Skip to contentSkip to conversation

RDMA/siw: fix pointer cast warning

PATCH In review Arnd Bergmann 4 participants 8 messages Dec 15, 2022 – Dec 16, 2022

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 RDMA/siw: fix pointer cast warning

From: Arnd Bergmann <[email protected]>

The previous build fix left a remaining issue in configurations with 64-bit dma_addr_t on 32-bit architectures:

drivers/infiniband/sw/siw/siw_qp_tx.c: In function 'siw_get_pblpage':
drivers/infiniband/sw/siw/siw_qp_tx.c:32:37: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
32 | return virt_to_page((void *)paddr);
| ^

Use the same double cast here that the driver uses elsewhere to convert between dma_addr_t and void*.

It took me a while to figure out why this driver does it like this, as there is no hardware access and it just stores kernel pointers in place of device addresses when communicating with the rdma core and with user space.

Fixes 0d1b756acf60 ("RDMA/siw: Pass a pointer to virt_to_page()")
---
 drivers/infiniband/sw/siw/siw_qp_tx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/sw/siw/siw_qp_tx.c +1 -1
index 7d47b521070b..05052b49107f 100644
@@ -29,7 +29,7 @@ static struct page *siw_get_pblpage(struct siw_mem *mem, u64 addr, int *idx)
 	dma_addr_t paddr = siw_pbl_get_buffer(pbl, offset, NULL, idx);
 
 	if (paddr)
-		return virt_to_page((void *)paddr);
+		return virt_to_page((void *)(uintptr_t)paddr);
 
 	return NULL;
 }
Signature · 1 line
2.35.1
Acked-by on v1 1/1
#1 lore raw
Quoted 50 lines

-----Original Message-----
From: Arnd Bergmann <[email protected]>
Sent: Thursday, 15 December 2022 18:04
To: Bernard Metzler <[email protected]>; Jason Gunthorpe <[email protected]>; Leon Romanovsky <[email protected]>; Linus Walleij <[email protected]>
Cc: Arnd Bergmann <[email protected]>; [email protected]; linux- [email protected]
Subject: [EXTERNAL] [PATCH] RDMA/siw: fix pointer cast warning

From: Arnd Bergmann <[email protected]>

The previous build fix left a remaining issue in configurations with 64-bit dma_addr_t on 32-bit architectures:

drivers/infiniband/sw/siw/siw_qp_tx.c: In function 'siw_get_pblpage':
drivers/infiniband/sw/siw/siw_qp_tx.c:32:37: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
32 | return virt_to_page((void *)paddr);
| ^

Use the same double cast here that the driver uses elsewhere to convert between dma_addr_t and void*.

It took me a while to figure out why this driver does it like this, as there is no hardware access and it just stores kernel pointers in place of device addresses when communicating with the rdma core and with user space.

Fixes 0d1b756acf60 ("RDMA/siw: Pass a pointer to virt_to_page()")
---
 drivers/infiniband/sw/siw/siw_qp_tx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/sw/siw/siw_qp_tx.c +1 -1
index 7d47b521070b..05052b49107f 100644
@@ -29,7 +29,7 @@ static struct page *siw_get_pblpage(struct siw_mem *mem,
 u64 addr, int *idx)
 	dma_addr_t paddr = siw_pbl_get_buffer(pbl, offset, NULL, idx);
 
 	if (paddr)
-		return virt_to_page((void *)paddr);
+		return virt_to_page((void *)(uintptr_t)paddr);
 
 	return NULL;
Quoted 1 line

}

Thanks Arnd, makes complete sense.

#2 lore raw
In reply to Arnd Bergmann

From: Arnd Bergmann

Quoted 13 lines

Sent: 15 December 2022 17:04

From: Arnd Bergmann <[email protected]>

The previous build fix left a remaining issue in configurations with 64-bit dma_addr_t on 32-bit architectures:

drivers/infiniband/sw/siw/siw_qp_tx.c: In function 'siw_get_pblpage':
drivers/infiniband/sw/siw/siw_qp_tx.c:32:37: error: cast to pointer from integer of different size [- Werror=int-to-pointer-cast]
32 | return virt_to_page((void *)paddr);
| ^

Quoted 7 lines

Use the same double cast here that the driver uses elsewhere to convert between dma_addr_t and void*.

It took me a while to figure out why this driver does it like this, as there is no hardware access and it just stores kernel pointers in place of device addresses when communicating with the rdma core and with user space.

I hope that doesn't mean it is relying on user space only giving it back valid values?

