Skip to contentSkip to conversation

iommu: rockchip: Add support for iommu v2

bug report Dan Carpenter 3 participants 3 messages Jun 18, 2021 Arm Rockchip

lore
Share & tools
lore
#0 lore raw

Hello Benjamin Gaignard,

The patch c55356c534aa: "iommu: rockchip: Add support for iommu v2" from Jun 4, 2021, leads to the following static checker warning:

	drivers/iommu/rockchip-iommu.c:552 rk_dte_addr_phys_v2()
	warn: potential shift truncation.  '0xff00000000 (0,4294967296-1095216660480) << 28'

drivers/iommu/rockchip-iommu.c
544 }
545
546 #define DT_HI_MASK GENMASK_ULL(39, 32)

                           ^^^^^^^^^^^^^^^^^^^

547 #define DT_SHIFT 28
548
549 static inline phys_addr_t rk_dte_addr_phys_v2(u32 addr) 550 {
551 return (phys_addr_t)(addr & RK_DTE_PT_ADDRESS_MASK) |

   552                 ((addr & DT_HI_MASK) << DT_SHIFT);
                         ^^^^^^^^^^^^^^^^^

Is addr supposed to be a u64?

553 }

regards,
dan carpenter

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#1 lore raw

Le 18/06/2021 à 11:15, Dan Carpenter a écrit :

Quoted 21 lines

Hello Benjamin Gaignard,

The patch c55356c534aa: "iommu: rockchip: Add support for iommu v2" from Jun 4, 2021, leads to the following static checker warning:

	drivers/iommu/rockchip-iommu.c:552 rk_dte_addr_phys_v2()
	warn: potential shift truncation.  '0xff00000000 (0,4294967296-1095216660480) << 28'

drivers/iommu/rockchip-iommu.c

    544  }
    545
    546  #define DT_HI_MASK GENMASK_ULL(39, 32)
                            ^^^^^^^^^^^^^^^^^^^

    547  #define DT_SHIFT   28
    548
    549  static inline phys_addr_t rk_dte_addr_phys_v2(u32 addr)
    550  {
    551          return (phys_addr_t)(addr & RK_DTE_PT_ADDRESS_MASK) |
    552                 ((addr & DT_HI_MASK) << DT_SHIFT);
                          ^^^^^^^^^^^^^^^^^
Quoted 1 line

Is addr supposed to be a u64?

The error is to use DT_HI_MASK here.
It should be:
#define DTE_BASE_HI_MASK GENMASK(11, 4)

static inline phys_addr_t rk_dte_addr_phys_v2(u32 addr)

{
	return (phys_addr_t)(addr & RK_DTE_PT_ADDRESS_MASK) |
	       ((addr & DTE_BASE_HI_MASK) << DT_SHIFT);
}

I will send a patch.
Thanks for the finding.

Regards,
Benjamin

Quoted 5 lines
    553  }

regards,
dan carpenter

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#2 lore raw
In reply to Dan Carpenter
Quoted 21 lines · Dan Carpenter

Hello Benjamin Gaignard,

The patch c55356c534aa: "iommu: rockchip: Add support for iommu v2" from Jun 4, 2021, leads to the following static checker warning:

	drivers/iommu/rockchip-iommu.c:552 rk_dte_addr_phys_v2()
	warn: potential shift truncation.  '0xff00000000 (0,4294967296-1095216660480) << 28'

drivers/iommu/rockchip-iommu.c

    544  }
    545
    546  #define DT_HI_MASK GENMASK_ULL(39, 32)
                            ^^^^^^^^^^^^^^^^^^^

    547  #define DT_SHIFT   28
    548
    549  static inline phys_addr_t rk_dte_addr_phys_v2(u32 addr)
    550  {
    551          return (phys_addr_t)(addr & RK_DTE_PT_ADDRESS_MASK) |
    552                 ((addr & DT_HI_MASK) << DT_SHIFT);
                          ^^^^^^^^^^^^^^^^^
Quoted 1 line

Is addr supposed to be a u64?

Hmm, the cast is in the wrong place to be useful, but the shifting and masking to unpack the high address bits also looks backwards anyway. I think the whole expression should probably be something more like:

	(((phys_addr_t)addr << DT_SHIFT) & DT_HI_MASK) |
	(addr & RK_DTE_PT_ADDRESS_MASK);

Robin.

Quoted 11 lines
    553  }

regards,
dan carpenter

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip