From patchwork Thu May 5 10:58:31 2011 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Anthony PERARD X-Patchwork-Id: 94248 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from lists.gnu.org (lists.gnu.org [140.186.70.17]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTPS id 3944DB6F4E for ; Thu, 5 May 2011 21:05:36 +1000 (EST) Received: from localhost ([::1]:36765 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1QHwMy-0004IA-AW for incoming@patchwork.ozlabs.org; Thu, 05 May 2011 07:05:20 -0400 Received: from eggs.gnu.org ([140.186.70.92]:49834) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1QHwHB-0002CP-MH for qemu-devel@nongnu.org; Thu, 05 May 2011 06:59:23 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1QHwHA-0003zj-1X for qemu-devel@nongnu.org; Thu, 05 May 2011 06:59:21 -0400 Received: from smtp.citrix.com ([66.165.176.89]:58460) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1QHwH9-0003tE-MZ for qemu-devel@nongnu.org; Thu, 05 May 2011 06:59:19 -0400 X-IronPort-AV: E=Sophos;i="4.64,318,1301889600"; d="scan'208";a="13957045" Received: from ftlpmailmx01.citrite.net ([10.13.107.65]) by FTLPIPO01.CITRIX.COM with ESMTP/TLS/RC4-MD5; 05 May 2011 06:59:19 -0400 Received: from smtp01.ad.xensource.com (10.219.128.104) by smtprelay.citrix.com (10.13.107.65) with Microsoft SMTP Server id 8.3.137.0; Thu, 5 May 2011 06:59:18 -0400 Received: from perard.uk.xensource.com (dhcp-3-28.uk.xensource.com [10.80.3.28]) by smtp01.ad.xensource.com (8.13.1/8.13.1) with ESMTP id p45AwmIs029731; Thu, 5 May 2011 03:59:17 -0700 From: To: QEMU-devel , Anthony Liguori , Alexander Graf Date: Thu, 5 May 2011 11:58:31 +0100 Message-ID: <1304593113-10689-17-git-send-email-anthony.perard@citrix.com> X-Mailer: git-send-email 1.7.2.5 In-Reply-To: <1304593113-10689-1-git-send-email-anthony.perard@citrix.com> References: <1304593113-10689-1-git-send-email-anthony.perard@citrix.com> MIME-Version: 1.0 X-detected-operating-system: by eggs.gnu.org: Genre and OS details not recognized. X-Received-From: 66.165.176.89 Cc: Arun Sharma , Anthony PERARD , Xen Devel , Stefano Stabellini Subject: [Qemu-devel] [PATCH V15 16/18] xen: Initialize event channels and io rings X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Sender: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org From: Arun Sharma Open and bind event channels; map ioreq and buffered ioreq rings. Signed-off-by: Arun Sharma Signed-off-by: Anthony PERARD Signed-off-by: Stefano Stabellini Acked-by: Alexander Graf --- hw/xen_common.h | 2 + xen-all.c | 417 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 419 insertions(+), 0 deletions(-) diff --git a/hw/xen_common.h b/hw/xen_common.h index dd3e896..a1958a0 100644 --- a/hw/xen_common.h +++ b/hw/xen_common.h @@ -107,4 +107,6 @@ static inline int xc_fd(xc_interface *xen_xc) } #endif +void destroy_hvm_domain(void); + #endif /* QEMU_HW_XEN_COMMON_H */ diff --git a/xen-all.c b/xen-all.c index cb01ab9..e849a38 100644 --- a/xen-all.c +++ b/xen-all.c @@ -6,6 +6,8 @@ * */ +#include + #include "hw/pci.h" #include "hw/xen_common.h" #include "hw/xen_backend.h" @@ -13,6 +15,58 @@ #include "xen-mapcache.h" #include "trace.h" +#include +#include + +//#define DEBUG_XEN + +#ifdef DEBUG_XEN +#define DPRINTF(fmt, ...) \ + do { fprintf(stderr, "xen: " fmt, ## __VA_ARGS__); } while (0) +#else +#define DPRINTF(fmt, ...) \ + do { } while (0) +#endif + +/* Compatibility with older version */ +#if __XEN_LATEST_INTERFACE_VERSION__ < 0x0003020a +static inline uint32_t xen_vcpu_eport(shared_iopage_t *shared_page, int i) +{ + return shared_page->vcpu_iodata[i].vp_eport; +} +static inline ioreq_t *xen_vcpu_ioreq(shared_iopage_t *shared_page, int vcpu) +{ + return &shared_page->vcpu_iodata[vcpu].vp_ioreq; +} +# define FMT_ioreq_size PRIx64 +#else +static inline uint32_t xen_vcpu_eport(shared_iopage_t *shared_page, int i) +{ + return shared_page->vcpu_ioreq[i].vp_eport; +} +static inline ioreq_t *xen_vcpu_ioreq(shared_iopage_t *shared_page, int vcpu) +{ + return &shared_page->vcpu_ioreq[vcpu]; +} +# define FMT_ioreq_size "u" +#endif + +#define BUFFER_IO_MAX_DELAY 100 + +typedef struct XenIOState { + shared_iopage_t *shared_page; + buffered_iopage_t *buffered_io_page; + QEMUTimer *buffered_io_timer; + /* the evtchn port for polling the notification, */ + evtchn_port_t *ioreq_local_port; + /* the evtchn fd for polling */ + XenEvtchn xce_handle; + /* which vcpu we are serving */ + int send_vcpu; + + Notifier exit; +} XenIOState; + /* Xen specific function for piix pci */ int xen_pci_slot_get_pirq(PCIDevice *pci_dev, int irq_num) @@ -133,8 +187,304 @@ void xen_vcpu_init(void) } } +/* get the ioreq packets from share mem */ +static ioreq_t *cpu_get_ioreq_from_shared_memory(XenIOState *state, int vcpu) +{ + ioreq_t *req = xen_vcpu_ioreq(state->shared_page, vcpu); + + if (req->state != STATE_IOREQ_READY) { + DPRINTF("I/O request not ready: " + "%x, ptr: %x, port: %"PRIx64", " + "data: %"PRIx64", count: %" FMT_ioreq_size ", size: %" FMT_ioreq_size "\n", + req->state, req->data_is_ptr, req->addr, + req->data, req->count, req->size); + return NULL; + } + + xen_rmb(); /* see IOREQ_READY /then/ read contents of ioreq */ + + req->state = STATE_IOREQ_INPROCESS; + return req; +} + +/* use poll to get the port notification */ +/* ioreq_vec--out,the */ +/* retval--the number of ioreq packet */ +static ioreq_t *cpu_get_ioreq(XenIOState *state) +{ + int i; + evtchn_port_t port; + + port = xc_evtchn_pending(state->xce_handle); + if (port != -1) { + for (i = 0; i < smp_cpus; i++) { + if (state->ioreq_local_port[i] == port) { + break; + } + } + + if (i == smp_cpus) { + hw_error("Fatal error while trying to get io event!\n"); + } + + /* unmask the wanted port again */ + xc_evtchn_unmask(state->xce_handle, port); + + /* get the io packet from shared memory */ + state->send_vcpu = i; + return cpu_get_ioreq_from_shared_memory(state, i); + } + + /* read error or read nothing */ + return NULL; +} + +static uint32_t do_inp(pio_addr_t addr, unsigned long size) +{ + switch (size) { + case 1: + return cpu_inb(addr); + case 2: + return cpu_inw(addr); + case 4: + return cpu_inl(addr); + default: + hw_error("inp: bad size: %04"FMT_pioaddr" %lx", addr, size); + } +} + +static void do_outp(pio_addr_t addr, + unsigned long size, uint32_t val) +{ + switch (size) { + case 1: + return cpu_outb(addr, val); + case 2: + return cpu_outw(addr, val); + case 4: + return cpu_outl(addr, val); + default: + hw_error("outp: bad size: %04"FMT_pioaddr" %lx", addr, size); + } +} + +static void cpu_ioreq_pio(ioreq_t *req) +{ + int i, sign; + + sign = req->df ? -1 : 1; + + if (req->dir == IOREQ_READ) { + if (!req->data_is_ptr) { + req->data = do_inp(req->addr, req->size); + } else { + uint32_t tmp; + + for (i = 0; i < req->count; i++) { + tmp = do_inp(req->addr, req->size); + cpu_physical_memory_write(req->data + (sign * i * req->size), + (uint8_t *) &tmp, req->size); + } + } + } else if (req->dir == IOREQ_WRITE) { + if (!req->data_is_ptr) { + do_outp(req->addr, req->size, req->data); + } else { + for (i = 0; i < req->count; i++) { + uint32_t