KVM Archive on lore.kernel.org
 help / color / Atom feed
From: Tony Krowiak <akrowiak@linux.ibm.com>
To: Harald Freudenberger <freude@linux.ibm.com>,
	linux-s390@vger.kernel.org, linux-kernel@vger.kernel.org,
	kvm@vger.kernel.org
Cc: borntraeger@de.ibm.com, cohuck@redhat.com, frankja@linux.ibm.com,
	david@redhat.com, mjrosato@linux.ibm.com, schwidefsky@de.ibm.com,
	heiko.carstens@de.ibm.com, pmorel@linux.ibm.com,
	pasic@linux.ibm.com, alex.williamson@redhat.com,
	kwankhede@nvidia.com
Subject: Re: [PATCH v4 5/7] s390: vfio-ap: allow assignment of unavailable AP resources to mdev device
Date: Wed, 19 Jun 2019 09:39:02 -0400
Message-ID: <c9c255e2-6059-3b0e-6f36-9ca5c208a003@linux.ibm.com> (raw)
In-Reply-To: <80639474-8d9f-2f95-be0b-ae44cbbac3fe@linux.ibm.com>

On 6/18/19 2:49 AM, Harald Freudenberger wrote:
> On 17.06.19 17:07, Tony Krowiak wrote:
>> On 6/17/19 6:05 AM, Harald Freudenberger wrote:
>>> On 13.06.19 21:39, Tony Krowiak wrote:
>>>> The AP architecture does not preclude assignment of AP resources that are
>>>> not available. Let's go ahead and implement this facet of the AP
>>>> architecture for linux guests.
>>>>
>>>> The current implementation does not allow assignment of AP adapters or
>>>> domains to an mdev device if the APQNs resulting from the assignment
>>>> reference AP queue devices that are not bound to the vfio_ap device
>>>> driver. This patch allows assignment of AP resources to the mdev device as
>>>> long as the APQNs resulting from the assignment are not reserved by the AP
>>>> BUS for use by the zcrypt device drivers and the APQNs are not assigned to
>>>> another mdev device.
>>>>
>>>> Signed-off-by: Tony Krowiak <akrowiak@linux.ibm.com>
>>>> ---
>>>>    drivers/s390/crypto/vfio_ap_ops.c | 231 ++++++++------------------------------
>>>>    1 file changed, 44 insertions(+), 187 deletions(-)
>>>>
>>>> diff --git a/drivers/s390/crypto/vfio_ap_ops.c b/drivers/s390/crypto/vfio_ap_ops.c
>>>> index 60efd3d7896d..9db86c0db52e 100644
>>>> --- a/drivers/s390/crypto/vfio_ap_ops.c
>>>> +++ b/drivers/s390/crypto/vfio_ap_ops.c
>>>> @@ -406,122 +406,6 @@ 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)
>>>> -{
>>>> -    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
>>>>     *
>>>> @@ -529,18 +413,26 @@ vfio_ap_mdev_verify_queues_reserved_for_apid(struct ap_matrix_mdev *matrix_mdev,
>>>>     * and AP queue indexes comprising the AP matrix are not configured for another
>>>>     * mediated device. AP queue sharing is not allowed.
>>>>     *
>>>> - * @matrix_mdev: the mediated matrix device
>>>> + * @mdev_apm: the mask identifying the adapters assigned to mdev
>>>> + * @mdev_apm: the mask identifying the adapters assigned to mdev
>>> I assume you wanted to write @mdev_aqm ... queues ... at the 2nd line.
>>
>> You assumed correctly. I also mean to say "domains assigned to mdev".
>>
>>>>     *
>>>>     * Returns 0 if the APQNs are not shared, otherwise; returns -EADDRINUSE.
>>>>     */
>>>> -static int vfio_ap_mdev_verify_no_sharing(struct ap_matrix_mdev *matrix_mdev)
>>>> +static int vfio_ap_mdev_verify_no_sharing(unsigned long *mdev_apm,
>>>> +                      unsigned long *mdev_aqm)
>>>>    {
>>>>        struct ap_matrix_mdev *lstdev;
>>>>        DECLARE_BITMAP(apm, AP_DEVICES);
>>>>        DECLARE_BITMAP(aqm, AP_DOMAINS);
>>>>          list_for_each_entry(lstdev, &matrix_dev->mdev_list, node) {
