From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by dpdk.org (Postfix) with ESMTP id 3D7D01B206; Fri, 12 Oct 2018 10:57:16 +0200 (CEST) Received: from smtp.corp.redhat.com (int-mx01.intmail.prod.int.phx2.redhat.com [10.5.11.11]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 7CB647E9D6; Fri, 12 Oct 2018 08:57:15 +0000 (UTC) Received: from [10.36.112.48] (ovpn-112-48.ams2.redhat.com [10.36.112.48]) by smtp.corp.redhat.com (Postfix) with ESMTPS id 23FDF79597; Fri, 12 Oct 2018 08:57:07 +0000 (UTC) From: Maxime Coquelin To: Ilya Maximets , dev@dpdk.org, tiwei.bie@intel.com, zhihong.wang@intel.com, jfreimann@redhat.com, nicknickolaev@gmail.com, bruce.richardson@intel.com, alejandro.lucero@netronome.com, dgilbert@redhat.com Cc: stable@dpdk.org References: <20181011092432.22275-1-maxime.coquelin@redhat.com> <20181011092432.22275-8-maxime.coquelin@redhat.com> <20181011155709eucas1p20da3d82e3f74a0f8b8bd06eae4baa0c5~cmQddRjzg2527125271eucas1p2_@eucas1p2.samsung.com> <377e0ee1-562d-e5cb-e411-e041f8c60ba5@redhat.com> <7df5b8ea-6570-e2ab-1983-e74d316e530d@redhat.com> Message-ID: <72d63d93-ba13-d228-d67a-0bbdbebeceaf@redhat.com> Date: Fri, 12 Oct 2018 10:57:05 +0200 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1 MIME-Version: 1.0 In-Reply-To: <7df5b8ea-6570-e2ab-1983-e74d316e530d@redhat.com> Content-Type: text/plain; charset=utf-8; format=flowed Content-Language: en-US Content-Transfer-Encoding: 8bit X-Scanned-By: MIMEDefang 2.79 on 10.5.11.11 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.26]); Fri, 12 Oct 2018 08:57:15 +0000 (UTC) Subject: Re: [dpdk-dev] [PATCH v6 07/19] vhost: add number of fds to vhost-user messages and use it X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 12 Oct 2018 08:57:16 -0000 On 10/12/2018 10:45 AM, Maxime Coquelin wrote: > > > On 10/12/2018 10:43 AM, Maxime Coquelin wrote: >> >> >> On 10/11/2018 05:59 PM, Ilya Maximets wrote: >>> On 11.10.2018 12:24, Maxime Coquelin wrote: >>>> As soon as some ancillary data (fds) are received, it is copied >>>> without checking its length. >>>> >>>> This patch adds the number of fds received to the message, >>>> which is set in read_vhost_message(). >>>> >>>> This is preliminary work to support sending fds to Qemu. >>>> >>>> Signed-off-by: Dr. David Alan Gilbert >>>> Signed-off-by: Maxime Coquelin >>>> --- >>>>   lib/librte_vhost/socket.c     | 25 ++++++++++++++++++++----- >>>>   lib/librte_vhost/vhost_user.c |  2 +- >>>>   lib/librte_vhost/vhost_user.h |  4 +++- >>>>   3 files changed, 24 insertions(+), 7 deletions(-) >>>> >>>> diff --git a/lib/librte_vhost/socket.c b/lib/librte_vhost/socket.c >>>> index d63031747..3b0287a26 100644 >>>> --- a/lib/librte_vhost/socket.c >>>> +++ b/lib/librte_vhost/socket.c >>>> @@ -94,18 +94,24 @@ static struct vhost_user vhost_user = { >>>>       .mutex = PTHREAD_MUTEX_INITIALIZER, >>>>   }; >>>> -/* return bytes# of read on success or negative val on failure. */ >>>> +/* >>>> + * return bytes# of read on success or negative val on failure. >>>> Update fdnum >>>> + * with number of fds read. >>>> + */ >>>>   int >>>> -read_fd_message(int sockfd, char *buf, int buflen, int *fds, int >>>> fd_num) >>>> +read_fd_message(int sockfd, char *buf, int buflen, int *fds, int >>>> max_fds, >>>> +        int *fd_num) >>>>   { >>>>       struct iovec iov; >>>>       struct msghdr msgh; >>>> -    size_t fdsize = fd_num * sizeof(int); >>>> -    char control[CMSG_SPACE(fdsize)]; >>>> +    char control[CMSG_SPACE(max_fds * sizeof(int))]; >>>>       struct cmsghdr *cmsg; >>>>       int got_fds = 0; >>>> +    int *tmp_fds; >>>>       int ret; >>>> +    *fd_num = 0; >>>> + >>>>       memset(&msgh, 0, sizeof(msgh)); >>>>       iov.iov_base = buf; >>>>       iov.iov_len  = buflen; >>>> @@ -131,13 +137,22 @@ read_fd_message(int sockfd, char *buf, int >>>> buflen, int *fds, int fd_num) >>>>           if ((cmsg->cmsg_level == SOL_SOCKET) && >>>>               (cmsg->cmsg_type == SCM_RIGHTS)) { >>>>               got_fds = (cmsg->cmsg_len - CMSG_LEN(0)) / sizeof(int); >>>> +            if (got_fds > max_fds) { >>> >>> Hmm. I just noticed that 'msg_controllen' is set to receive >>> not more than max_fds descriptors. So, this case should not >>> be possible. We will receive MSG_CTRUNC and return before >>> the loop. >> >> Maybe it is better to remove check for MSG_CTRUNC. > s/remove/rework/ > >> IIUC, if MSG_CTRUNC happens, we may have to close anyway the ones >> received. >> >> Do you agree? So it seems that other use of MSG_CTRUNC in DPDK and QEMU does care to close the ones that would have been received and just return an error. I propose to do the same for now, an remove the got_fds > max_fds part. >>> +                RTE_LOG(ERR, VHOST_CONFIG, >>>> +                    "Received msg contains more fds than >>>> supported\n"); >>>> +                tmp_fds = (int *)CMSG_DATA(cmsg); >>>> +                while (got_fds--) >>>> +                    close(tmp_fds[got_fds]); >>>> +                return -1; >>>> +            } >>>> +            *fd_num = got_fds; >>>>               memcpy(fds, CMSG_DATA(cmsg), got_fds * sizeof(int)); >>>>               break; >>>>           } >>>>       } >>>>       /* Clear out unused file descriptors */ >>>> -    while (got_fds < fd_num) >>>> +    while (got_fds < max_fds) >>>>           fds[got_fds++] = -1; >>>>       return ret; >>>> diff --git a/lib/librte_vhost/vhost_user.c >>>> b/lib/librte_vhost/vhost_user.c >>>> index 83d3e6321..c1c5f35ff 100644 >>>> --- a/lib/librte_vhost/vhost_user.c >>>> +++ b/lib/librte_vhost/vhost_user.c >>>> @@ -1509,7 +1509,7 @@ read_vhost_message(int sockfd, struct >>>> VhostUserMsg *msg) >>>>       int ret; >>>>       ret = read_fd_message(sockfd, (char *)msg, VHOST_USER_HDR_SIZE, >>>> -        msg->fds, VHOST_MEMORY_MAX_NREGIONS); >>>> +        msg->fds, VHOST_MEMORY_MAX_NREGIONS, &msg->fd_num); >>>>       if (ret <= 0) >>>>           return ret; >>>> diff --git a/lib/librte_vhost/vhost_user.h >>>> b/lib/librte_vhost/vhost_user.h >>>> index 62654f736..9a91d496b 100644 >>>> --- a/lib/librte_vhost/vhost_user.h >>>> +++ b/lib/librte_vhost/vhost_user.h >>>> @@ -132,6 +132,7 @@ typedef struct VhostUserMsg { >>>>           VhostUserVringArea area; >>>>       } payload; >>>>       int fds[VHOST_MEMORY_MAX_NREGIONS]; >>>> +    int fd_num; >>>>   } __attribute((packed)) VhostUserMsg; >>>>   #define VHOST_USER_HDR_SIZE offsetof(VhostUserMsg, payload.u64) >>>> @@ -155,7 +156,8 @@ int vhost_user_iotlb_miss(struct virtio_net >>>> *dev, uint64_t iova, uint8_t perm); >>>>   int vhost_user_host_notifier_ctrl(int vid, bool enable); >>>>   /* socket.c */ >>>> -int read_fd_message(int sockfd, char *buf, int buflen, int *fds, >>>> int fd_num); >>>> +int read_fd_message(int sockfd, char *buf, int buflen, int *fds, >>>> int max_fds, >>>> +        int *fd_num); >>>>   int send_fd_message(int sockfd, char *buf, int buflen, int *fds, >>>> int fd_num); >>>>   #endif >>>>