tmp = 0; + + cpu_physical_memory_read(req->data + (sign * i * req->size), + (uint8_t*) &tmp, req->size); + do_outp(req->addr, req->size, tmp); + } + } + } +} + +static void cpu_ioreq_move(ioreq_t *req) +{ + int i, sign; + + sign = req->df ? -1 : 1; + + if (!req->data_is_ptr) { + if (req->dir == IOREQ_READ) { + for (i = 0; i < req->count; i++) { + cpu_physical_memory_read(req->addr + (sign * i * req->size), + (uint8_t *) &req->data, req->size); + } + } else if (req->dir == IOREQ_WRITE) { + for (i = 0; i < req->count; i++) { + cpu_physical_memory_write(req->addr + (sign * i * req->size), + (uint8_t *) &req->data, req->size); + } + } + } else { + target_ulong tmp; + + if (req->dir == IOREQ_READ) { + for (i = 0; i < req->count; i++) { + cpu_physical_memory_read(req->addr + (sign * i * req->size), + (uint8_t*) &tmp, req->size); + cpu_physical_memory_write(req->data + (sign * i * req->size), + (uint8_t*) &tmp, req->size); + } + } else if (req->dir == IOREQ_WRITE) { + for (i = 0; i < req->count; i++) { + cpu_physical_memory_read(req->data + (sign * i * req->size), + (uint8_t*) &tmp, req->size); + cpu_physical_memory_write(req->addr + (sign * i * req->size), + (uint8_t*) &tmp, req->size); + } + } + } +} + +static void handle_ioreq(ioreq_t *req) +{ + if (!req->data_is_ptr && (req->dir == IOREQ_WRITE) && + (req->size < sizeof (target_ulong))) { + req->data &= ((target_ulong) 1 << (8 * req->size)) - 1; + } + + switch (req->type) { + case IOREQ_TYPE_PIO: + cpu_ioreq_pio(req); + break; + case IOREQ_TYPE_COPY: + cpu_ioreq_move(req); + break; + case IOREQ_TYPE_TIMEOFFSET: + break; + case IOREQ_TYPE_INVALIDATE: + qemu_invalidate_map_cache(); + break; + default: + hw_error("Invalid ioreq type 0x%x\n", req->type); + } +} + +static void handle_buffered_iopage(XenIOState *state) +{ + buf_ioreq_t *buf_req = NULL; + ioreq_t req; + int qw; + + if (!state->buffered_io_page) { + return; + } + + while (state->buffered_io_page->read_pointer != state->buffered_io_page->write_pointer) { + buf_req = &state->buffered_io_page->buf_ioreq[ + state->buffered_io_page->read_pointer % IOREQ_BUFFER_SLOT_NUM]; + req.size = 1UL << buf_req->size; + req.count = 1; + req.addr = buf_req->addr; + req.data = buf_req->data; + req.state = STATE_IOREQ_READY; + req.dir = buf_req->dir; + req.df = 1; + req.type = buf_req->type; + req.data_is_ptr = 0; + qw = (req.size == 8); + if (qw) { + buf_req = &state->buffered_io_page->buf_ioreq[ + (state->buffered_io_page->read_pointer + 1) % IOREQ_BUFFER_SLOT_NUM]; + req.data |= ((uint64_t)buf_req->data) << 32; + } + + handle_ioreq(&req); + + xen_mb(); + state->buffered_io_page->read_pointer += qw ? 2 : 1; + } +} + +static void handle_buffered_io(void *opaque) +{ + XenIOState *state = opaque; + + handle_buffered_iopage(state); + qemu_mod_timer(state->buffered_io_timer, + BUFFER_IO_MAX_DELAY + qemu_get_clock_ms(rt_clock)); +} + +static void cpu_handle_ioreq(void *opaque) +{ + XenIOState *state = opaque; + ioreq_t *req = cpu_get_ioreq(state); + + handle_buffered_iopage(state); + if (req) { + handle_ioreq(req); + + if (req->state != STATE_IOREQ_INPROCESS) { + fprintf(stderr, "Badness in I/O request ... not in service?!: " + "%x, ptr: %x, port: %"PRIx64", " + "data: %"PRIx64", count: %" FMT_ioreq_size ", size: %" FMT_ioreq_size "\n", + req->state, req->data_is_ptr, req->addr, + req->data, req->count, req->size); + destroy_hvm_domain(); + return; + } + + xen_wmb(); /* Update ioreq contents /then/ update state. */ + + /* + * We do this before we send the response so that the tools + * have the opportunity to pick up on the reset before the + * guest resumes and does a hlt with interrupts disabled which + * causes Xen to powerdown the domain. + */ + if (vm_running) { + if (qemu_shutdown_requested_get()) { + destroy_hvm_domain(); + } + if (qemu_reset_requested_get()) { + qemu_system_reset(); + } + } + + req->state = STATE_IORESP_READY; + xc_evtchn_notify(state->xce_handle, state->ioreq_local_port[state->send_vcpu]); + } +} + +static void xen_main_loop_prepare(XenIOState *state) +{ + int evtchn_fd = -1; + + if (state->xce_handle != XC_HANDLER_INITIAL_VALUE) { + evtchn_fd = xc_evtchn_fd(state->xce_handle); + } + + state->buffered_io_timer = qemu_new_timer_ms(rt_clock, handle_buffered_io, + state); + qemu_mod_timer(state->buffered_io_timer, qemu_get_clock_ms(rt_clock)); + + if (evtchn_fd != -1) { + qemu_set_fd_handler(evtchn_fd, cpu_handle_ioreq, NULL, state); + } +} + + /* Initialise Xen */ +static void xen_vm_change_state_handler(void *opaque, int running, int reason) +{ + XenIOState *state = opaque; + if (running) { + xen_main_loop_prepare(state); + } +} + +static void xen_exit_notifier(Notifier *n) +{ + XenIOState *state = container_of(n, XenIOState, exit); + + xc_evtchn_close(state->xce_handle); +} + int xen_init(void) { xen_xc = xen_xc_interface_open(0, 0, 0); @@ -148,9 +498,76 @@ int xen_init(void) int xen_hvm_init(void) { + int i, rc; + unsigned long ioreq_pfn; + XenIOState *state; + + state = qemu_mallocz(sizeof (XenIOState)); + + state->xce_handle = xen_xc_evtchn_open(NULL, 0); + if (state->xce_handle == XC_HANDLER_INITIAL_VALUE) { + perror("xen: event channel open"); + return -errno; + } + + state->exit.notify = xen_exit_notifier; + qemu_add_exit_notifier(&state->exit); + + xc_get_hvm_param(xen_xc, xen_domid, HVM_PARAM_IOREQ_PFN, &ioreq_pfn); + DPRINTF("shared page at pfn %lx\n", ioreq_pfn); + state->shared_page = xc_map_foreign_range(xen_xc, xen_domid, XC_PAGE_SIZE, + PROT_READ|PROT_WRITE, ioreq_pfn); + if (state->shared_page == NULL) { + hw_error("map shared IO page returned error %d handle=" XC_INTERFACE_FMT, + errno, xen_xc); + } + + xc_get_hvm_param(xen_xc, xen_domid, HVM_PARAM_BUFIOREQ_PFN, &ioreq_pfn); + DPRINTF("buffered io page at pfn %lx\n", ioreq_pfn); + state->buffered_io_page = xc_map_foreign_range(xen_xc, xen_domid, XC_PAGE_SIZE, + PROT_READ|PROT_WRITE, ioreq_pfn); + if (state->buffered_io_page == NULL) { + hw_error("map buffered IO page returned error %d", errno); + } + + state->ioreq_local_port = qemu_mallocz(smp_cpus * sizeof (evtchn_port_t)); + + /* FIXME: how about if we overflow the page here? */ + for (i = 0; i < smp_cpus; i++) { + rc = xc_evtchn_bind_interdomain(state->xce_handle, xen_domid, + xen_vcpu_eport(state->shared_page, i)); + if (rc == -1) { + fprintf(stderr, "bind interdomain ioctl error %d\n", errno); + return -1; + } + state->ioreq_local_port[i] = rc; + } + /* Init RAM management */ qemu_map_cache_init(); xen_ram_init(ram_size); + qemu_add_vm_change_state_handler(xen_vm_change_state_handler, state); + return 0; } + +void destroy_hvm_domain(void) +{ + XenXC xc_handle; + int sts; + + xc_handle = xen_xc_interface_open(0, 0, 0); + if (xc_handle == XC_HANDLER_INITIAL_VALUE) { + fprintf(stderr, "Cannot acquire xenctrl handle\n"); + } else { + sts = xc_domain_shutdown(xc_handle, xen_domid, SHUTDOWN_poweroff); + if (sts != 0) { + fprintf(stderr, "? xc_domain_shutdown failed to issue poweroff, " + "sts %d, %s\n", sts, strerror(errno)); + } else { + fprintf(stderr, "Issued domain %d poweroff\n", xen_domid); + } + xc_interface_close(xc_handle); + } +}