linux-kernel - Re: [PATCH for-next v3 2/2] RDMA/rxe: Call rxe_set_mtu after rxe_register

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

Date:   Tue, 19 Sep 2023 09:11:43 +0800
From:   Zhu Yanjun <zyjzyj2000@...il.com>
To:     "Zhijian Li (Fujitsu)" <lizhijian@...itsu.com>
Cc:     Leon Romanovsky <leon@...nel.org>,
        "linux-rdma@...r.kernel.org" <linux-rdma@...r.kernel.org>,
        "jgg@...pe.ca" <jgg@...pe.ca>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "rpearsonhpe@...il.com" <rpearsonhpe@...il.com>,
        "Daisuke Matsuda (Fujitsu)" <matsuda-daisuke@...itsu.com>,
        Zhu Yanjun <yanjun.zhu@...ux.dev>
Subject: Re: [PATCH for-next v3 2/2] RDMA/rxe: Call rxe_set_mtu after rxe_register_device

On Tue, Sep 19, 2023 at 8:57 AM Zhijian Li (Fujitsu)
<lizhijian@...itsu.com> wrote:
>
>
>
> On 18/09/2023 20:37, Leon Romanovsky wrote:
> > On Mon, Sep 18, 2023 at 10:05:43AM +0800, Li Zhijian wrote:
> >> rxe_set_mtu() will call rxe_info_dev() to print message, and
> >> rxe_info_dev() expects dev_name(rxe->ib_dev->dev) has been assigned.
> >>
> >> Previously since dev_name() is not set, when a new rxe link is being
> >> added, 'null' will be used as the dev_name like:
> >>
> >> "(null): rxe_set_mtu: Set mtu to 1024"
> >>
> >> Move rxe_register_device() earlier to assign the correct dev_name
> >> so that it can be read by rxe_set_mtu() later.
> >
> > I would expect removal of that print line instead of moving
> > rxe_register_device().
>
>
> I also struggled with this point. The last option is keep it as it is.
> Once rxe is registered, this print will work fine.

I delved into the source code. About moving rxe_register_device, I
could not find any harm to the driver.
So I think this is also a solution to this problem.

Zhu Yanjun

>
> Thanks
> Zhijian
>
>
> >
> > Thanks
> >
> >>
> >> And it's safe to do such change since mtu will not be used during the
> >> rxe_register_device()
> >>
> >> After this change, the message becomes:
> >> "rxe_eth0: rxe_set_mtu: Set mtu to 4096"
> >>
> >> Fixes: 9ac01f434a1e ("RDMA/rxe: Extend dbg log messages to err and info")
> >> Reviewed-by: Daisuke Matsuda <matsuda-daisuke@...itsu.com>
> >> Reviewed-by: Zhu Yanjun <yanjun.zhu@...ux.dev>
> >> Signed-off-by: Li Zhijian <lizhijian@...itsu.com>
> >> ---
> >>   drivers/infiniband/sw/rxe/rxe.c | 5 ++++-
> >>   1 file changed, 4 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/infiniband/sw/rxe/rxe.c b/drivers/infiniband/sw/rxe/rxe.c
> >> index a086d588e159..8a43c0c4f8d8 100644
> >> --- a/drivers/infiniband/sw/rxe/rxe.c
> >> +++ b/drivers/infiniband/sw/rxe/rxe.c
> >> @@ -169,10 +169,13 @@ void rxe_set_mtu(struct rxe_dev *rxe, unsigned int ndev_mtu)
> >>    */
> >>   int rxe_add(struct rxe_dev *rxe, unsigned int mtu, const char *ibdev_name)
> >>   {
> >> +    int ret;
> >> +
> >>      rxe_init(rxe);
> >> +    ret = rxe_register_device(rxe, ibdev_name);
> >>      rxe_set_mtu(rxe, mtu);
> >>
> >> -    return rxe_register_device(rxe, ibdev_name);
> >> +    return ret;
> >>   }
> >>
> >>   static int rxe_newlink(const char *ibdev_name, struct net_device *ndev)
> >> --
> >> 2.29.2
> >>