David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)

#3 lore raw
Quoted 17 lines · David Laight

From: Arnd Bergmann

Quoted 16 lines

Sent: 15 December 2022 17:04

From: Arnd Bergmann <[email protected]>

The previous build fix left a remaining issue in configurations with 64-bit dma_addr_t on 32-bit architectures:

drivers/infiniband/sw/siw/siw_qp_tx.c: In function 'siw_get_pblpage':
drivers/infiniband/sw/siw/siw_qp_tx.c:32:37: error: cast to pointer from integer of different size [- Werror=int-to-pointer-cast]
32 | return virt_to_page((void *)paddr);
| ^

Use the same double cast here that the driver uses elsewhere to convert between dma_addr_t and void*.

Quoted 7 lines
Quoted 4 lines

It took me a while to figure out why this driver does it like this, as there is no hardware access and it just stores kernel pointers in place of device addresses when communicating with the rdma core and with user space.

I hope that doesn't mean it is relying on user space only giving it back valid values?

It looks to me like this driver totally trusts userspace.

Yours,
Linus Walleij

#4 lore raw
Quoted 38 lines

-----Original Message-----
From: Linus Walleij <[email protected]>
Sent: Friday, 16 December 2022 08:47
To: David Laight <[email protected]>
Cc: Arnd Bergmann <[email protected]>; Bernard Metzler <[email protected]>; Jason Gunthorpe <[email protected]>; Leon Romanovsky <[email protected]>; Arnd Bergmann <[email protected]>; [email protected]; linux- [email protected]
Subject: [EXTERNAL] Re: [PATCH] RDMA/siw: fix pointer cast warning

Quoted 10 lines · David Laight

From: Arnd Bergmann

Quoted 9 lines

Sent: 15 December 2022 17:04

From: Arnd Bergmann <[email protected]>

The previous build fix left a remaining issue in configurations with 64-bit dma_addr_t on 32-bit architectures:

drivers/infiniband/sw/siw/siw_qp_tx.c: In function 'siw_get_pblpage':
drivers/infiniband/sw/siw/siw_qp_tx.c:32:37: error: cast to pointer

from integer of different size [-

Quoted 12 lines
Quoted 11 lines

Werror=int-to-pointer-cast]
32 | return virt_to_page((void *)paddr);
| ^

Use the same double cast here that the driver uses elsewhere to convert between dma_addr_t and void*.

It took me a while to figure out why this driver does it like this, as there is no hardware access and it just stores kernel pointers in place of device addresses when communicating with the rdma core and with user space.

Quoted 5 lines
Quoted 2 lines

I hope that doesn't mean it is relying on user space only giving it back valid values?

It looks to me like this driver totally trusts userspace.

Shame on me. Yes, somehow, an access_ok((void __user *)start, len) is missing! Let me fix that when I am back at my desk. Seems it needs immediate action.

Many thanks for pointing that out!
Bernard.

Quoted 2 lines

Yours,
Linus Walleij

#5 lore raw

From: Bernard Metzler

Quoted 41 lines

Sent: 16 December 2022 10:01

Quoted 39 lines

-----Original Message-----
From: Linus Walleij <[email protected]>
Sent: Friday, 16 December 2022 08:47
To: David Laight <[email protected]>
Cc: Arnd Bergmann <[email protected]>; Bernard Metzler <[email protected]>; Jason Gunthorpe <[email protected]>; Leon Romanovsky <[email protected]>; Arnd Bergmann <[email protected]>; [email protected]; linux- [email protected]
Subject: [EXTERNAL] Re: [PATCH] RDMA/siw: fix pointer cast warning

Quoted 10 lines · David Laight

From: Arnd Bergmann

Quoted 9 lines

Sent: 15 December 2022 17:04

From: Arnd Bergmann <[email protected]>

The previous build fix left a remaining issue in configurations with 64-bit dma_addr_t on 32-bit architectures:

drivers/infiniband/sw/siw/siw_qp_tx.c: In function 'siw_get_pblpage':
drivers/infiniband/sw/siw/siw_qp_tx.c:32:37: error: cast to pointer

from integer of different size [-

Quoted 14 lines
Quoted 11 lines

Werror=int-to-pointer-cast]
32 | return virt_to_page((void *)paddr);
| ^

Use the same double cast here that the driver uses elsewhere to convert between dma_addr_t and void*.

It took me a while to figure out why this driver does it like this, as there is no hardware access and it just stores kernel pointers in place of device addresses when communicating with the rdma core and with user space.

I hope that doesn't mean it is relying on user space only giving it back valid values?

Quoted 6 lines
Quoted 2 lines

It looks to me like this driver totally trusts userspace.

Shame on me. Yes, somehow, an access_ok((void __user *)start, len) is missing! Let me fix that when I am back at my desk. Seems it needs immediate action.

