From: Eric Farman <farman@linux.ibm.com> To: Halil Pasic <pasic@linux.ibm.com> Cc: linux-s390@vger.kernel.org, Cornelia Huck <cohuck@redhat.com>, Alex Williamson <alex.williamson@redhat.com>, Pierre Morel <pmorel@linux.ibm.com>, kvm@vger.kernel.org, Farhan Ali <alifm@linux.ibm.com>, qemu-devel@nongnu.org, qemu-s390x@nongnu.org Subject: Re: [PATCH v2 2/5] vfio-ccw: concurrent I/O handling Date: Fri, 25 Jan 2019 15:21:55 -0500 [thread overview] Message-ID: <cf22b682-199e-f314-244e-81b1c39b5f8c@linux.ibm.com> (raw) In-Reply-To: <20190125135807.17c59cdd@oc2783563651> On 01/25/2019 07:58 AM, Halil Pasic wrote: > On Thu, 24 Jan 2019 21:25:10 -0500 > Eric Farman <farman@linux.ibm.com> wrote: > >>> private = dev_get_drvdata(mdev_parent_dev(mdev)); >>> - if (private->state != VFIO_CCW_STATE_IDLE) >>> + if (private->state == VFIO_CCW_STATE_NOT_OPER || >>> + private->state == VFIO_CCW_STATE_STANDBY) >>> return -EACCES; >>> + if (!mutex_trylock(&private->io_mutex)) >>> + return -EAGAIN; >> >> Ah, I see Halil's difficulty here. >> >> It is true there is a race condition today, and that this doesn't >> address it. That's fine, add it to the todo list. But even with that, >> I don't see what the mutex is enforcing? > > It is protecting the io regions. AFAIU the idea was that only one > thread is accessing the io region(s) at a time to prevent corruption and > reading half-morphed data. > >> Two simultaneous SSCHs will be >> serialized (one will get kicked out with a failed trylock() call), while >> still leaving the window open between cc=0 on the SSCH and the >> subsequent interrupt. In the latter case, a second SSCH will come >> through here, do the copy_from_user below, and then jump to fsm_io_busy >> to return EAGAIN. Do we really want to stomp on io_region in that case? > > I'm not sure I understood you correctly. The interrupt handler does not > take the lock before writing to the io_region. That is one race but it is > easy to fix. > > The bigger problem is that between the interrupt handler has written IRB > area and userspace has read it we may end up destroying it by stomping on > it (to use your words). The userspace reading a wrong (given todays qemu > zeroed out) IRB could lead to follow on problems. I wasn't thinking about a race between the start and interrupt handler, but rather between two near-simultaneous starts. Looking at it more closely, the orb and scsw structs as well as the ret_code field in ccw_io_region are only referenced under the protection of the new mutex (within fsm_io_request, for example), which I guess is the point. So that leaves us with just the irb fields, which you'd mentioned a couple days ago (and which I was trying to ignore since it'd seems to have been discussed enough at the time). So I withdraw my concerns on this point. For now. ;-) > >> Why can't we simply return EAGAIN if state==BUSY? >> > > Sure we can. That would essentially go back to the old way of things: > if not idle return with error. I think this happens both before and after this series. With this series, we just update the io_region with things that are never used because we're busy. Just the error code returned would change > form EACCESS to EAGAIN. Which Isn't necessarily a win, because > conceptually here should be never two interleaved io_requests/start > commands hitting the module. > > >>> >>> region = private->io_region; >>> - if (copy_from_user((void *)region + *ppos, buf, count)) >>> - return -EFAULT; >>> + if (copy_from_user((void *)region + *ppos, buf, count)) { >>> + ret = -EFAULT; >>> + goto out_unlock; >>> + } >
WARNING: multiple messages have this Message-ID (diff)
From: Eric Farman <farman@linux.ibm.com> To: Halil Pasic <pasic@linux.ibm.com> Cc: Cornelia Huck <cohuck@redhat.com>, Farhan Ali <alifm@linux.ibm.com>, Pierre Morel <pmorel@linux.ibm.com>, linux-s390@vger.kernel.org, kvm@vger.kernel.org, qemu-devel@nongnu.org, qemu-s390x@nongnu.org, Alex Williamson <alex.williamson@redhat.com> Subject: Re: [Qemu-devel] [PATCH v2 2/5] vfio-ccw: concurrent I/O handling Date: Fri, 25 Jan 2019 15:21:55 -0500 [thread overview] Message-ID: <cf22b682-199e-f314-244e-81b1c39b5f8c@linux.ibm.com> (raw) In-Reply-To: <20190125135807.17c59cdd@oc2783563651> On 01/25/2019 07:58 AM, Halil Pasic wrote: > On Thu, 24 Jan 2019 21:25:10 -0500 > Eric Farman <farman@linux.ibm.com> wrote: > >>> private = dev_get_drvdata(mdev_parent_dev(mdev)); >>> - if (private->state != VFIO_CCW_STATE_IDLE) >>> + if (private->state == VFIO_CCW_STATE_NOT_OPER || >>> + private->state == VFIO_CCW_STATE_STANDBY) >>> return -EACCES; >>> + if (!mutex_trylock(&private->io_mutex)) >>> + return -EAGAIN; >> >> Ah, I see Halil's difficulty here. >> >> It is true there is a race condition today, and that this doesn't >> address it. That's fine, add it to the todo list. But even with that, >> I don't see what the mutex is enforcing? > > It is protecting the io regions. AFAIU the idea was that only one > thread is accessing the io region(s) at a time to prevent corruption and > reading half-morphed data. > >> Two simultaneous SSCHs will be >> serialized (one will get kicked out with a failed trylock() call), while >> still leaving the window open between cc=0 on the SSCH and the >> subsequent interrupt. In the latter case, a second SSCH will come >> through here, do the copy_from_user below, and then jump to fsm_io_busy >> to return EAGAIN. Do we really want to stomp on io_region in that case? > > I'm not sure I understood you correctly. The interrupt handler does not > take the lock before writing to the io_region. That is one race but it is > easy to fix. > > The bigger problem is that between the interrupt handler has written IRB > area and userspace has read it we may end up destroying it by stomping on > it (to use your words). The userspace reading a wrong (given todays qemu > zeroed out) IRB could lead to follow on problems. I wasn't thinking about a race between the start and interrupt handler, but rather between two near-simultaneous starts. Looking at it more closely, the orb and scsw structs as well as the ret_code field in ccw_io_region are only referenced under the protection of the new mutex (within fsm_io_request, for example), which I guess is the point. So that leaves us with just the irb fields, which you'd mentioned a couple days ago (and which I was trying to ignore since it'd seems to have been discussed enough at the time). So I withdraw my concerns on this point. For now. ;-) > >> Why can't we simply return EAGAIN if state==BUSY? >> > > Sure we can. That would essentially go back to the old way of things: > if not idle return with error. I think this happens both before and after this series. With this series, we just update the io_region with things that are never used because we're busy. Just the error code returned would change > form EACCESS to EAGAIN. Which Isn't necessarily a win, because > conceptually here should be never two interleaved io_requests/start > commands hitting the module. > > >>> >>> region = private->io_region; >>> - if (copy_from_user((void *)region + *ppos, buf, count)) >>> - return -EFAULT; >>> + if (copy_from_user((void *)region + *ppos, buf, count)) { >>> + ret = -EFAULT; >>> + goto out_unlock; >>> + } >