>>>> -        if (matrix_mdev == lstdev)
>>>> +        /*
>>>> +         * If either of the input masks belongs to the mdev to which an
>>>> +         * AP resource is being assigned, then we don't need to verify
>>>> +         * that mdev's masks.
>>>> +         */
>>>> +        if ((mdev_apm == lstdev->matrix.apm) ||
>>>> +            (mdev_aqm == lstdev->matrix.aqm))
>>>>                continue;
>>>>              memset(apm, 0, sizeof(apm));
>>>> @@ -550,12 +442,10 @@ static int vfio_ap_mdev_verify_no_sharing(struct ap_matrix_mdev *matrix_mdev)
>>>>             * We work on full longs, as we can only exclude the leftover
>>>>             * bits in non-inverse order. The leftover is all zeros.
>>>>             */
>>>> -        if (!bitmap_and(apm, matrix_mdev->matrix.apm,
>>>> -                lstdev->matrix.apm, AP_DEVICES))
>>>> +        if (!bitmap_and(apm, mdev_apm, lstdev->matrix.apm, AP_DEVICES))
>>>>                continue;
>>>>    -        if (!bitmap_and(aqm, matrix_mdev->matrix.aqm,
>>>> -                lstdev->matrix.aqm, AP_DOMAINS))
>>>> +        if (!bitmap_and(aqm, mdev_aqm, lstdev->matrix.aqm, AP_DOMAINS))
>>>>                continue;
>>>>              return -EADDRINUSE;
>>>> @@ -564,6 +454,17 @@ static int vfio_ap_mdev_verify_no_sharing(struct ap_matrix_mdev *matrix_mdev)
>>>>        return 0;
>>>>    }
>>>>    +static int vfio_ap_mdev_validate_masks(unsigned long *apm, unsigned long *aqm)
>>>> +{
>>>> +    int ret;
>>>> +
>>>> +    ret = ap_apqn_in_matrix_owned_by_def_drv(apm, aqm);
>>>> +    if (ret)
>>>> +        return (ret == 1) ? -EADDRNOTAVAIL : ret;
>>>> +
>>>> +    return vfio_ap_mdev_verify_no_sharing(apm, aqm);
>>>> +}
>>>> +
>>>>    /**
>>>>     * assign_adapter_store
>>>>     *
>>>> @@ -602,6 +503,7 @@ static ssize_t assign_adapter_store(struct device *dev,
>>>>    {
>>>>        int ret;
>>>>        unsigned long apid;
>>>> +    DECLARE_BITMAP(apm, AP_DEVICES);
>>>>        struct mdev_device *mdev = mdev_from_dev(dev);
>>>>        struct ap_matrix_mdev *matrix_mdev = mdev_get_drvdata(mdev);
>>>>    @@ -616,32 +518,19 @@ static ssize_t assign_adapter_store(struct device *dev,
>>>>        if (apid > matrix_mdev->matrix.apm_max)
>>>>            return -ENODEV;
>>>>    -    /*
>>>> -     * Set the bit in the AP mask (APM) corresponding to the AP adapter
>>>> -     * number (APID). The bits in the mask, from most significant to least
>>>> -     * significant bit, correspond to APIDs 0-255.
>>>> -     */
>>>> -    mutex_lock(&matrix_dev->lock);
>>>> -
>>>> -    ret = vfio_ap_mdev_verify_queues_reserved_for_apid(matrix_mdev, apid);
>>>> -    if (ret)
>>>> -        goto done;
>>>> +    memset(apm, 0, ARRAY_SIZE(apm) * sizeof(*apm));
>>> What about just memset(apm, 0, sizeof(apm));
>>
>> apm is a pointer to an array of unsigned long, so sizeof(apm) will
>> yield the number of bytes in the pointer (8), not the number of bytes in
>> the array (32); or am I wrong about that?
> The
>    DECLARE_BITMAP(apm, AP_DEVICES);
> with the macro definition in types.h:
>    #define DECLARE_BITMAP(name,bits) \
>      unsigned long name[BITS_TO_LONGS(bits)]
> gives this expanded code:
>    unsigned long apm[BITS_TO_LONGS(AP_DEVICES)];
> which is an ordinary unsigned long array.
> So sizeof(apm) gives the size of the array in bytes:
>    sizeof(apm) = 32
> I know this is weird, but it is common practice.

