From: David Hildenbrand <david@redhat.com>
To: qemu-devel@nongnu.org
Cc: qemu-s390x@nongnu.org, Collin Walling <walling@linux.ibm.com>,
Thomas Huth <thuth@redhat.com>,
Christian Borntraeger <borntraeger@de.ibm.com>,
Cornelia Huck <cohuck@redhat.com>,
Richard Henderson <rth@twiddle.net>
Subject: Re: [Qemu-devel] [PATCH v1 5/6] s390x/pci: Introduce unplug requests and split unplug handler
Date: Thu, 10 Jan 2019 17:12:04 +0100 [thread overview]
Message-ID: <9fef18b2-9d28-c66f-cc3a-0595d5e3572a@redhat.com> (raw)
In-Reply-To: <20190109165349.23631-6-david@redhat.com>
On 09.01.19 17:53, David Hildenbrand wrote:
> PCI on s390x is really weird and how it was modeled in QEMU might not have
> been the right choice. Anyhow, right now it is the case that:
> - Hotplugging a PCI device will silently create a zPCI device
> (if none is provided)
> - Hotunplugging a zPCI device will unplug the PCI device (if any)
> - Hotunplugging a PCI device will unplug also the zPCI device
>
> However both devices are handled via a single hotplug handler call. This
> is problematic for various reasons:
> 1. Unplugging via the zPCI device allows to unplug PCI bridges as
> checks are not performed - bad.
> 2. Unplugging via the zPCI device allows to unplug devices that are not
> hot removable. (check performed in qdev_unplug()) - bad.
> 3. Hotplug handler chains are not possible for the unplug case. In the
> future, the machine might want to override hotplug handlers, to
> process device specific stuff and to then branch off to the actual
> hotplug handler. We need separate hotplug handler calls for both the
> PCI and zPCI device to make this work reliably. All other PCI
> implementations are already prepared to handle this correctly, only
> s390x is missing.
>
> Therefore, introduce the unplug_request handler and properly perform
> unplug checks by redirecting to the separate unplug_request handlers.
> When finally unplugging, perform two separate hotplug_handler_unplug()
> calls, first for the PCI device, followed by the zPCI device. This now
> nicely splits unplugging paths for both devices.
>
> The redirect part is a little hairy, as the user is allowed to trigger
> unplug either via the PCI or the zPCI device. So redirect always to the
> PCI unplug request handler first and remember if that check has been
> performed in the zPCI device. Redirect then to the zPCI device unplug
> request handler to perform the magic. Otherwise we would have a redirect
> loop.
>
> Signed-off-by: David Hildenbrand <david@redhat.com>
> ---
> hw/s390x/s390-pci-bus.c | 162 +++++++++++++++++++++++++++-------------
> hw/s390x/s390-pci-bus.h | 1 +
> 2 files changed, 112 insertions(+), 51 deletions(-)
>
> diff --git a/hw/s390x/s390-pci-bus.c b/hw/s390x/s390-pci-bus.c
> index f27c5852db..bcd8565a14 100644
> --- a/hw/s390x/s390-pci-bus.c
> +++ b/hw/s390x/s390-pci-bus.c
> @@ -148,6 +148,23 @@ out:
> psccb->header.response_code = cpu_to_be16(rc);
> }
>
> +static void s390_pci_perform_unplug(S390PCIBusDevice *pbdev)
> +{
> + HotplugHandler *hotplug_ctrl;
> +
I'll turn this into
if (pbdev->pdev)
and drop the g_assert(). Straight forward than.
> + /* Unplug the PCI device */
> + if (pbdev->state == ZPCI_FS_STANDBY) {
> + hotplug_ctrl = qdev_get_hotplug_handler(DEVICE(pbdev->pdev));
> + hotplug_handler_unplug(hotplug_ctrl, DEVICE(pbdev->pdev),
> + &error_abort);
> + }
> + g_assert(pbdev->state == ZPCI_FS_RESERVED);
> +
> + /* Unplug the zPCI device */
> + hotplug_ctrl = qdev_get_hotplug_handler(DEVICE(pbdev));
> + hotplug_handler_unplug(hotplug_ctrl, DEVICE(pbdev), &error_abort);
> +}
> +
> void s390_pci_sclp_deconfigure(SCCB *sccb)
> {
> IoaCfgSccb *psccb = (IoaCfgSccb *)sccb;
> @@ -179,7 +196,7 @@ void s390_pci_sclp_deconfigure(SCCB *sccb)
