# \[PATCH 1/1\] virtio/s390: fix vritio-ccw device teardown - Subsystems: VIRTIO DRIVERS FOR S390, S390 ARCHITECTURE, THE REST - Maintainers: Cornelia Huck, Halil Pasic, Eric Farman, Heiko Carstens, Vasily Gorbik, Alexander Gordeev, Linus Torvalds - List: virtualization - 8 messages, 2 authors - Page: https://ratatoskr.run/virtualization/2021/09/1509846/t ## \[PATCH 1/1\] virtio/s390: fix vritio-ccw device teardown - From: Halil Pasic \ - Date: 2021-09-15 21:58:03 - Message: https://ratatoskr.run/virtualization/2021/09/1509846.md ```text Since commit 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and classic notifiers") we were supposed to make sure that virtio_ccw_release_dev() completes before the ccw device, and the attached dma pool are torn down, but unfortunately we did not. Before that commit it used to be OK to delay cleaning up the memory allocated by virtio-ccw indefinitely (which isn't really intuitive for guys used to destruction happens in reverse construction order). To accomplish this let us take a reference on the ccw device before we allocate the dma_area and give it up after dma_area was freed. Signed-off-by: Halil Pasic Fixes: 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and classic notifiers") Reported-by: --- I'm not certain this is the only hot-unplug and teardonw related problem with virtio-ccw. Some things that are not perfectly clear to me: * What would happen if we observed an hot-unplug while we are doing wait_event() in ccw_io_helper()? Do we get stuck? I don't thin we are guaranteed to receive an irq for a subchannel that is gone. * cdev->online seems to be manipulated under cdev->ccwlock, but in virtio_ccw_remove() we look at it to decide should we clean up or not. What is the idea there? I guess we want to avoid doing if nothing is there or twice. But I don't understand how stuff interlocks. * Can virtio_ccw_remove() get called while !cdev->online and virtio_ccw_online() is running on a different cpu? If yes, what would happen then? The main addresse of these questions is Conny ;). An alternative to this approach would be to inc and dec the refcount in ccw_device_dma_zalloc() and ccw_device_dma_free() respectively. --- drivers/s390/virtio/virtio_ccw.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/s390/virtio/virtio_ccw.c b/drivers/s390/virtio/virtio_ccw.c index d35e7a3f7067..99141df3259b 100644 --- a/drivers/s390/virtio/virtio_ccw.c +++ b/drivers/s390/virtio/virtio_ccw.c @@ -1006,10 +1006,12 @@ static void virtio_ccw_release_dev(struct device *_d) { struct virtio_device *dev = dev_to_virtio(_d); struct virtio_ccw_device *vcdev = to_vc_device(dev); + struct ccw_device *cdev = READ_ONCE(vcdev->cdev); ccw_device_dma_free(vcdev->cdev, vcdev->dma_area, sizeof(*vcdev->dma_area)); kfree(vcdev); + put_device(&cdev->dev); } static int irb_is_error(struct irb *irb) @@ -1262,6 +1264,7 @@ static int virtio_ccw_online(struct ccw_device *cdev) struct virtio_ccw_device *vcdev; unsigned long flags; + get_device(&cdev->dev); vcdev = kzalloc(sizeof(*vcdev), GFP_KERNEL); if (!vcdev) { dev_warn(&cdev->dev, "Could not get memory for virtio\n"); @@ -1315,6 +1318,7 @@ static int virtio_ccw_online(struct ccw_device *cdev) sizeof(*vcdev->dma_area)); } kfree(vcdev); + put_device(&cdev->dev); return ret; } base-commit: 3ca706c189db861b2ca2019a0901b94050ca49d8 -- 2.25.1 _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ``` ## Re: \[PATCH 1/1\] virtio/s390: fix vritio-ccw device teardown - From: Halil Pasic \ - Date: 2021-09-15 22:00:54 - Message: https://ratatoskr.run/virtualization/2021/09/1509848.md ```text s/vritio/virtio/ (subject) [..] On Wed, 15 Sep 2021 23:57:42 +0200 Halil Pasic [ref] wrote: > Since commit 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and > classic notifiers") we were supposed to make sure that > virtio_ccw_release_dev() completes before the ccw device, and the > attached dma pool are torn down, but unfortunately we did not. > Before that commit it used to be OK to delay cleaning up the memory > allocated by virtio-ccw indefinitely (which isn't really intuitive for > guys used to destruction happens in reverse construction order). > [..] _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ``` ## Re: \[PATCH 1/1\] virtio/s390: fix vritio-ccw device teardown - From: Cornelia Huck \ - Date: 2021-09-16 08:59:28 - Message: https://ratatoskr.run/virtualization/2021/09/1510187.md ```text On Wed, Sep 15 2021, Halil Pasic [ref] wrote: > Since commit 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and > classic notifiers") we were supposed to make sure that > virtio_ccw_release_dev() completes before the ccw device, and the > attached dma pool are torn down, but unfortunately we did not. > Before that commit it used to be OK to delay cleaning up the memory > allocated by virtio-ccw indefinitely (which isn't really intuitive for > guys used to destruction happens in reverse construction order). > > To accomplish this let us take a reference on the ccw device before we > allocate the dma_area and give it up after dma_area was freed. > > Signed-off-by: Halil Pasic > Fixes: 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and > classic notifiers") > Reported-by: > --- > > I'm not certain this is the only hot-unplug and teardonw related problem > with virtio-ccw. > > Some things that are not perfectly clear to me: > * What would happen if we observed an hot-unplug while we are doing > wait_event() in ccw_io_helper()? Do we get stuck? I don't thin we > are guaranteed to receive an irq for a subchannel that is gone. Hm. I think we may need to do a wake_up during remove handling. > * cdev->online seems to be manipulated under cdev->ccwlock, but > in virtio_ccw_remove() we look at it to decide should we clean up > or not. What is the idea there? I guess we want to avoid doing > if nothing is there or twice. But I don't understand how stuff > interlocks. We only created the virtio device when we onlined the ccw device. Do you have a better idea how to check for that? (And yes, I'm not sure the locking is correct.) > * Can virtio_ccw_remove() get called while !cdev->online and > virtio_ccw_online() is running on a different cpu? If yes, what would > happen then? All of the remove/online/... etc. callbacks are invoked via the ccw bus code. We have to trust that it gets it correct :) (Or have the common I/O layer maintainers double-check it.) > > The main addresse of these questions is Conny ;). > > An alternative to this approach would be to inc and dec the refcount > in ccw_device_dma_zalloc() and ccw_device_dma_free() respectively. Yeah, I also thought about that. This would give us more get/put operations, but might be the safer option. > > --- > drivers/s390/virtio/virtio_ccw.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/s390/virtio/virtio_ccw.c b/drivers/s390/virtio/virtio_ccw.c > index d35e7a3f7067..99141df3259b 100644 > --- a/drivers/s390/virtio/virtio_ccw.c > +++ b/drivers/s390/virtio/virtio_ccw.c > @@ -1006,10 +1006,12 @@ static void virtio_ccw_release_dev(struct device *_d) > { > struct virtio_device *dev = dev_to_virtio(_d); > struct virtio_ccw_device *vcdev = to_vc_device(dev); > + struct ccw_device *cdev = READ_ONCE(vcdev->cdev); > > ccw_device_dma_free(vcdev->cdev, vcdev->dma_area, > sizeof(*vcdev->dma_area)); > kfree(vcdev); > + put_device(&cdev->dev); > } > > static int irb_is_error(struct irb *irb) > @@ -1262,6 +1264,7 @@ static int virtio_ccw_online(struct ccw_device *cdev) > struct virtio_ccw_device *vcdev; > unsigned long flags; > > + get_device(&cdev->dev); > vcdev = kzalloc(sizeof(*vcdev), GFP_KERNEL); > if (!vcdev) { > dev_warn(&cdev->dev, "Could not get memory for virtio\n"); > @@ -1315,6 +1318,7 @@ static int virtio_ccw_online(struct ccw_device *cdev) > sizeof(*vcdev->dma_area)); > } > kfree(vcdev); > + put_device(&cdev->dev); > return ret; > } > > > base-commit: 3ca706c189db861b2ca2019a0901b94050ca49d8 > -- > 2.25.1 _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ``` ## Re: \[PATCH 1/1\] virtio/s390: fix vritio-ccw device teardown - From: Halil Pasic \ - Date: 2021-09-16 13:18:49 - Message: https://ratatoskr.run/virtualization/2021/09/1510563.md ```text On Thu, 16 Sep 2021 10:59:15 +0200 Cornelia Huck [ref] wrote: > > Since commit 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and > > classic notifiers") we were supposed to make sure that > > virtio_ccw_release_dev() completes before the ccw device, and the > > attached dma pool are torn down, but unfortunately we did not. > > Before that commit it used to be OK to delay cleaning up the memory > > allocated by virtio-ccw indefinitely (which isn't really intuitive for > > guys used to destruction happens in reverse construction order). > > > > To accomplish this let us take a reference on the ccw device before we > > allocate the dma_area and give it up after dma_area was freed. > > > > Signed-off-by: Halil Pasic > > Fixes: 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and > > classic notifiers") > > Reported-by: > > --- > > > > I'm not certain this is the only hot-unplug and teardonw related problem > > with virtio-ccw. > > > > Some things that are not perfectly clear to me: > > * What would happen if we observed an hot-unplug while we are doing > > wait_event() in ccw_io_helper()? Do we get stuck? I don't thin we > > are guaranteed to receive an irq for a subchannel that is gone. > > Hm. I think we may need to do a wake_up during remove handling. My guess is that the BQL is saving us from ever seeing this with QEMU as the hypervisor-userspace. Nevertheless I don't think we should rely on that. > > > * cdev->online seems to be manipulated under cdev->ccwlock, but > > in virtio_ccw_remove() we look at it to decide should we clean up > > or not. What is the idea there? I guess we want to avoid doing > > if nothing is there or twice. But I don't understand how stuff > > interlocks. > > We only created the virtio device when we onlined the ccw device. Do you > have a better idea how to check for that? (And yes, I'm not sure the > locking is correct.) > Thanks, if I find time for it, I will try to understand this better and come back with my findings. > > * Can virtio_ccw_remove() get called while !cdev->online and > > virtio_ccw_online() is running on a different cpu? If yes, what would > > happen then? > > All of the remove/online/... etc. callbacks are invoked via the ccw bus > code. We have to trust that it gets it correct :) (Or have the common > I/O layer maintainers double-check it.) > Vineeth, what is your take on this? Are the struct ccw_driver virtio_ccw_remove and the virtio_ccw_online callbacks mutually exclusive. Please notice that we may initiate the onlining by calling ccw_device_set_online() from a workqueue. @Conny: I'm not sure what is your definition of 'it gets it correct'... I doubt CIO can make things 100% foolproof in this area. > > > > The main addresse of these questions is Conny ;). In any case, I think we can go step by step. I would like the issue this patch intends to address, addressed first. Then we can think about the rest. > > > > An alternative to this approach would be to inc and dec the refcount > > in ccw_device_dma_zalloc() and ccw_device_dma_free() respectively. > > Yeah, I also thought about that. This would give us more get/put > operations, but might be the safer option. My understanding is, that having the ccw device go away while in a middle of doing ccw stuff (about to submit, or waiting for a channel program, or whatever) was bad before. So my intuition tells me that drivers should manage explicitly. Yes virtio_ccw happens to have dma memory whose lifetime is more or less the lifetime of struct virtio_ccw, but that may not be always the case. Thanks for your comments! Regards, Halil _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ``` ## Re: \[PATCH 1/1\] virtio/s390: fix vritio-ccw device teardown - From: Cornelia Huck \ - Date: 2021-09-17 08:40:31 - Message: https://ratatoskr.run/virtualization/2021/09/1512676.md ```text On Thu, Sep 16 2021, Halil Pasic [ref] wrote: > On Thu, 16 Sep 2021 10:59:15 +0200 > Cornelia Huck [ref] wrote: > >> > Since commit 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and >> > classic notifiers") we were supposed to make sure that >> > virtio_ccw_release_dev() completes before the ccw device, and the >> > attached dma pool are torn down, but unfortunately we did not. >> > Before that commit it used to be OK to delay cleaning up the memory >> > allocated by virtio-ccw indefinitely (which isn't really intuitive for >> > guys used to destruction happens in reverse construction order). >> > >> > To accomplish this let us take a reference on the ccw device before we >> > allocate the dma_area and give it up after dma_area was freed. >> > >> > Signed-off-by: Halil Pasic >> > Fixes: 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and >> > classic notifiers") >> > Reported-by: >> > --- >> > >> > I'm not certain this is the only hot-unplug and teardonw related problem >> > with virtio-ccw. >> > >> > Some things that are not perfectly clear to me: >> > * What would happen if we observed an hot-unplug while we are doing >> > wait_event() in ccw_io_helper()? Do we get stuck? I don't thin we >> > are guaranteed to receive an irq for a subchannel that is gone. >> >> Hm. I think we may need to do a wake_up during remove handling. > > My guess is that the BQL is saving us from ever seeing this with QEMU > as the hypervisor-userspace. Nevertheless I don't think we should rely > on that. I agree. Let's do that via a separate patch. > >> >> > * cdev->online seems to be manipulated under cdev->ccwlock, but >> > in virtio_ccw_remove() we look at it to decide should we clean up >> > or not. What is the idea there? I guess we want to avoid doing >> > if nothing