netdev - RE: [PATCH net-next 01/10] net/smc: add missing dev

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <VI1PR0502MB300821C0E577822DC66E1A83D1700@VI1PR0502MB3008.eurprd05.prod.outlook.com>
Date:   Thu, 5 Oct 2017 14:44:52 +0000
From:   Parav Pandit <parav@...lanox.com>
To:     Ursula Braun <ubraun@...ux.vnet.ibm.com>
CC:     "davem@...emloft.net" <davem@...emloft.net>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "linux-rdma@...r.kernel.org" <linux-rdma@...r.kernel.org>,
        "linux-s390@...r.kernel.org" <linux-s390@...r.kernel.org>,
        "jwi@...ux.vnet.ibm.com" <jwi@...ux.vnet.ibm.com>,
        "schwidefsky@...ibm.com" <schwidefsky@...ibm.com>,
        "heiko.carstens@...ibm.com" <heiko.carstens@...ibm.com>,
        "raspl@...ux.vnet.ibm.com" <raspl@...ux.vnet.ibm.com>
Subject: RE: [PATCH net-next 01/10] net/smc: add missing dev_put

Hi Ursula,

> -----Original Message-----
> From: Ursula Braun [mailto:ubraun@...ux.vnet.ibm.com]
> Sent: Thursday, October 05, 2017 2:54 AM
> To: Parav Pandit <parav@...lanox.com>
> Cc: davem@...emloft.net; netdev@...r.kernel.org; linux-
> rdma@...r.kernel.org; linux-s390@...r.kernel.org; jwi@...ux.vnet.ibm.com;
> schwidefsky@...ibm.com; heiko.carstens@...ibm.com;
> raspl@...ux.vnet.ibm.com
> Subject: Re: [PATCH net-next 01/10] net/smc: add missing dev_put
> 
> 
> 
> On 10/02/2017 10:36 PM, Parav Pandit wrote:
> > Hi Ursula, Dave, Hans,
> >
> >> -----Original Message-----
> >> From: linux-rdma-owner@...r.kernel.org [mailto:linux-rdma-
> >> owner@...r.kernel.org] On Behalf Of Ursula Braun
> >> Sent: Wednesday, September 20, 2017 6:58 AM
> >> To: davem@...emloft.net
> >> Cc: netdev@...r.kernel.org; linux-rdma@...r.kernel.org; linux-
> >> s390@...r.kernel.org; jwi@...ux.vnet.ibm.com; schwidefsky@...ibm.com;
> >> heiko.carstens@...ibm.com; raspl@...ux.vnet.ibm.com;
> >> ubraun@...ux.vnet.ibm.com
> >> Subject: [PATCH net-next 01/10] net/smc: add missing dev_put
> >>
> >> From: Hans Wippel <hwippel@...ux.vnet.ibm.com>
> >>
> >> In the infiniband part, SMC currently uses get_netdev which calls
> >> dev_hold on the returned net device. However, the SMC code never
> >> calls dev_put on that net device resulting in a wrong reference count.
> >>
> >> This patch adds a dev_put after the usage of the net device to fix the issue.
> >>
> >> Signed-off-by: Hans Wippel <hwippel@...ux.vnet.ibm.com>
> >> Signed-off-by: Ursula Braun <ubraun@...ux.vnet.ibm.com>
> >> ---
> >>  net/smc/smc_ib.c | 1 +
> >>  1 file changed, 1 insertion(+)
> >>
> >> diff --git a/net/smc/smc_ib.c b/net/smc/smc_ib.c index
> >> 547e0e113b17..0b5852299158 100644
> >> --- a/net/smc/smc_ib.c
> >> +++ b/net/smc/smc_ib.c
> >> @@ -380,6 +380,7 @@ static int smc_ib_fill_gid_and_mac(struct
> >> smc_ib_device *smcibdev, u8 ibport)
> >>  	ndev = smcibdev->ibdev->get_netdev(smcibdev->ibdev, ibport);
> >>  	if (ndev) {
> >>  		memcpy(&smcibdev->mac, ndev->dev_addr, ETH_ALEN);
> >> +		dev_put(ndev);
> >
> > I am sorry for providing late comments. smc_ib_fill_gid_and_mac() is not
> coded correctly.
> > Few fixes are needed.
> > 1. ULP such as SMC should not open code/deference any function pointer like
> get_netdev() of the IB device.
> > 2. Replace ib_query_gid(..., NULL)
> > With
> > ib_query_gid(..., gid_attr);
> >
> > Use gid_attr.ndev to get the MAC address.
> > Do dev_put(gid_attr.ndev);
> >
> > Code should look like below,
> >
> > struct ib_gid_attr gid_attr;
> >
> > rc = ib_query_gid(..., &gid_attr);
> > if (rc || !gid_addr.ndev)
> > 	return -ENODEV;
> > else
> > 	memcpy(smcibdev->mac, ndev->dev_addr, ETH_ALEN);
> > dev_put(gid_addr.ndev);
> >
> 
> Thanks, Parav!
> Following your fix ideas I plan to change the function into this one:
> 
> static int smc_ib_fill_gid_and_mac(struct smc_ib_device *smcibdev, u8 ibport) {
>         struct ib_gid_attr gattr;
>         int rc;
> 
>         rc = ib_query_gid(smcibdev->ibdev, ibport, 0,
>                           &smcibdev->gid[ibport - 1], &gattr);
>         /* the SMC protocol requires specification of the roce MAC address;
>          * if net_device cannot be determined, it can be derived from gid 0
>          */
>         if (rc)
>                 return rc;
> 
>         if (gattr.ndev) {
>                 memcpy(&smcibdev->mac, gattr.ndev->dev_addr, ETH_ALEN);
>                 dev_put(gattr.ndev);
>         } else {
>                 memcpy(&smcibdev->mac[ibport - 1][0],
>                        &smcibdev->gid[ibport - 1].raw[8], 3);
>                 memcpy(&smcibdev->mac[ibport - 1][3],
>                        &smcibdev->gid[ibport - 1].raw[13], 3);
>                 smcibdev->mac[ibport - 1][0] &= ~0x02;
This else part is not needed. You can fail the call as suggested above, inline below too.
if (rc || !gid_addr.ndev)
 	return -ENODEV;
There must be valid netdev for RoCE gid.
Rest code looks fine.