net/af_iucv: fix virtual vs physical address confusion

[PATCH] net/af_iucv: fix virtual vs physical address confusion

Posted by Alexander Gordeev 1 year, 11 months ago

Fix virtual vs physical address confusion. This does not fix a bug
since virtual and physical address spaces are currently the same.

Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
---
 net/iucv/af_iucv.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/net/iucv/af_iucv.c b/net/iucv/af_iucv.c
index 498a0c35b7bb..4aa1c72e6c49 100644
--- a/net/iucv/af_iucv.c
+++ b/net/iucv/af_iucv.c
@@ -1060,13 +1060,12 @@ static int iucv_sock_sendmsg(struct socket *sock, struct msghdr *msg,
 			int i;
 
 			/* skip iucv_array lying in the headroom */
-			iba[0].address = (u32)(addr_t)skb->data;
+			iba[0].address = (u32)virt_to_phys(skb->data);
 			iba[0].length = (u32)skb_headlen(skb);
 			for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
 				skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
 
-				iba[i + 1].address =
-					(u32)(addr_t)skb_frag_address(frag);
+				iba[i + 1].address = (u32)virt_to_phys(skb_frag_address(frag));
 				iba[i + 1].length = (u32)skb_frag_size(frag);
 			}
 			err = pr_iucv->message_send(iucv->path, &txmsg,
@@ -1162,13 +1161,12 @@ static void iucv_process_message(struct sock *sk, struct sk_buff *skb,
 			struct iucv_array *iba = (struct iucv_array *)skb->head;
 			int i;
 
-			iba[0].address = (u32)(addr_t)skb->data;
+			iba[0].address = (u32)virt_to_phys(skb->data);
 			iba[0].length = (u32)skb_headlen(skb);
 			for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
 				skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
 
-				iba[i + 1].address =
-					(u32)(addr_t)skb_frag_address(frag);
+				iba[i + 1].address = (u32)virt_to_phys(skb_frag_address(frag));
 				iba[i + 1].length = (u32)skb_frag_size(frag);
 			}
 			rc = pr_iucv->message_receive(path, msg,
-- 
2.40.1

Re: [PATCH] net/af_iucv: fix virtual vs physical address confusion

Posted by Alexandra Winter 1 year, 11 months ago


On 15.02.24 09:05, Alexander Gordeev wrote:
> Fix virtual vs physical address confusion. This does not fix a bug
> since virtual and physical address spaces are currently the same.
> 
> Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
> ---
>  net/iucv/af_iucv.c | 10 ++++------
>  1 file changed, 4 insertions(+), 6 deletions(-)
> 
> diff --git a/net/iucv/af_iucv.c b/net/iucv/af_iucv.c
> index 498a0c35b7bb..4aa1c72e6c49 100644
> --- a/net/iucv/af_iucv.c
> +++ b/net/iucv/af_iucv.c
> @@ -1060,13 +1060,12 @@ static int iucv_sock_sendmsg(struct socket *sock, struct msghdr *msg,
>  			int i;
>  
>  			/* skip iucv_array lying in the headroom */
> -			iba[0].address = (u32)(addr_t)skb->data;
> +			iba[0].address = (u32)virt_to_phys(skb->data);
>  			iba[0].length = (u32)skb_headlen(skb);
>  			for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
>  				skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
>  
> -				iba[i + 1].address =
> -					(u32)(addr_t)skb_frag_address(frag);
> +				iba[i + 1].address = (u32)virt_to_phys(skb_frag_address(frag));
>  				iba[i + 1].length = (u32)skb_frag_size(frag);
>  			}
>  			err = pr_iucv->message_send(iucv->path, &txmsg,
> @@ -1162,13 +1161,12 @@ static void iucv_process_message(struct sock *sk, struct sk_buff *skb,
>  			struct iucv_array *iba = (struct iucv_array *)skb->head;
>  			int i;
>  
> -			iba[0].address = (u32)(addr_t)skb->data;
> +			iba[0].address = (u32)virt_to_phys(skb->data);
>  			iba[0].length = (u32)skb_headlen(skb);
>  			for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
>  				skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
>  
> -				iba[i + 1].address =
> -					(u32)(addr_t)skb_frag_address(frag);
> +				iba[i + 1].address = (u32)virt_to_phys(skb_frag_address(frag));
>  				iba[i + 1].length = (u32)skb_frag_size(frag);
>  			}
>  			rc = pr_iucv->message_receive(path, msg,


Reviewed-by: Alexandra Winter <wintera@linux.ibm.com>

I would have preferred to do all the translations in __iucv_* functions in iucv.c,
but I understand that for __iucv_message_receive() this would mean significant changes.

Re: [PATCH] net/af_iucv: fix virtual vs physical address confusion

Posted by Jakub Kicinski 1 year, 11 months ago

On Thu, 15 Feb 2024 14:36:57 +0100 Alexandra Winter wrote:
> I would have preferred to do all the translations in __iucv_* functions in iucv.c,
> but I understand that for __iucv_message_receive() this would mean significant changes. 

FWIW we're assuming this is going via the s390 tree.
Please let us know if you prefer networking to pick it up.

Re: [PATCH] net/af_iucv: fix virtual vs physical address confusion

Posted by Alexandra Winter 1 year, 11 months ago


On 21.02.24 23:31, Jakub Kicinski wrote:
> On Thu, 15 Feb 2024 14:36:57 +0100 Alexandra Winter wrote:
>> I would have preferred to do all the translations in __iucv_* functions in iucv.c,
>> but I understand that for __iucv_message_receive() this would mean significant changes. 
> 
> FWIW we're assuming this is going via the s390 tree.
> Please let us know if you prefer networking to pick it up.
> 

Hello Jakub and maintainers,

please take this patch via net-next, as it belongs to the net/iucv module.
Please excuse that it was not flagged with [PATCH net-next]

Kind regards
Alexandra

Re: [PATCH] net/af_iucv: fix virtual vs physical address confusion

Posted by Jakub Kicinski 1 year, 11 months ago

On Thu, 22 Feb 2024 15:41:23 +0100 Alexandra Winter wrote:
> > FWIW we're assuming this is going via the s390 tree.
> > Please let us know if you prefer networking to pick it up.
> 
> Hello Jakub and maintainers,
> 
> please take this patch via net-next, as it belongs to the net/iucv module.
> Please excuse that it was not flagged with [PATCH net-next]

Done! Hopefully nobody screams at us for the use of virt_to_phys() :)