is there or twice. But I don't understand how stuff >> > interlocks. >> >> We only created the virtio device when we onlined the ccw device. Do you >> have a better idea how to check for that? (And yes, I'm not sure the >> locking is correct.) >> > > Thanks, if I find time for it, I will try to understand this better and > come back with my findings. > >> > * Can virtio_ccw_remove() get called while !cdev->online and >> > virtio_ccw_online() is running on a different cpu? If yes, what would >> > happen then? >> >> All of the remove/online/... etc. callbacks are invoked via the ccw bus >> code. We have to trust that it gets it correct :) (Or have the common >> I/O layer maintainers double-check it.) >> > > Vineeth, what is your take on this? Are the struct ccw_driver > virtio_ccw_remove and the virtio_ccw_online callbacks mutually > exclusive. Please notice that we may initiate the onlining by > calling ccw_device_set_online() from a workqueue. > > @Conny: I'm not sure what is your definition of 'it gets it correct'... > I doubt CIO can make things 100% foolproof in this area. Not 100% foolproof, but "don't online a device that is in the progress of going away" seems pretty basic to me. > >> > >> > The main addresse of these questions is Conny ;). > > In any case, I think we can go step by step. I would like the issue > this patch intends to address, addressed first. Then we can think > about the rest. > >> > >> > An alternative to this approach would be to inc and dec the refcount >> > in ccw_device_dma_zalloc() and ccw_device_dma_free() respectively. >> >> Yeah, I also thought about that. This would give us more get/put >> operations, but might be the safer option. > > My understanding is, that having the ccw device go away while in a > middle of doing ccw stuff (about to submit, or waiting for a channel > program, or whatever) was bad before. What do you mean with "was bad before"? > So my intuition tells me that > drivers should manage explicitly. Yes virtio_ccw happens to have dma > memory whose lifetime is more or less the lifetime of struct virtio_ccw, > but that may not be always the case. I'm not sure what you're getting at here. Regardless of the lifetime of the dma memory, it depends on the presence of the ccw device to which it is tied. This means that the ccw device must not be released while the dma memory is alive. We can use the approach in your patch here due to the lifetime of the dma memory that virtio-ccw allocates when we start using the device and frees when we stop using the device, or we can use get/put with every allocate/release dma memory pair, which should be safe for everyone? _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ``` ## Re: \[PATCH 1/1\] virtio/s390: fix vritio-ccw device teardown - From: Halil Pasic \ - Date: 2021-09-19 22:39:51 - Message: https://ratatoskr.run/virtualization/2021/09/1514188.md ```text On Fri, 17 Sep 2021 10:40:20 +0200 Cornelia Huck [ref] wrote: > On Thu, Sep 16 2021, Halil Pasic [ref] wrote: > > > On Thu, 16 Sep 2021 10:59:15 +0200 > > Cornelia Huck [ref] wrote: > > > >> > Since commit 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and > >> > classic notifiers") we were supposed to make sure that > >> > virtio_ccw_release_dev() completes before the ccw device, and the > >> > attached dma pool are torn down, but unfortunately we did not. > >> > Before that commit it used to be OK to delay cleaning up the memory > >> > allocated by virtio-ccw indefinitely (which isn't really intuitive for > >> > guys used to destruction happens in reverse construction order). > >> > > >> > To accomplish this let us take a reference on the ccw device before we > >> > allocate the dma_area and give it up after dma_area was freed. > >> > > >> > Signed-off-by: Halil Pasic > >> > Fixes: 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and > >> > classic notifiers") > >> > Reported-by: > >> > --- > >> > > >> > I'm not certain this is the only hot-unplug and teardonw related problem > >> > with virtio-ccw. > >> > > >> > Some things that are not perfectly clear to me: > >> > * What would happen if we observed an hot-unplug while we are doing > >> > wait_event() in ccw_io_helper()? Do we get stuck? I don't thin we > >> > are guaranteed to receive an irq for a subchannel that is gone. > >> > >> Hm. I think we may need to do a wake_up during remove handling. > > > > My guess is that the BQL is saving us from ever seeing this with QEMU > > as the hypervisor-userspace. Nevertheless I don't think we should rely > > on that. > > I agree. Let's do that via a separate patch. > I understand