> rc = SCLP_RC_NORMAL_COMPLETION;
>
> if (pbdev->release_timer) {
> - qdev_unplug(DEVICE(pbdev->pdev), NULL);
> + s390_pci_perform_unplug(pbdev);
> }
> }
> out:
> @@ -217,6 +234,24 @@ S390PCIBusDevice *s390_pci_find_dev_by_target(S390pciState *s,
> return NULL;
> }
>
> +static S390PCIBusDevice *s390_pci_find_dev_by_pci(S390pciState *s,
> + PCIDevice *pci_dev)
> +{
> + S390PCIBusDevice *pbdev;
> +
> + if (!pci_dev) {
> + return NULL;
> + }
> +
> + QTAILQ_FOREACH(pbdev, &s->zpci_devs, link) {
> + if (pbdev->pdev == pci_dev) {
> + return pbdev;
> + }
> + }
> +
> + return NULL;
> +}
> +
> S390PCIBusDevice *s390_pci_find_dev_by_idx(S390pciState *s, uint32_t idx)
> {
> return g_hash_table_lookup(s->zpci_table, &idx);
> @@ -937,74 +972,98 @@ static void s390_pcihost_timer_cb(void *opaque)
> pbdev->state = ZPCI_FS_STANDBY;
> s390_pci_generate_plug_event(HP_EVENT_CONFIGURED_TO_STBRES,
> pbdev->fh, pbdev->fid);
> - qdev_unplug(DEVICE(pbdev), NULL);
> + s390_pci_perform_unplug(pbdev);
> }
>
> static void s390_pcihost_unplug(HotplugHandler *hotplug_dev, DeviceState *dev,
> Error **errp)
> {
> S390pciState *s = S390_PCI_HOST_BRIDGE(hotplug_dev);
> - PCIDevice *pci_dev = NULL;
> - PCIBus *bus;
> - int32_t devfn;
> S390PCIBusDevice *pbdev = NULL;
>
> + if (object_dynamic_cast(OBJECT(dev), TYPE_PCI_DEVICE)) {
> + PCIDevice *pci_dev = PCI_DEVICE(dev);
> + PCIBus *bus;
> + int32_t devfn;
> +
> + pbdev = s390_pci_find_dev_by_pci(s, PCI_DEVICE(dev));
> + g_assert(pbdev);
> +
> + s390_pci_generate_plug_event(HP_EVENT_STANDBY_TO_RESERVED,
> + pbdev->fh, pbdev->fid);
> + bus = pci_get_bus(pci_dev);
> + devfn = pci_dev->devfn;
> + object_unparent(OBJECT(pci_dev));
> +
> + s390_pci_msix_free(pbdev);
> + s390_pci_iommu_free(s, bus, devfn);
> + pbdev->pdev = NULL;
> + pbdev->state = ZPCI_FS_RESERVED;
> + } else if (object_dynamic_cast(OBJECT(dev), TYPE_S390_PCI_DEVICE)) {
> + pbdev = S390_PCI_DEVICE(dev);
> +
> + if (pbdev->release_timer) {
> + timer_del(pbdev->release_timer);
> + timer_free(pbdev->release_timer);
> + pbdev->release_timer = NULL;
> + }
> + pbdev->fid = 0;
> + QTAILQ_REMOVE(&s->zpci_devs, pbdev, link);
> + g_hash_table_remove(s->zpci_table, &pbdev->idx);
> + object_unparent(OBJECT(pbdev));
> + }
> +}
> +
> +static void s390_pcihost_unplug_request(HotplugHandler *hotplug_dev,
> + DeviceState *dev,
> + Error **errp)
> +{
> + S390pciState *s = S390_PCI_HOST_BRIDGE(hotplug_dev);
> + S390PCIBusDevice *pbdev;
> +
> if (object_dynamic_cast(OBJECT(dev), TYPE_PCI_BRIDGE)) {
> error_setg(errp, "PCI bridge hot unplug currently not supported");
> - return;
> } else if (object_dynamic_cast(OBJECT(dev), TYPE_PCI_DEVICE)) {
> - pci_dev = PCI_DEVICE(dev);
> -
> - QTAILQ_FOREACH(pbdev, &s->zpci_devs, link) {
> - if (pbdev->pdev == pci_dev) {
> - break;
> - }
> - }
> - assert(pbdev != NULL);
> + /*
> + * Redirect the unplug request to the zPCI device and remember that
> + * we've checked the PCI device already (to prevent endless recursion).
> + */
> + pbdev = s390_pci_find_dev_by_pci(s, PCI_DEVICE(dev));
> + g_assert(pbdev);
> + pbdev->pci_unplug_request_processed = true;
> + qdev_unplug(DEVICE(pbdev), errp);
> } else if (object_dynamic_cast(OBJECT(dev), TYPE_S390_PCI_DEVICE)) {
> pbdev = S390_PCI_DEVICE(dev);
> - pci_dev = pbdev->pdev;
> - }
>
> - switch (pbdev->state) {
> - case ZPCI_FS_RESERVED:
> - goto out;
> - case ZPCI_FS_STANDBY:
> - break;
> - default:
> - if (pbdev->release_timer) {
> + /*
> + * If unplug was initially requested for the zPCI device, we
> + * first have to redirect to the PCI device, which will in return
> + * redirect back to us after performing its checks (if the request
> + * is not blocked, e.g. because it's a PCI bridge).