It must be the compiler figures out that apm points to
an array. Although probably not necessary, I will make
the change.

> 
> Funnily look at this code:
>    int blubber[10];
>    int *ptr = blubber;
> sizeof(blubber) gives 10*sizeof(int) but
> sizeof(ptr) gives the size of the pointer pointing to
> the blubber array.
> 
>>
>>>> +    set_bit_inv(apid, apm);
>>>>    +    mutex_lock(&matrix_dev->lock);
>>>> +    ret = vfio_ap_mdev_validate_masks(apm, matrix_mdev->matrix.aqm);
>>>> +    if (ret) {
>>>> +        mutex_unlock(&matrix_dev->lock);
>>>> +        return ret;
>>>> +    }
>>>>        set_bit_inv(apid, matrix_mdev->matrix.apm);
>>>> -
>>>> -    ret = vfio_ap_mdev_verify_no_sharing(matrix_mdev);
>>>> -    if (ret)
>>>> -        goto share_err;
>>>> -
>>>> -    ret = count;
>>>> -    goto done;
>>>> -
>>>> -share_err:
>>>> -    clear_bit_inv(apid, matrix_mdev->matrix.apm);
>>>> -done:
>>>>        mutex_unlock(&matrix_dev->lock);
>>>>    -    return ret;
>>>> +    return count;
>>>>    }
>>>>    static DEVICE_ATTR_WO(assign_adapter);
>>>>    @@ -690,26 +579,6 @@ 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
>>>>     *
>>>> @@ -748,6 +617,7 @@ static ssize_t assign_domain_store(struct device *dev,
>>>>    {
>>>>        int ret;
>>>>        unsigned long apqi;
>>>> +    DECLARE_BITMAP(aqm, AP_DEVICES);
>>> AP_DEVICES -> AP_DOMAINS
>>
>> Copy and paste error, it should be AP_DOMAINS.
>>
>>>>        struct mdev_device *mdev = mdev_from_dev(dev);
>>>>        struct ap_matrix_mdev *matrix_mdev = mdev_get_drvdata(mdev);
>>>>        unsigned long max_apqi = matrix_mdev->matrix.aqm_max;
>>>> @@ -762,27 +632,19 @@ static ssize_t assign_domain_store(struct device *dev,
>>>>        if (apqi > max_apqi)
>>>>            return -ENODEV;
>>>>    -    mutex_lock(&matrix_dev->lock);
>>>> -
>>>> -    ret = vfio_ap_mdev_verify_queues_reserved_for_apqi(matrix_mdev, apqi);
>>>> -    if (ret)
>>>> -        goto done;
>>>> +    memset(aqm, 0, ARRAY_SIZE(aqm) * sizeof(*aqm));
>>> memset(aqm, 0, sizeof(aqm));
>>
>> See response above.
>>
>>>> +    set_bit_inv(apqi, aqm);
>>>>    +    mutex_lock(&matrix_dev->lock);
>>>> +    ret = vfio_ap_mdev_validate_masks(matrix_mdev->matrix.apm, aqm);
>>>> +    if (ret) {
>>>> +        mutex_unlock(&matrix_dev->lock);
>>>> +        return ret;
>>>> +    }
>>>>        set_bit_inv(apqi, matrix_mdev->matrix.aqm);
>>>> -
>>>> -    ret = vfio_ap_mdev_verify_no_sharing(matrix_mdev);
>>>> -    if (ret)
>>>> -        goto share_err;
>>>> -
>>>> -    ret = count;
>>>> -    goto done;
>>>> -
>>>> -share_err:
>>>> -    clear_bit_inv(apqi, matrix_mdev->matrix.aqm);
>>>> -done:
>>>>        mutex_unlock(&matrix_dev->lock);
>>>>    -    return ret;
>>>> +    return count;
>>>>    }
>>>>    static DEVICE_ATTR_WO(assign_domain);
>>>>    @@ -868,11 +730,6 @@ static ssize_t assign_control_domain_store(struct device *dev,
>>>>        if (id > matrix_mdev->matrix.adm_max)
>>>>            return -ENODEV;
>>>>    -    /* Set the bit in the ADM (bitmask) corresponding to the AP control
>>>> -     * domain number (id). The bits in the mask, from most significant to
>>>> -     * least significant, correspond to IDs 0 up to the one less than the
>>>> -     * number of control domains that can be assigned.
>>>> -     */
>>>>        mutex_lock(&matrix_dev->lock);
>>>>        set_bit_inv(id, matrix_mdev->matrix.adm);
>>>>        mutex_unlock(&matrix_dev->lock);
>>>
>>
> 


  reply index

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-06-13 19:39 [PATCH v4 0/7] s390: vfio-ap: dynamic configuration support Tony Krowiak
2019-06-13 19:39 ` [PATCH v4 1/7] s390: vfio-ap: Refactor vfio_ap driver probe and remove callbacks Tony Krowiak
2019-06-17  8:27   ` Harald Freudenberger
2019-06-17 14:24     ` Tony Krowiak
2019-06-18 16:14   ` Cornelia Huck
2019-06-19 12:31     ` Tony Krowiak
2019-06-13 19:39 ` [PATCH v4 2/7] s390: vfio-ap: wait for queue empty on queue reset Tony Krowiak
2019-06-17  8:47   ` Harald Freudenberger
2019-06-17 14:29     ` Tony Krowiak
2019-06-13 19:39 ` [PATCH v4 3/7] s390: zcrypt: driver callback to indicate resource in use Tony Krowiak
2019-06-17  9:28   ` Harald Freudenberger
2019-06-17 14:37     ` Tony Krowiak
2019-06-18 16:25   ` Cornelia Huck
2019-06-19 13:04     ` Tony Krowiak
2019-06-26 21:13       ` Tony Krowiak
2019-06-27  7:25         ` Cornelia Huck
2019-06-27 12:59           ` Tony Krowiak
2019-07-01 19:26       ` Cornelia Huck
2019-07-08 14:27         ` Tony Krowiak
2019-07-09 10:49           ` Cornelia Huck
2019-07-09 21:11             ` Tony Krowiak
2019-06-13 19:39 ` [PATCH v4 4/7] s390: vfio-ap: implement in-use callback for vfio_ap driver Tony Krowiak
2019-06-13 19:39 ` [PATCH v4 5/7] s390: vfio-ap: allow assignment of unavailable AP resources to mdev device Tony Krowiak
2019-06-17 10:05   ` Harald Freudenberger
2019-06-17 15:07     ` Tony Krowiak
2019-06-18  6:49       ` Harald Freudenberger
2019-06-19 13:39         ` Tony Krowiak [this message]
2019-06-13 19:39 ` [PATCH v4 6/7] s390: vfio-ap: allow hot plug/unplug of AP resources using " Tony Krowiak
2019-06-13 19:39 ` [PATCH v4 7/7] s390: vfio-ap: update documentation Tony Krowiak
2019-06-17 11:42   ` Harald Freudenberger
2019-06-17 15:21     ` Tony Krowiak
2019-07-09 15:30 ` [PATCH v4 0/7] s390: vfio-ap: dynamic configuration support Halil Pasic

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=c9c255e2-6059-3b0e-6f36-9ca5c208a003@linux.ibm.com \
    --to=akrowiak@linux.ibm.com \
    --cc=alex.williamson@redhat.com \
    --cc=borntraeger@de.ibm.com \
    --cc=cohuck@redhat.com \
    --cc=david@redhat.com \
    --cc=frankja@linux.ibm.com \
    --cc=freude@linux.ibm.com \
    --cc=heiko.carstens@de.ibm.com \
    --cc=kvm@vger.kernel.org \
    --cc=kwankhede@nvidia.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-s390@vger.kernel.org \
    --cc=mjrosato@linux.ibm.com \
    --cc=pasic@linux.ibm.com \
    --cc=pmorel@linux.ibm.com \
    --cc=schwidefsky@de.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

KVM Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/kvm/0 kvm/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 kvm kvm/ https://lore.kernel.org/kvm \
		kvm@vger.kernel.org
	public-inbox-index kvm

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.kvm


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git