you would like us to finish the discussion on the alternate approach before giving an r-b for this patch, right? > > > >> > >> > * cdev->online seems to be manipulated under cdev->ccwlock, but > >> > in virtio_ccw_remove() we look at it to decide should we clean up > >> > or not. What is the idea there? I guess we want to avoid doing > >> > if nothing is there or twice. But I don't understand how stuff > >> > interlocks. > >> > >> We only created the virtio device when we onlined the ccw device. Do you > >> have a better idea how to check for that? (And yes, I'm not sure the > >> locking is correct.) > >> > > > > Thanks, if I find time for it, I will try to understand this better and > > come back with my findings. > > > >> > * Can virtio_ccw_remove() get called while !cdev->online and > >> > virtio_ccw_online() is running on a different cpu? If yes, what would > >> > happen then? > >> > >> All of the remove/online/... etc. callbacks are invoked via the ccw bus > >> code. We have to trust that it gets it correct :) (Or have the common > >> I/O layer maintainers double-check it.) > >> > > > > Vineeth, what is your take on this? Are the struct ccw_driver > > virtio_ccw_remove and the virtio_ccw_online callbacks mutually > > exclusive. Please notice that we may initiate the onlining by > > calling ccw_device_set_online() from a workqueue. > > > > @Conny: I'm not sure what is your definition of 'it gets it correct'... > > I doubt CIO can make things 100% foolproof in this area. > > Not 100% foolproof, but "don't online a device that is in the progress > of going away" seems pretty basic to me. > I hope Vineeth will chime in on this. > > > >> > > >> > The main addresse of these questions is Conny ;). > > > > In any case, I think we can go step by step. I would like the issue > > this patch intends to address, addressed first. Then we can think > > about the rest. > > > >> > > >> > An alternative to this approach would be to inc and dec the refcount > >> > in ccw_device_dma_zalloc() and ccw_device_dma_free() respectively. > >> > >> Yeah, I also thought about that. This would give us more get/put > >> operations, but might be the safer option. > > > > My understanding is, that having the ccw device go away while in a > > middle of doing ccw stuff (about to submit, or waiting for a channel > > program, or whatever) was bad before. > > What do you mean with "was bad before"? Using an already invalid pointer to the ccw device is always bad. I'm not sure what prevented this from happening before commit 48720ba56891. I'm aware of the fact that virtio_ccw_release_dev() didn't use to deference the vcdev->cdev before that commit, so we didn't have this exact problem. Can you tell me, how did we use to ensure that all dereferences of vcdev->cdev are legit, i.e. happened while the ccw device is still fully alive before commit 48720ba56891? > > > So my intuition tells me that > > drivers should manage explicitly. Yes virtio_ccw happens to have dma > > memory whose lifetime is more or less the lifetime of struct virtio_ccw, > > but that may not be always the case. > > I'm not sure what you're getting at here. Regardless of the lifetime of > the dma memory, it depends on the presence of the ccw device to which it > is tied. This means that the ccw device must not be released while the > dma memory is alive. We can use the approach in your patch here due to > the lifetime of the dma memory that virtio-ccw allocates when we start > using the device and frees when we stop using the device, or we can use > get/put with every allocate/release dma memory pair, which should be > safe for everyone? > What I mean is that ccw_device_dma_[zalloc,free]() take a pointer to the ccw_device. If we get/put in those we can ensure that, provided the alloc and the free calls are properly paired, the device will be still alive (and the pointer valid) for the free, if it was valid for the alloc. But it does not ensure that each and every call to alloc is with a valid pointer, or that other uses of the pointer are OK. So I don't think it is completely safe for everyone, because we could try to use a pointer to a ccw device when not having any dma memory allocated from its pool. This patch takes reference to cdev before the pointer is published via vcdev->cdev and drops the reference after *vcdev is freed. The idea is that the pointee basically outlives the pointer. (Without having a full understanding of how things are synchronized). Regards, Halil _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ``` ## Re: \[PATCH 1/1\] virtio/s390: fix vritio-ccw device teardown - From: Cornelia Huck \ - Date: 2021-09-20 10:30:59 - Message: https://ratatoskr.run/virtualization/2021/09/1514473.md ```text On Mon, Sep 20 2021, Halil Pasic [ref] wrote: > On Fri, 17 Sep 2021 10:40:20 +0200 > Cornelia Huck [ref] wrote: > >> On Thu, Sep 16 2021, Halil Pasic [ref] wrote: >> >> > On Thu, 16 Sep 2021 10:59:15 +0200 >> > Cornelia Huck [ref] wrote: >> > >> >> > Since commit 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and >> >> > classic notifiers") we were supposed to make sure that >> >> > virtio_ccw_release_dev() completes before the ccw device, and the >> >> > attached dma pool are torn down, but unfortunately we did not. >> >> > Before that commit it used to be OK to delay cleaning up the memory >> >> > allocated by virtio-ccw indefinitely (which isn't really intuitive for >> >> > guys used to destruction happens in reverse construction order). >> >> > >> >> > To accomplish this let us take a reference on the ccw device before we >> >> > allocate the dma_area and give it up after dma_area was freed. >> >> > >> >> > Signed-off-by: Halil Pasic >> >> > Fixes: 48720ba56891 ("virtio/s390: use DMA memory for ccw I/O and >> >> > classic notifiers") >> >> > Reported-by: >> >> > --- >> >> > >> >> > I'm not certain this is the only hot-unplug and teardonw related problem >> >> > with virtio-ccw. >> >> > >> >> > Some things that are not perfectly clear to me: >> >> > * What would happen if we observed an hot-unplug while we are doing >> >> > wait_event() in ccw_io_helper()? Do we get stuck? I don't thin we >> >> > are guaranteed to receive an irq for a subchannel that is gone. >> >> >> >> Hm. I think we may need to do a wake_up during remove handling. >> > >> > My guess is that the BQL is saving us from ever seeing this with QEMU >> > as the hypervisor-userspace. Nevertheless I don't think we should rely >> > on that. >> >> I agree. Let's do that via a separate patch. >> > > I understand you would like us to finish the discussion on the alternate > approach before giving an r-b for this patch, right? Yes, exactly. (...) >> >> > An alternative to this approach would be to inc and dec the refcount >> >> > in ccw_device_dma_zalloc() and ccw_device_dma_free() respectively. >> >> >> >> Yeah, I also thought about that. This would give us more get/put >> >> operations, but might be the safer option. >> > >> > My understanding is, that having the ccw device go away while in a >> > middle of doing ccw stuff (about to submit, or waiting for a channel >> > program, or whatever) was bad before. >> >> What do you mean with "was bad before"? > > Using an already invalid pointer to the ccw device is always bad. I'm > not sure what prevented this from happening before commit 48720ba56891. > I'm aware of the fact that virtio_ccw_release_dev() didn't use to > deference the vcdev->cdev before that commit, so we didn't have this > exact problem. Can you tell me, how did we use to ensure that all > dereferences of vcdev->cdev are legit, i.e. happened while the > ccw device is still fully alive before commit 48720ba56891? I'm not sure what that commit is having to do with lifetimes, it did not change anything, only added the extra interaction for the dma buffer. Basically, the vcdev is supposed to be around while the ccw device is online (with a tail end until references have been given up, of course.) It embeds a virtio device that has the ccw device as a parent, which will give us a reference on the ccw device as long as the virtio device is alive. Any interactions with the ccw device (except freeing the dma buffer) are limited to the time where we still have a reference to it via the virtio device. > >> >> > So my intuition tells me that >> > drivers should manage explicitly. Yes virtio_ccw happens to have dma >> > memory whose lifetime is more or less the lifetime of struct virtio_ccw, >> > but that may not be always the case. >> >> I'm not sure what you're getting at here. Regardless of the lifetime of >> the dma memory, it depends on the presence of the ccw device to which it >> is tied. This means that the ccw device must not be released while the >> dma memory is alive. We can use the approach in your patch here due to >> the lifetime of the dma memory that virtio-ccw allocates when we start >> using the device and frees when we stop using the device, or we can use >> get/put with every allocate/release dma memory pair, which should be >> safe for everyone? >> > > What I mean is that ccw_device_dma_[zalloc,free]() take a pointer to the > ccw_device. If we get/put in those we can ensure that, provided the > alloc and the