next prev parent reply other threads:[~2019-01-25 20:21 UTC|newest] Thread overview: 134+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-01-21 11:03 [PATCH v2 0/5] vfio-ccw: support hsch/csch (kernel part) Cornelia Huck 2019-01-21 11:03 ` [Qemu-devel] " Cornelia Huck 2019-01-21 11:03 ` [PATCH v2 1/5] vfio-ccw: make it safe to access channel programs Cornelia Huck 2019-01-21 11:03 ` [Qemu-devel] " Cornelia Huck 2019-01-22 14:56 ` Halil Pasic 2019-01-22 14:56 ` [Qemu-devel] " Halil Pasic 2019-01-22 15:19 ` Cornelia Huck 2019-01-22 15:19 ` [Qemu-devel] " Cornelia Huck 2019-01-21 11:03 ` [PATCH v2 2/5] vfio-ccw: concurrent I/O handling Cornelia Huck 2019-01-21 11:03 ` [Qemu-devel] " Cornelia Huck 2019-01-21 20:20 ` Halil Pasic 2019-01-21 20:20 ` [Qemu-devel] " Halil Pasic 2019-01-22 10:29 ` Cornelia Huck 2019-01-22 10:29 ` [Qemu-devel] " Cornelia Huck 2019-01-22 11:17 ` Halil Pasic 2019-01-22 11:17 ` [Qemu-devel] " Halil Pasic 2019-01-22 11:53 ` Cornelia Huck 2019-01-22 11:53 ` [Qemu-devel] " Cornelia Huck 2019-01-22 12:46 ` Halil Pasic 2019-01-22 12:46 ` [Qemu-devel] " Halil Pasic 2019-01-22 17:26 ` Cornelia Huck 2019-01-22 17:26 ` [Qemu-devel] " Cornelia Huck 2019-01-22 19:03 ` Halil Pasic 2019-01-22 19:03 ` [Qemu-devel] " Halil Pasic 2019-01-23 10:34 ` Cornelia Huck 2019-01-23 10:34 ` [Qemu-devel] " Cornelia Huck 2019-01-23 13:06 ` Halil Pasic 2019-01-23 13:06 ` [Qemu-devel] " Halil Pasic 2019-01-23 13:34 ` Cornelia Huck 2019-01-23 13:34 ` [Qemu-devel] " Cornelia Huck 2019-01-24 19:16 ` Eric Farman 2019-01-24 19:16 ` [Qemu-devel] " Eric Farman 2019-01-25 10:13 ` Cornelia Huck 2019-01-25 10:13 ` [Qemu-devel] " Cornelia Huck 2019-01-22 18:33 ` Halil Pasic 2019-01-22 18:33 ` [Qemu-devel] " Halil Pasic 2019-01-23 10:21 ` Cornelia Huck 2019-01-23 10:21 ` [Qemu-devel] " Cornelia Huck 2019-01-23 13:30 ` Halil Pasic 2019-01-23 13:30 ` [Qemu-devel] " Halil Pasic 2019-01-24 10:05 ` Cornelia Huck 2019-01-24 10:05 ` [Qemu-devel] " Cornelia Huck 2019-01-24 10:08 ` Pierre Morel 2019-01-24 10:08 ` [Qemu-devel] " Pierre Morel 2019-01-24 10:19 ` Cornelia Huck 2019-01-24 10:19 ` [Qemu-devel] " Cornelia Huck 2019-01-24 11:18 ` Pierre Morel 2019-01-24 11:18 ` [Qemu-devel] " Pierre Morel 2019-01-24 11:45 ` Halil Pasic 2019-01-24 11:45 ` [Qemu-devel] " Halil Pasic 2019-01-24 19:14 ` Eric Farman 2019-01-24 19:14 ` [Qemu-devel] " Eric Farman 2019-01-25 2:25 ` Eric Farman 2019-01-25 2:25 ` [Qemu-devel] " Eric Farman 2019-01-25 2:37 ` Eric Farman 2019-01-25 2:37 ` [Qemu-devel] " Eric Farman 2019-01-25 10:24 ` Cornelia Huck 2019-01-25 10:24 ` [Qemu-devel] " Cornelia Huck 2019-01-25 12:58 ` Cornelia Huck 2019-01-25 12:58 ` [Qemu-devel] " Cornelia Huck 2019-01-25 14:01 ` Halil Pasic 2019-01-25 14:01 ` [Qemu-devel] " Halil Pasic 2019-01-25 14:21 ` Cornelia Huck 2019-01-25 14:21 ` [Qemu-devel] " Cornelia Huck 2019-01-25 16:04 ` Halil Pasic 2019-01-25 16:04 ` [Qemu-devel] " Halil