> + */
> + if (pbdev->pdev && !pbdev->pci_unplug_request_processed) {
> + qdev_unplug(DEVICE(pbdev->pdev), errp);
> return;
> }
> - s390_pci_generate_plug_event(HP_EVENT_DECONFIGURE_REQUEST,
> - pbdev->fh, pbdev->fid);
> - pbdev->release_timer = timer_new_ns(QEMU_CLOCK_VIRTUAL,
> - s390_pcihost_timer_cb,
> - pbdev);
> - timer_mod(pbdev->release_timer,
> - qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + HOT_UNPLUG_TIMEOUT);
> - return;
> - }
> + pbdev->pci_unplug_request_processed = false;
>
> - if (pbdev->release_timer) {
> - timer_del(pbdev->release_timer);
> - timer_free(pbdev->release_timer);
> - pbdev->release_timer = NULL;
> + switch (pbdev->state) {
> + case ZPCI_FS_STANDBY:
> + case ZPCI_FS_RESERVED:
> + s390_pci_perform_unplug(pbdev);
> + break;
> + default:
> + if (pbdev->release_timer) {
> + return;
> + }
> + s390_pci_generate_plug_event(HP_EVENT_DECONFIGURE_REQUEST,
> + pbdev->fh, pbdev->fid);
> + pbdev->release_timer = timer_new_ns(QEMU_CLOCK_VIRTUAL,
> + s390_pcihost_timer_cb, pbdev);
> + timer_mod(pbdev->release_timer,
> + qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + HOT_UNPLUG_TIMEOUT);
> + }
> }
> -
> - s390_pci_generate_plug_event(HP_EVENT_STANDBY_TO_RESERVED,
> - pbdev->fh, pbdev->fid);
> - bus = pci_get_bus(pci_dev);
> - devfn = pci_dev->devfn;
> - object_unparent(OBJECT(pci_dev));
> - s390_pci_msix_free(pbdev);
> - s390_pci_iommu_free(s, bus, devfn);
> - pbdev->pdev = NULL;
> - pbdev->state = ZPCI_FS_RESERVED;
> -out:
> - pbdev->fid = 0;
> - QTAILQ_REMOVE(&s->zpci_devs, pbdev, link);
> - g_hash_table_remove(s->zpci_table, &pbdev->idx);
> - object_unparent(OBJECT(pbdev));
> }
>
> static void s390_pci_enumerate_bridge(PCIBus *bus, PCIDevice *pdev,
> @@ -1054,6 +1113,7 @@ static void s390_pcihost_class_init(ObjectClass *klass, void *data)
> dc->realize = s390_pcihost_realize;
> hc->pre_plug = s390_pcihost_pre_plug;
> hc->plug = s390_pcihost_plug;
> + hc->unplug_request = s390_pcihost_unplug_request;
> hc->unplug = s390_pcihost_unplug;
> msi_nonbroken = true;
> }
> diff --git a/hw/s390x/s390-pci-bus.h b/hw/s390x/s390-pci-bus.h
> index f47a0f2da5..6987c31df2 100644
> --- a/hw/s390x/s390-pci-bus.h
> +++ b/hw/s390x/s390-pci-bus.h
> @@ -307,6 +307,7 @@ struct S390PCIBusDevice {
> IndAddr *summary_ind;
> IndAddr *indicator;
> QEMUTimer *release_timer;
> + bool pci_unplug_request_processed;
> QTAILQ_ENTRY(S390PCIBusDevice) link;
> };
>
>
--
Thanks,
David / dhildenb
next prev parent reply other threads:[~2019-01-10 16:12 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-01-09 16:53 [Qemu-devel] [PATCH v1 0/6] s390/pci: hotplug handler fixes and reworks David Hildenbrand
2019-01-09 16:53 ` [Qemu-devel] [PATCH v1 1/6] s390x/pci: Use hotplug_dev instead of looking up the host bridge David Hildenbrand
2019-01-09 16:53 ` [Qemu-devel] [PATCH v1 2/6] s390x/pci: Move some hotplug checks to the pre_plug handler David Hildenbrand
2019-01-09 16:53 ` [Qemu-devel] [PATCH v1 3/6] s390x/pci: Always delete and free the release_timer David Hildenbrand
2019-01-09 16:53 ` [Qemu-devel] [PATCH v1 4/6] s390x/pci: Ignore the unplug call if we already have a release_timer David Hildenbrand
2019-01-09 16:53 ` [Qemu-devel] [PATCH v1 5/6] s390x/pci: Introduce unplug requests and split unplug handler David Hildenbrand
2019-01-10 16:12 ` David Hildenbrand [this message]
2019-01-09 16:53 ` [Qemu-devel] [PATCH v1 6/6] s390x/pci: unplug remaining devices on pcihost reset David Hildenbrand
Reply instructions:
You may reply publicly 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=9fef18b2-9d28-c66f-cc3a-0595d5e3572a@redhat.com \
--to=david@redhat.com \
--cc=borntraeger@de.ibm.com \
--cc=cohuck@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=qemu-s390x@nongnu.org \
--cc=rth@twiddle.net \
--cc=thuth@redhat.com \
--cc=walling@linux.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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).