word
> drivers/pci/access.c, 918: 
> 	pci_bus_read_config_word in pci_read_config_word
> drivers/block/umem.c, 630: 
> 	pci_read_config_word in mm_interrupt (interrupt handler)
> 
> [FUNC] usleep_range
> drivers/pci/dwc/pcie-designware.c, 181: 
> 	usleep_range in dw_pcie_prog_outbound_atu
> drivers/pci/dwc/pcie-designware-host.c, 479: 
> 	dw_pcie_prog_outbound_atu in dw_pcie_rd_other_conf
> drivers/pci/dwc/pcie-designware-host.c, 561: 
> 	dw_pcie_rd_other_conf in dw_pcie_rd_conf
> drivers/pci/access.c, 66: 
> 	[FUNC_PTR]dw_pcie_rd_conf in pci_bus_read_config_word
> drivers/pci/access.c, 918: 
> 	pci_bus_read_config_word in pci_read_config_word
> drivers/ata/pata_efar.c, 115: 
> 	pci_read_config_word in efar_set_piomode
> drivers/ata/pata_efar.c, 113: 
> 	_raw_spin_lock_irqsave in efar_set_piomod
> 
> [FUNC] usleep_range
> drivers/pci/dwc/pcie-designware.c, 181: 
> 	usleep_range in dw_pcie_prog_outbound_atu
> drivers/pci/dwc/pcie-designware-host.c, 479: 
> 	dw_pcie_prog_outbound_atu in dw_pcie_rd_other_conf
> drivers/pci/dwc/pcie-designware-host.c, 561: 
> 	dw_pcie_rd_other_conf in dw_pcie_rd_conf
> drivers/pci/access.c, 66: 
> 	[FUNC_PTR]dw_pcie_rd_conf in pci_bus_read_config_word
> drivers/pci/access.c, 918: 
> 	pci_bus_read_config_word in pci_read_config_word
> drivers/block/mtip32xx/mtip32xx.c, 158: 
> 	pci_read_config_word in mtip_check_surprise_removal
> drivers/block/mtip32xx/mtip32xx.c, 843: 
> 	mtip_check_surprise_removal in mtip_handle_irq
> drivers/block/mtip32xx/mtip32xx.c, 879: 
> 	mtip_handle_irq in mtip_irq_handler (interrupt handler)
> 
> [FUNC] usleep_range
> drivers/pci/dwc/pcie-designware.c, 181: 
> 	usleep_range in dw_pcie_prog_outbound_atu
> drivers/pci/dwc/pcie-designware-host.c, 479: 
> 	dw_pcie_prog_outbound_atu in dw_pcie_rd_other_conf
> drivers/pci/dwc/pcie-designware-host.c, 561: 
> 	dw_pcie_rd_other_conf in dw_pcie_rd_conf
> drivers/pci/access.c, 66: 
> 	[FUNC_PTR]dw_pcie_rd_conf in pci_bus_read_config_word
> drivers/pci/access.c, 918: 
> 	pci_bus_read_config_word in pci_read_config_word
> drivers/gpu/vga/vgaarb.c, 645: 
> 	pci_read_config_word in vga_arbiter_add_pci_device
> drivers/gpu/vga/vgaarb.c, 629: 
> 	_raw_spin_lock_irqsave in vga_arbiter_add_pci_device
> 
> [FUNC] usleep_range
> drivers/pci/dwc/pcie-designware.c, 181: 
> 	usleep_range in dw_pcie_prog_outbound_atu
> drivers/pci/dwc/pcie-designware-host.c, 479: 
> 	dw_pcie_prog_outbound_atu in dw_pcie_rd_other_conf
> drivers/pci/dwc/pcie-designware-host.c, 561: 
> 	dw_pcie_rd_other_conf in dw_pcie_rd_conf
> drivers/pci/access.c, 66: 
> 	[FUNC_PTR]dw_pcie_rd_conf in pci_bus_read_config_word
> drivers/pci/access.c, 918: 
> 	pci_bus_read_config_word in pci_read_config_word
> drivers/pci/ats.c, 139: 
> 	pci_read_config_word in pci_ats_queue_depth
> drivers/iommu/intel-iommu.c, 1519: 
> 	pci_ats_queue_depth in iommu_enable_dev_iotlb
> drivers/iommu/intel-iommu.c, 5295: 
> 	iommu_enable_dev_iotlb in intel_iommu_enable_pasid
> drivers/iommu/intel-iommu.c, 5241: 
> 	_raw_spin_lock_irqsave in intel_iommu_enable_pasid
> 
> To fix this bug, usleep_range() is replaced with udelay().
> 
> This bug is found by my static analysis tool DSAC.
> 
> Signed-off-by: Jia-Ju Bai <baijiaju1990@xxxxxxxxx>
> ---
>  drivers/pci/controller/dwc/pcie-designware.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/controller/dwc/pcie-designware.c b/drivers/pci/controller/dwc/pcie-designware.c
> index 778c4f76a884..7f50f7e51543 100644
> --- a/drivers/pci/controller/dwc/pcie-designware.c
> +++ b/drivers/pci/controller/dwc/pcie-designware.c
> @@ -135,7 +135,7 @@ static void dw_pcie_prog_outbound_atu_unroll(struct dw_pcie *pci, int index,
>  		if (val & PCIE_ATU_ENABLE)
>  			return;
>  
> -		usleep_range(LINK_WAIT_IATU_MIN, LINK_WAIT_IATU_MAX);
> +		udelay(LINK_WAIT_IATU_MAX);

You might use the mdelay instead of udelay, however this fix and others were
already made on patch on [1]

[1] https://lkml.org/lkml/2018/9/10/284

Regards,
Gustavo

>  	}
>  	dev_err(pci->dev, "Outbound iATU is not being enabled\n");
>  }
>