free calls are properly paired, the device will be still > alive (and the pointer valid) for the free, if it was valid for the > alloc. But it does not ensure that each and every call to alloc is with > a valid pointer, or that other uses of the pointer are OK. So I don't > think it is completely safe for everyone, because we could try to use > a pointer to a ccw device when not having any dma memory allocated from > its pool. But the problem is the dma memory, right? Also, it is the same issue for any potential caller of the ccw_device_dma_* interfaces. > > This patch takes reference to cdev before the pointer is published via > vcdev->cdev and drops the reference after *vcdev is freed. The idea is > that the pointee basically outlives the pointer. (Without having a full > understanding of how things are synchronized). I don't think we have to care about accessing ->cdev (see above.) Plus, as we give up the dma memory at the very last point, we would also give up the reference via that memory at the very last point, so I'm not sure what additional problems could come up. _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ``` ## Re: \[PATCH 1/1\] virtio/s390: fix vritio-ccw device teardown - From: Halil Pasic \ - Date: 2021-09-20 13:28:06 - Message: https://ratatoskr.run/virtualization/2021/09/1514664.md ```text On Mon, 20 Sep 2021 12:30:39 +0200 Cornelia Huck [ref] wrote: > On Mon, Sep 20 2021, Halil Pasic [ref] wrote: [..] > > Basically, the vcdev is supposed to be around while the ccw device is > online (with a tail end until references have been given up, of course.) > It embeds a virtio device that has the ccw device as a parent, which > will give us a reference on the ccw device as long as the virtio device > is alive. Any interactions with the ccw device (except freeing the dma > buffer) are limited to the time where we still have a reference to it > via the virtio device. > I didn't remember that device_add() takes a reference to the parent, and that device_del() before device_put(dev) and remove callback. > > > >> > >> > So my intuition tells me that > >> > drivers should manage explicitly. Yes virtio_ccw happens to have dma > >> > memory whose lifetime is more or less the lifetime of struct virtio_ccw, > >> > but that may not be always the case. > >> > >> I'm not sure what you're getting at here. Regardless of the lifetime of > >> the dma memory, it depends on the presence of the ccw device to which it > >> is tied. This means that the ccw device must not be released while the > >> dma memory is alive. We can use the approach in your patch here due to > >> the lifetime of the dma memory that virtio-ccw allocates when we start > >> using the device and frees when we stop using the device, or we can use > >> get/put with every allocate/release dma memory pair, which should be > >> safe for everyone? > >> > > > > What I mean is that ccw_device_dma_[zalloc,free]() take a pointer to the > > ccw_device. If we get/put in those we can ensure that, provided the > > alloc and the free calls are properly paired, the device will be still > > alive (and the pointer valid) for the free, if it was valid for the > > alloc. But it does not ensure that each and every call to alloc is with > > a valid pointer, or that other uses of the pointer are OK. So I don't > > think it is completely safe for everyone, because we could try to use > > a pointer to a ccw device when not having any dma memory allocated from > > its pool. > > But the problem is the dma memory, right? Also, it is the same issue for > any potential caller of the ccw_device_dma_* interfaces. I tend to agree, my argument was based on the assumption that we did not use to take a reference to the ccw device in virtio_ccw_online(), but we do via register_virtio_device(). This reference however gets dropped right before virtio_ccw_release_dev() is called. > > > > > This patch takes reference to cdev before the pointer is published via > > vcdev->cdev and drops the reference after *vcdev is freed. The idea is > > that the pointee basically outlives the pointer. (Without having a full > > understanding of how things are synchronized). > > I don't think we have to care about accessing ->cdev (see above.) Plus, > as we give up the dma memory at the very last point, we would also give > up the reference via that memory at the very last point, so I'm not sure > what additional problems could come up. I understand now. Let me think about it some more. I'm wonderning about leafs. Will come back at you shortly. Regards, Halil _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ```