From patchwork Tue Jul 7 21:32:41 2009 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Gregory Haskins X-Patchwork-Id: 34520 Received: from vger.kernel.org (vger.kernel.org [209.132.176.167]) by demeter.kernel.org (8.14.2/8.14.2) with ESMTP id n67LWxWD027907 for ; Tue, 7 Jul 2009 21:32:59 GMT Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757668AbZGGVc6 (ORCPT ); Tue, 7 Jul 2009 17:32:58 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1757875AbZGGVc5 (ORCPT ); Tue, 7 Jul 2009 17:32:57 -0400 Received: from victor.provo.novell.com ([137.65.250.26]:54610 "EHLO victor.provo.novell.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757668AbZGGVc5 (ORCPT ); Tue, 7 Jul 2009 17:32:57 -0400 Received: from dev.haskins.net (prv-ext-foundry1int.gns.novell.com [137.65.251.240]) by victor.provo.novell.com with ESMTP (TLS encrypted); Tue, 07 Jul 2009 15:32:47 -0600 Received: from dev.haskins.net (localhost [127.0.0.1]) by dev.haskins.net (Postfix) with ESMTP id 478CB4641EA; Tue, 7 Jul 2009 17:32:41 -0400 (EDT) From: Gregory Haskins Subject: [QEMU-KVM PATCH] qemu-kvm: add ioeventfd support To: kvm@vger.kernel.org Cc: mst@redhat.com, avi@redhat.com Date: Tue, 07 Jul 2009 17:32:41 -0400 Message-ID: <20090707213241.23998.91195.stgit@dev.haskins.net> In-Reply-To: <20090707213127.23998.68153.stgit@dev.haskins.net> References: <20090707213127.23998.68153.stgit@dev.haskins.net> User-Agent: StGIT/0.14.3 MIME-Version: 1.0 Sender: kvm-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: kvm@vger.kernel.org An ioeventfd allows an eventfd to attach to a specific PIO/MMIO region in the guest. Any guest-writes to that region will trigger an eventfd signal. Signed-off-by: Gregory Haskins --- libkvm-all.h | 44 +++++++++++++++++++++++++++++++++++++ qemu-kvm.c | 69 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 113 insertions(+), 0 deletions(-) -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html diff --git a/libkvm-all.h b/libkvm-all.h index e16646c..b21bb7e 100644 --- a/libkvm-all.h +++ b/libkvm-all.h @@ -943,6 +943,50 @@ int kvm_get_irq_route_gsi(kvm_context_t kvm); */ int kvm_irqfd(kvm_context_t kvm, int gsi, int flags); +enum { + ioeventfd_option_pio, + ioeventfd_option_datamatch, +}; + +#define IOEVENTFD_FLAG_PIO (1 << ioeventfd_option_pio) +#define IOEVENTFD_FLAG_DATAMATCH (1 << ioeventfd_option_datamatch) + +/*! + * \brief Assign an eventfd to an IO port (PIO or MMIO) + * + * Assigns an eventfd based file-descriptor to a specific PIO or MMIO + * address range. Any guest writes to the specified range will generate + * an eventfd signal. + * + * A data-match pointer can be optionally provided in "datamatch" and only + * writes which match this value exactly will generate an event. The length + * of the datamatch is established by the length of the overall IO range, and + * therefore must be in a natural byte-width for the IO routines of your + * particular architecture (e.g. 1, 2, 4, or 8 bytes on x86_64). + * + * \param kvm Pointer to the current kvm_context + * \param addr The IO address + * \param len The length of the IO region at the address + * \param fd The eventfd file-descriptor + * \param datamatch A optional data-match token + * \param flags FLAG_PIO: PIO, else MMIO, FLAG_DATAMATCH, datamatch valid + */ +int kvm_assign_ioeventfd(kvm_context_t kvm, unsigned long addr, size_t len, + int fd, __u64 datamatch, int flags); + +/*! + * \brief Deassign an ioeventfd from a previously registered IO port + * + * Deassigns an ioeventfd previously registered with kvm_assign_ioeventfd() + * + * \param kvm Pointer to the current kvm_context + * \param addr The IO address to deassign + * \param fd The eventfd file-descriptor + * \param flags FLAG_PIO: PIO, else MMIO + */ +int kvm_deassign_ioeventfd(kvm_context_t kvm, unsigned long addr, + int fd, int flags); + #ifdef KVM_CAP_DEVICE_MSIX int kvm_assign_set_msix_nr(kvm_context_t kvm, struct kvm_assigned_msix_nr *msix_nr); diff --git a/qemu-kvm.c b/qemu-kvm.c index 982ad33..ae77e0a 100644 --- a/qemu-kvm.c +++ b/qemu-kvm.c @@ -1688,6 +1688,75 @@ int kvm_irqfd(kvm_context_t kvm, int gsi, int flags) } #endif /* KVM_CAP_IRQFD */ + +#ifdef KVM_CAP_IOEVENTFD + +#include + +int kvm_assign_ioeventfd(kvm_context_t kvm, unsigned long addr, size_t len, + int fd, __u64 datamatch, int flags) +{ + int r; + int type = flags & IOEVENTFD_FLAG_PIO; + struct kvm_ioeventfd data = { + .datamatch = datamatch, + .addr = addr, + .len = len, + .fd = fd, + }; + + data.flags |= flags & IOEVENTFD_FLAG_DATAMATCH ? + KVM_IOEVENTFD_FLAG_DATAMATCH : 0; + data.flags |= type ? KVM_IOEVENTFD_FLAG_PIO : 0; + + if (!kvm_check_extension(kvm, KVM_CAP_IOEVENTFD)) + return -ENOENT; + + r = ioctl(kvm->vm_fd, KVM_IOEVENTFD, &data); + if (r == -1) + r = -errno; + return r; +} + +int kvm_deassign_ioeventfd(kvm_context_t kvm, unsigned long addr, int fd, + int flags) +{ + int r; + int type = flags & IOEVENTFD_FLAG_PIO; + struct kvm_ioeventfd data = { + .addr = addr, + .fd = fd, + .flags = KVM_IOEVENTFD_FLAG_DEASSIGN | + (type ? KVM_IOEVENTFD_FLAG_PIO : 0), + }; + + if (!kvm_check_extension(kvm, KVM_CAP_IOEVENTFD)) + return -ENOENT; + + r = ioctl(kvm->vm_fd, KVM_IOEVENTFD, &data); + if (r == -1) + r = -errno; + return r; +} + +#else /* KVM_CAP_IOEVENTFD */ + +int kvm_assign_ioeventfd(kvm_context_t kvm, unsigned long addr, size_t len, + int fd, __u64 datamatch, int flags) +{ + return -ENOSYS; +} + +int kvm_deassign_ioeventfd(kvm_context_t kvm, unsigned long addr, int fd, + int flags) +{ + return -ENOSYS; +} + +#endif /* KVM_CAP_IOEVENTFD */ + + + static inline unsigned long kvm_get_thread_id(void) { return syscall(SYS_gettid);