Pasic 2019-01-28 17:13 ` Cornelia Huck 2019-01-28 17:13 ` [Qemu-devel] " Cornelia Huck 2019-01-28 19:30 ` Halil Pasic 2019-01-28 19:30 ` [Qemu-devel] " Halil Pasic 2019-01-29 9:58 ` Cornelia Huck 2019-01-29 9:58 ` [Qemu-devel] " Cornelia Huck 2019-01-29 19:39 ` Halil Pasic 2019-01-29 19:39 ` [Qemu-devel] " Halil Pasic 2019-01-30 13:29 ` Cornelia Huck 2019-01-30 13:29 ` [Qemu-devel] " Cornelia Huck 2019-01-30 14:32 ` Farhan Ali 2019-01-30 14:32 ` [Qemu-devel] " Farhan Ali 2019-01-28 17:09 ` Cornelia Huck 2019-01-28 17:09 ` [Qemu-devel] " Cornelia Huck 2019-01-28 19:15 ` Halil Pasic 2019-01-28 19:15 ` [Qemu-devel] " Halil Pasic 2019-01-28 21:48 ` Eric Farman 2019-01-28 21:48 ` [Qemu-devel] " Eric Farman 2019-01-29 10:20 ` Cornelia Huck 2019-01-29 10:20 ` [Qemu-devel] " Cornelia Huck 2019-01-29 14:14 ` Eric Farman 2019-01-29 14:14 ` [Qemu-devel] " Eric Farman 2019-01-29 18:53 ` Cornelia Huck 2019-01-29 18:53 ` [Qemu-devel] " Cornelia Huck 2019-01-29 10:10 ` Cornelia Huck 2019-01-29 10:10 ` [Qemu-devel] " Cornelia Huck 2019-01-25 15:57 ` Eric Farman 2019-01-25 15:57 ` [Qemu-devel] " Eric Farman 2019-01-28 17:24 ` Cornelia Huck 2019-01-28 17:24 ` [Qemu-devel] " Cornelia Huck 2019-01-28 21:50 ` Eric Farman 2019-01-28 21:50 ` [Qemu-devel] " Eric Farman 2019-01-25 20:22 ` Eric Farman 2019-01-25 20:22 ` [Qemu-devel] " Eric Farman 2019-01-28 17:31 ` Cornelia Huck 2019-01-28 17:31 ` [Qemu-devel] " Cornelia Huck 2019-01-25 13:09 ` Halil Pasic 2019-01-25 13:09 ` [Qemu-devel] " Halil Pasic 2019-01-25 12:58 ` Halil Pasic 2019-01-25 12:58 ` [Qemu-devel] " Halil Pasic 2019-01-25 20:21 ` Eric Farman [this message] 2019-01-25 20:21 ` Eric Farman 2019-01-21 11:03 ` [PATCH v2 3/5] vfio-ccw: add capabilities chain Cornelia Huck 2019-01-21 11:03 ` [Qemu-devel] " Cornelia Huck 2019-01-23 15:57 ` [qemu-s390x] " Halil Pasic 2019-01-23 15:57 ` [Qemu-devel] " Halil Pasic 2019-01-25 16:19 ` Eric Farman 2019-01-25 16:19 ` [Qemu-devel] " Eric Farman 2019-01-25 21:00 ` Eric Farman 2019-01-25 21:00 ` [Qemu-devel] " Eric Farman 2019-01-28 17:34 ` Cornelia Huck 2019-01-28 17:34 ` [Qemu-devel] " Cornelia Huck 2019-01-21 11:03 ` [PATCH v2 4/5] s390/cio: export hsch to modules Cornelia Huck 2019-01-21 11:03 ` [Qemu-devel] " Cornelia Huck 2019-01-22 15:21 ` [qemu-s390x] " Halil Pasic 2019-01-22 15:21 ` [Qemu-devel] " Halil Pasic 2019-01-21 11:03 ` [PATCH v2 5/5] vfio-ccw: add handling for async channel instructions Cornelia Huck 2019-01-21 11:03 ` [Qemu-devel] " Cornelia Huck 2019-01-23 15:51 ` Halil Pasic 2019-01-23 15:51 ` [Qemu-devel] " Halil Pasic 2019-01-24 10:06 ` Cornelia Huck 2019-01-24 10:06 ` [Qemu-devel] " Cornelia Huck 2019-01-24 10:37 ` Halil Pasic 2019-01-24 10:37 ` [Qemu-devel] " Halil Pasic 2019-01-25 21:00 ` Eric Farman 2019-01-25 21:00 ` [Qemu-devel] " Eric Farman 2019-01-28 17:40 ` Cornelia Huck 2019-01-28 17:40 ` [Qemu-devel] " Cornelia Huck
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=cf22b682-199e-f314-244e-81b1c39b5f8c@linux.ibm.com \ --to=farman@linux.ibm.com \ --cc=alex.williamson@redhat.com \ --cc=alifm@linux.ibm.com \ --cc=cohuck@redhat.com \ --cc=kvm@vger.kernel.org \ --cc=linux-s390@vger.kernel.org \ --cc=pasic@linux.ibm.com \ --cc=pmorel@linux.ibm.com \ --cc=qemu-devel@nongnu.org \ --cc=qemu-s390x@nongnu.org \ /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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.