That wasn't the sort of issue I was thinking about.
I was worried that it was putting the addresses of kernel memory into buffers written to userspace and then later reading the addresses back from userspace and accessing them.

David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)

#6 lore raw
Quoted 59 lines

-----Original Message-----
From: David Laight <[email protected]>
Sent: Friday, 16 December 2022 11:23
To: Bernard Metzler <[email protected]>; Linus Walleij <[email protected]>
Cc: Arnd Bergmann <[email protected]>; Jason Gunthorpe <[email protected]>; Leon Romanovsky <[email protected]>; Arnd Bergmann <[email protected]>; linux- [email protected]; [email protected]
Subject: [EXTERNAL] RE: Re: [PATCH] RDMA/siw: fix pointer cast warning

From: Bernard Metzler

Quoted 7 lines

Sent: 16 December 2022 10:01

Quoted 5 lines

-----Original Message-----
From: Linus Walleij <[email protected]>
Sent: Friday, 16 December 2022 08:47
To: David Laight <[email protected]>
Cc: Arnd Bergmann <[email protected]>; Bernard Metzler

<[email protected]>;

Quoted 9 lines
Quoted 9 lines

Jason Gunthorpe <[email protected]>; Leon Romanovsky <[email protected]>; Arnd Bergmann <[email protected]>; [email protected]; linux- [email protected]
Subject: [EXTERNAL] Re: [PATCH] RDMA/siw: fix pointer cast warning

Quoted 1 line · David Laight

From: Arnd Bergmann

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

Sent: 15 December 2022 17:04

From: Arnd Bergmann <[email protected]>

The previous build fix left a remaining issue in configurations with 64-bit dma_addr_t on 32-bit architectures:

drivers/infiniband/sw/siw/siw_qp_tx.c: In function

'siw_get_pblpage':

Quoted 20 lines
Quoted 19 lines
Quoted 1 line
Quoted 1 line

drivers/infiniband/sw/siw/siw_qp_tx.c:32:37: error: cast to pointer

from integer of different size [-

Quoted 7 lines
Quoted 7 lines

Werror=int-to-pointer-cast]
32 | return virt_to_page((void *)paddr);
| ^

Use the same double cast here that the driver uses elsewhere to convert between dma_addr_t and void*.

Quoted 7 lines
Quoted 4 lines

It took me a while to figure out why this driver does it like this, as there is no hardware access and it just stores kernel pointers in place of device addresses when communicating with the rdma core and with user space.

I hope that doesn't mean it is relying on user space only giving it back valid values?

It looks to me like this driver totally trusts userspace.

Quoted 9 lines
Quoted 3 lines

Shame on me. Yes, somehow, an access_ok((void __user *)start, len) is missing! Let me fix that when I am back at my desk. Seems it needs immediate action.

That wasn't the sort of issue I was thinking about.
I was worried that it was putting the addresses of kernel memory into buffers written to userspace and then later reading the addresses back from userspace and accessing them.

Oh, no, that is not the case. The address mapping is not accessible from userspace. It is local to the kernel driver only to translate user virtual addresses to kernel pages during transmit/receive of application data. The user only knows about its own VA's it uses in its work requests, and during buffer registration.

BUT, you pointed me to something bad. Checking the users permission to register requested memory with the driver is definitely missing. I was under the wrong impression it would be checked by used pin_user_pages(), but that is not the case. pin_user_pages_fast() does that check, other drivers using it, and it looks like a good fix.

Other rdma drivers, like infiniband/hw/usnic, wich also use pin_user_pages(), may suffer from the same problem. There is no checking of user permissions for the memory being registered from user land.

Thanks very much,
Bernard.

Quoted 6 lines

David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)

Reviewed-by on v1 1/1
#7 lore raw
In reply to Arnd Bergmann
Quoted 13 lines · Arnd Bergmann

From: Arnd Bergmann <[email protected]>

The previous build fix left a remaining issue in configurations with 64-bit dma_addr_t on 32-bit architectures:

drivers/infiniband/sw/siw/siw_qp_tx.c: In function 'siw_get_pblpage':
drivers/infiniband/sw/siw/siw_qp_tx.c:32:37: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
32 | return virt_to_page((void *)paddr);
| ^

Use the same double cast here that the driver uses elsewhere to convert between dma_addr_t and void*.

Quoted 7 lines

It took me a while to figure out why this driver does it like this, as there is no hardware access and it just stores kernel pointers in place of device addresses when communicating with the rdma core and with user space.

Fixes 0d1b756acf60 ("RDMA/siw: Pass a pointer to virt_to_page()")

This driver is a big confusion for me too.

Yours,
Linus Walleij