From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: [PATCH v2 3/8] s390: vfio-ap: allow assignment of unavailable AP resources to mdev device References: <1555796980-27920-1-git-send-email-akrowiak@linux.ibm.com> <1555796980-27920-4-git-send-email-akrowiak@linux.ibm.com> <2c76d792-00aa-a334-2243-b0f8b91295b2@linux.ibm.com> From: Tony Krowiak Date: Tue, 23 Apr 2019 09:19:00 -0400 MIME-Version: 1.0 In-Reply-To: <2c76d792-00aa-a334-2243-b0f8b91295b2@linux.ibm.com> Content-Type: text/plain; charset="utf-8"; format="flowed" Content-Language: en-US Content-Transfer-Encoding: 8bit Message-Id: <22607063-1194-8fa6-2941-6acecddc3761@linux.ibm.com> Sender: kvm-owner@vger.kernel.org List-Archive: List-Post: To: pmorel@linux.ibm.com, linux-s390@vger.kernel.org, linux-kernel@vger.kernel.org, kvm@vger.kernel.org Cc: freude@linux.ibm.com, borntraeger@de.ibm.com, cohuck@redhat.com, frankja@linux.ibm.com, david@redhat.com, schwidefsky@de.ibm.com, heiko.carstens@de.ibm.com, pasic@linux.ibm.com, alex.williamson@redhat.com, kwankhede@nvidia.com List-ID: On 4/23/19 8:46 AM, Pierre Morel wrote: > On 20/04/2019 23:49, Tony Krowiak wrote: >> The AP architecture does not preclude assignment of AP resources that are >> not yet in the AP configuration (i.e., not available or not online). >> Let's go ahead and implement this facet of the AP architecture for linux >> guests. >> >> Access to AP resources is controlled by bit masks in a guest's SIE >> state description (i.e., the control block containing the state of a >> guest). When an AP instruction is executed on the guest, the firmware >> combines the masks from the guest's SIE state description with the masks >> from the host by doing a logical bitwise AND of the masks to create what >> are defined as effective masks. The effective masks determine which AP >> resources can be accessed by the guest. >> >> Effective masking allows the assignment of AP resources to a guest even >> if the underlying device is not in the AP configuration. If the device >> subsequently becomes configured, the guest will automatically have access >> to the associated AP resources (e.g., AP queues). >> >> The current implementation does not allow assignment of AP resources to >> an mdev device if the AP queues identified by the assignment are not >> bound to the vfio_ap device driver. This patch allows assignment of AP >> resources to the mdev device even if the AP queues are not bound to the >> vfio_ap device driver, as long as the AP queues are not reserved by the >> AP BUS for use by the zcrypt device drivers. If the queues are >> subsequently >> bound to the vfio_ap device driver while a running guest is using the >> mdev >> device, the guest will automatically gain access to the queues. Effective >> masking prevents access to queues until such time as they become >> available. >> >> Signed-off-by: Tony Krowiak >> --- >>   drivers/s390/crypto/vfio_ap_ops.c | 160 >> +++++--------------------------------- >>   1 file changed, 18 insertions(+), 142 deletions(-) >> >> diff --git a/drivers/s390/crypto/vfio_ap_ops.c >> b/drivers/s390/crypto/vfio_ap_ops.c >> index eae4ee24460a..221491a9ba95 100644 >> --- a/drivers/s390/crypto/vfio_ap_ops.c >> +++ b/drivers/s390/crypto/vfio_ap_ops.c >> @@ -113,122 +113,6 @@ static struct attribute_group >> *vfio_ap_mdev_type_groups[] = { >>       NULL, >>   }; >> -struct vfio_ap_queue_reserved { >> -    unsigned long *apid; >> -    unsigned long *apqi; >> -    bool reserved; >> -}; >> - >> -/** >> - * vfio_ap_has_queue >> - * >> - * @dev: an AP queue device >> - * @data: a struct vfio_ap_queue_reserved reference >> - * >> - * Flags whether the AP queue device (@dev) has a queue ID containing >> the APQN, >> - * apid or apqi specified in @data: >> - * >> - * - If @data contains both an apid and apqi value, then @data will >> be flagged >> - *   as reserved if the APID and APQI fields for the AP queue device >> matches >> - * >> - * - If @data contains only an apid value, @data will be flagged as >> - *   reserved if the APID field in the AP queue device matches >> - * >> - * - If @data contains only an apqi value, @data will be flagged as >> - *   reserved if the APQI field in the AP queue device matches >> - * >> - * Returns 0 to indicate the input to function succeeded. Returns >> -EINVAL if >> - * @data does not contain either an apid or apqi. >> - */ >> -static int vfio_ap_has_queue(struct device *dev, void *data) >> -{ >> -    struct vfio_ap_queue_reserved *qres = data; >> -    struct ap_queue *ap_queue = to_ap_queue(dev); >> -    ap_qid_t qid; >> -    unsigned long id; >> - >> -    if (qres->apid && qres->apqi) { >> -        qid = AP_MKQID(*qres->apid, *qres->apqi); >> -        if (qid == ap_queue->qid) >> -            qres->reserved = true; >> -    } else if (qres->apid && !qres->apqi) { >> -        id = AP_QID_CARD(ap_queue->qid); >> -        if (id == *qres->apid) >> -            qres->reserved = true; >> -    } else if (!qres->apid && qres->apqi) { >> -        id = AP_QID_QUEUE(ap_queue->qid); >> -        if (id == *qres->apqi) >> -            qres->reserved = true; >> -    } else { >> -        return -EINVAL; >> -    } >> - >> -    return 0; >> -} >> - >> -/** >> - * vfio_ap_verify_queue_reserved >> - * >> - * @matrix_dev: a mediated matrix device >> - * @apid: an AP adapter ID >> - * @apqi: an AP queue index >> - * >> - * Verifies that the AP queue with @apid/@apqi is reserved by the >> VFIO AP device >> - * driver according to the following rules: >> - * >> - * - If both @apid and @apqi are not NULL, then there must be an AP >> queue >> - *   device bound to the vfio_ap driver with the APQN identified by >> @apid and >> - *   @apqi >> - * >> - * - If only @apid is not NULL, then there must be an AP queue device >> bound >> - *   to the vfio_ap driver with an APQN containing @apid >> - * >> - * - If only @apqi is not NULL, then there must be an AP queue device >> bound >> - *   to the vfio_ap driver with an APQN containing @apqi >> - * >> - * Returns 0 if the AP queue is reserved; otherwise, returns >> -EADDRNOTAVAIL. >> - */ >> -static int vfio_ap_verify_queue_reserved(unsigned long *apid, >> -                     unsigned long *apqi) >> -{ >> -    int ret; >> -    struct vfio_ap_queue_reserved qres; >> - >> -    qres.apid = apid; >> -    qres.apqi = apqi; >> -    qres.reserved = false; >> - >> -    ret = driver_for_each_device(&matrix_dev->vfio_ap_drv->driver, NULL, >> -                     &qres, vfio_ap_has_queue); >> -    if (ret) >> -        return ret; >> - >> -    if (qres.reserved) >> -        return 0; >> - >> -    return -EADDRNOTAVAIL; >> -} >> - >> -static int >> -vfio_ap_mdev_verify_queues_reserved_for_apid(struct ap_matrix_mdev >> *matrix_mdev, >> -                         unsigned long apid) > > You suppress this function but it is still in use in > assign_adapter_store... Right you are. I must have been removed in a subsequent patch, because this code was compiled and tested. It should have been removed here though. > > >> -{ >> -    int ret; >> -    unsigned long apqi; >> -    unsigned long nbits = matrix_mdev->matrix.aqm_max + 1; >> - >> -    if (find_first_bit_inv(matrix_mdev->matrix.aqm, nbits) >= nbits) >> -        return vfio_ap_verify_queue_reserved(&apid, NULL); >> - >> -    for_each_set_bit_inv(apqi, matrix_mdev->matrix.aqm, nbits) { >> -        ret = vfio_ap_verify_queue_reserved(&apid, &apqi); >> -        if (ret) >> -            return ret; >> -    } >> - >> -    return 0; >> -} >> - >>   /** >>    * vfio_ap_mdev_verify_no_sharing >>    * >> @@ -359,9 +243,17 @@ static ssize_t assign_adapter_store(struct device >> *dev, >>       set_bit_inv(apid, matrix_mdev->matrix.apm); >> +    ret = ap_apqn_in_matrix_owned_by_def_drv(matrix_mdev->matrix.apm, >> +                         matrix_mdev->matrix.aqm); >> + >> +    /* If any APQN is owned by a default driver */ >> +    ret = (ret == 1) ? -EADDRNOTAVAIL : ret; >> +    if (ret) >> +        goto error; >> + >>       ret = vfio_ap_mdev_verify_no_sharing(matrix_mdev); >>       if (ret) >> -        goto share_err; >> +        goto error; >>       if (matrix_mdev->shadow_crycb) >>           set_bit_inv(apid, matrix_mdev->shadow_crycb->apm); >> @@ -369,7 +261,7 @@ static ssize_t assign_adapter_store(struct device >> *dev, >>       ret = count; >>       goto done; >> -share_err: >> +error: >>       clear_bit_inv(apid, matrix_mdev->matrix.apm); >>   done: >>       mutex_unlock(&matrix_dev->lock); >> @@ -426,26 +318,6 @@ static ssize_t unassign_adapter_store(struct >> device *dev, >>   } >>   static DEVICE_ATTR_WO(unassign_adapter); >> -static int >> -vfio_ap_mdev_verify_queues_reserved_for_apqi(struct ap_matrix_mdev >> *matrix_mdev, >> -                         unsigned long apqi) >> -{ >> -    int ret; >> -    unsigned long apid; >> -    unsigned long nbits = matrix_mdev->matrix.apm_max + 1; >> - >> -    if (find_first_bit_inv(matrix_mdev->matrix.apm, nbits) >= nbits) >> -        return vfio_ap_verify_queue_reserved(NULL, &apqi); >> - >> -    for_each_set_bit_inv(apid, matrix_mdev->matrix.apm, nbits) { >> -        ret = vfio_ap_verify_queue_reserved(&apid, &apqi); >> -        if (ret) >> -            return ret; >> -    } >> - >> -    return 0; >> -} >> - >>   /** >>    * assign_domain_store >>    * >> @@ -500,12 +372,16 @@ static ssize_t assign_domain_store(struct device >> *dev, >>       mutex_lock(&matrix_dev->lock); >> -    ret = vfio_ap_mdev_verify_queues_reserved_for_apqi(matrix_mdev, >> apqi); >> -    if (ret) >> -        goto done; >> - >>       set_bit_inv(apqi, matrix_mdev->matrix.aqm); >> +    ret = ap_apqn_in_matrix_owned_by_def_drv(matrix_mdev->matrix.apm, >> +                         matrix_mdev->matrix.aqm); >> + >> +    /* If any APQN is owned by a default driver */ >> +    ret = (ret == 1) ? -EADDRNOTAVAIL : ret; >> +    if (ret) >> +        goto error; > > forget to add the error label ? Apparently I did. I'm not sure what happened here since I compiled and tested the code. I must have messed something up when formatting the patches. > >> + >>       ret = vfio_ap_mdev_verify_no_sharing(matrix_mdev); >>       if (ret) >>           goto share_err; >> > >