summaryrefslogtreecommitdiffstats
path: root/sys/dev/pci/virtio.c (follow)
Commit message (Collapse)AuthorAgeFilesLines
* non-PCI virtio files have been moved to sys/dev/pvreyk2017-01-211-903/+0
|
* virtio: Move interrupt handler into transport specific codesf2016-07-141-5/+3
| | | | | | | | For MSI-X (and also possibly for other transports), the interrupt handler must do different things. Move it out of virtio.c and into virtio_pci. ARM part tested by patrick@
* Make virtio_enqueue_commit return voidsf2015-12-051-4/+2
| | | | The return value was unused
* KNF fixessf2015-07-181-53/+45
| | | | No object file changes besides line numbers in KASSERTs
* remove uneeded pci includesjsg2015-06-111-4/+2
|
* Axe vq_free_entry_lockedsf2015-04-101-13/+5
| | | | A leftover from netbsd
* Fix typo in macro namesf2015-04-101-12/+12
|
* Remove some includes include-what-you-use claims don'tjsg2015-03-141-2/+1
| | | | | | | have any direct symbols used. Tested for indirect use by compiling amd64/i386/sparc64 kernels. ok tedu@ deraadt@
* include atomic.h to get atomic operations (virtio_membar). ok sftedu2014-12-161-1/+2
|
* publish_avail_idx does not read the used ring, no need to callsf2014-10-081-3/+1
| | | | bus_dmamap_sync for it.
* Add missing memory barriers to virtio.sf2014-10-081-1/+12
| | | | | | The virtio_membar_sync() calls are required even on x86. The *_producer() and *_consumer() calls are only relevant on architectures with less strict memory ordering.
* Some reallocarray() use; review Jean-Philippe Ouellet, patrick keshishianderaadt2014-07-131-3/+3
| | | | ok tedu
* add a size argument to free. will be used soon, but for now default to 0.tedu2014-07-121-2/+2
| | | | after discussions with beck deraadt kettenis.
* fix format string when compiled with VIRTIO_DEBUGjasper2014-06-261-3/+3
| | | | ok sf@
* Fix hang with virtio event_idx featuresf2014-06-151-8/+5
| | | | | | | When using the RING_EVENT_IDX feature, we must first call publish_avail_idx() and then read VQ_AVAIL_EVENT(vq), or there is a race condition that may cause us to miss that the host needs to be notified. This resulted in an occasional hang of network in vio(4).
* Rename write_queue_address -> setup_queuesf2013-03-101-5/+5
| | | | | | | | | For transports supporting several interrupts, this function needs to setup per-queue interrupts in addition of setting the queue address. No logic change. OK jasper@
* if_vio: Add some more debug output if VIRTIO_DEBUG is definedsf2013-03-101-1/+3
| | | | | | | Add GuestCSum feature string virtio: Make some printfs depend on VIRTIO_DEBUG OK jasper@
* Remove excessive sys/cdefs.h inclusionderaadt2012-12-051-3/+1
| | | | ok guenther millert kettenis
* Add $OpenBSD$ CVS Ids.reyk2012-10-121-0/+1
| | | | ok sf@
* Add new drivers for virtio network (vio) and block devices (vioblk, the diskssf2012-09-191-0/+918
attach as scsi disks). These are paravirtualized devices offered by some hypervisors like kvm and virtualbox. The virtio transport driver has the pci specific parts separated out. This will make it easier to add support for mmio (e.g. for ARM) later. OK mikeb OK jasper "commit what you have" deraadt
Copyright © 1996 – 2023 Jason A. Donenfeld. All Rights Reverse Engineered.