X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=lib%2Flibrte_eal%2Flinuxapp%2Feal%2Feal_interrupts.c;h=3e9ac41ee582119871f938817210f264a95a6dab;hb=d5c7a09edbe709b50ecde4269b80b9a07627c876;hp=a4b9506a9cd7a864de5b91ac8f58bebcd7a495af;hpb=67b6d3039e9edbc4624c878c6930be5e126e8b58;p=dpdk.git

diff --git a/lib/librte_eal/linuxapp/eal/eal_interrupts.c b/lib/librte_eal/linuxapp/eal/eal_interrupts.c
index a4b9506a9c..3e9ac41ee5 100644
--- a/lib/librte_eal/linuxapp/eal/eal_interrupts.c
+++ b/lib/librte_eal/linuxapp/eal/eal_interrupts.c
@@ -45,6 +45,8 @@
 #include <sys/signalfd.h>
 #include <sys/ioctl.h>
 #include <sys/eventfd.h>
+#include <assert.h>
+#include <stdbool.h>
 
 #include <rte_common.h>
 #include <rte_interrupts.h>
@@ -56,14 +58,13 @@
 #include <rte_lcore.h>
 #include <rte_atomic.h>
 #include <rte_branch_prediction.h>
-#include <rte_ring.h>
 #include <rte_debug.h>
 #include <rte_log.h>
-#include <rte_mempool.h>
 #include <rte_pci.h>
 #include <rte_malloc.h>
 #include <rte_errno.h>
 #include <rte_spinlock.h>
+#include <rte_pause.h>
 
 #include "eal_private.h"
 #include "eal_vfio.h"
@@ -137,7 +138,7 @@ static pthread_t intr_thread;
 
 /* enable legacy (INTx) interrupts */
 static int
-vfio_enable_intx(struct rte_intr_handle *intr_handle) {
+vfio_enable_intx(const struct rte_intr_handle *intr_handle) {
 	struct vfio_irq_set *irq_set;
 	char irq_set_buf[IRQ_SET_BUF_LEN];
 	int len, ret;
@@ -184,7 +185,7 @@ vfio_enable_intx(struct rte_intr_handle *intr_handle) {
 
 /* disable legacy (INTx) interrupts */
 static int
-vfio_disable_intx(struct rte_intr_handle *intr_handle) {
+vfio_disable_intx(const struct rte_intr_handle *intr_handle) {
 	struct vfio_irq_set *irq_set;
 	char irq_set_buf[IRQ_SET_BUF_LEN];
 	int len, ret;
@@ -195,14 +196,14 @@ vfio_disable_intx(struct rte_intr_handle *intr_handle) {
 	irq_set = (struct vfio_irq_set *) irq_set_buf;
 	irq_set->argsz = len;
 	irq_set->count = 1;
-	irq_set->flags = VFIO_IRQ_SET_DATA_NONE | VFIO_IRQ_SET_ACTION_UNMASK;
+	irq_set->flags = VFIO_IRQ_SET_DATA_NONE | VFIO_IRQ_SET_ACTION_MASK;
 	irq_set->index = VFIO_PCI_INTX_IRQ_INDEX;
 	irq_set->start = 0;
 
 	ret = ioctl(intr_handle->vfio_dev_fd, VFIO_DEVICE_SET_IRQS, irq_set);
 
 	if (ret) {
-		RTE_LOG(ERR, EAL, "Error unmasking INTx interrupts for fd %d\n",
+		RTE_LOG(ERR, EAL, "Error masking INTx interrupts for fd %d\n",
 						intr_handle->fd);
 		return -1;
 	}
@@ -227,7 +228,7 @@ vfio_disable_intx(struct rte_intr_handle *intr_handle) {
 
 /* enable MSI interrupts */
 static int
-vfio_enable_msi(struct rte_intr_handle *intr_handle) {
+vfio_enable_msi(const struct rte_intr_handle *intr_handle) {
 	int len, ret;
 	char irq_set_buf[IRQ_SET_BUF_LEN];
 	struct vfio_irq_set *irq_set;
@@ -256,7 +257,7 @@ vfio_enable_msi(struct rte_intr_handle *intr_handle) {
 
 /* disable MSI interrupts */
 static int
-vfio_disable_msi(struct rte_intr_handle *intr_handle) {
+vfio_disable_msi(const struct rte_intr_handle *intr_handle) {
 	struct vfio_irq_set *irq_set;
 	char irq_set_buf[IRQ_SET_BUF_LEN];
 	int len, ret;
@@ -281,7 +282,7 @@ vfio_disable_msi(struct rte_intr_handle *intr_handle) {
 
 /* enable MSI-X interrupts */
 static int
-vfio_enable_msix(struct rte_intr_handle *intr_handle) {
+vfio_enable_msix(const struct rte_intr_handle *intr_handle) {
 	int len, ret;
 	char irq_set_buf[MSIX_IRQ_SET_BUF_LEN];
 	struct vfio_irq_set *irq_set;
@@ -291,18 +292,18 @@ vfio_enable_msix(struct rte_intr_handle *intr_handle) {
 
 	irq_set = (struct vfio_irq_set *) irq_set_buf;
 	irq_set->argsz = len;
-	if (!intr_handle->max_intr)
-		intr_handle->max_intr = 1;
-	else if (intr_handle->max_intr > RTE_MAX_RXTX_INTR_VEC_ID)
-		intr_handle->max_intr = RTE_MAX_RXTX_INTR_VEC_ID + 1;
-
-	irq_set->count = intr_handle->max_intr;
+	/* 0 < irq_set->count < RTE_MAX_RXTX_INTR_VEC_ID + 1 */
+	irq_set->count = intr_handle->max_intr ?
+		(intr_handle->max_intr > RTE_MAX_RXTX_INTR_VEC_ID + 1 ?
+		RTE_MAX_RXTX_INTR_VEC_ID + 1 : intr_handle->max_intr) : 1;
 	irq_set->flags = VFIO_IRQ_SET_DATA_EVENTFD | VFIO_IRQ_SET_ACTION_TRIGGER;
 	irq_set->index = VFIO_PCI_MSIX_IRQ_INDEX;
 	irq_set->start = 0;
 	fd_ptr = (int *) &irq_set->data;
-	memcpy(fd_ptr, intr_handle->efds, sizeof(intr_handle->efds));
-	fd_ptr[intr_handle->max_intr - 1] = intr_handle->fd;
+	/* INTR vector offset 0 reserve for non-efds mapping */
+	fd_ptr[RTE_INTR_VEC_ZERO_OFFSET] = intr_handle->fd;
+	memcpy(&fd_ptr[RTE_INTR_VEC_RXTX_OFFSET], intr_handle->efds,
+		sizeof(*intr_handle->efds) * intr_handle->nb_efd);
 
 	ret = ioctl(intr_handle->vfio_dev_fd, VFIO_DEVICE_SET_IRQS, irq_set);
 
@@ -317,7 +318,7 @@ vfio_enable_msix(struct rte_intr_handle *intr_handle) {
 
 /* disable MSI-X interrupts */
 static int
-vfio_disable_msix(struct rte_intr_handle *intr_handle) {
+vfio_disable_msix(const struct rte_intr_handle *intr_handle) {
 	struct vfio_irq_set *irq_set;
 	char irq_set_buf[MSIX_IRQ_SET_BUF_LEN];
 	int len, ret;
@@ -342,7 +343,7 @@ vfio_disable_msix(struct rte_intr_handle *intr_handle) {
 #endif
 
 static int
-uio_intx_intr_disable(struct rte_intr_handle *intr_handle)
+uio_intx_intr_disable(const struct rte_intr_handle *intr_handle)
 {
 	unsigned char command_high;
 
@@ -366,7 +367,7 @@ uio_intx_intr_disable(struct rte_intr_handle *intr_handle)
 }
 
 static int
-uio_intx_intr_enable(struct rte_intr_handle *intr_handle)
+uio_intx_intr_enable(const struct rte_intr_handle *intr_handle)
 {
 	unsigned char command_high;
 
@@ -390,7 +391,7 @@ uio_intx_intr_enable(struct rte_intr_handle *intr_handle)
 }
 
 static int
-uio_intr_disable(struct rte_intr_handle *intr_handle)
+uio_intr_disable(const struct rte_intr_handle *intr_handle)
 {
 	const int value = 0;
 
@@ -404,7 +405,7 @@ uio_intr_disable(struct rte_intr_handle *intr_handle)
 }
 
 static int
-uio_intr_enable(struct rte_intr_handle *intr_handle)
+uio_intr_enable(const struct rte_intr_handle *intr_handle)
 {
 	const int value = 1;
 
@@ -418,7 +419,7 @@ uio_intr_enable(struct rte_intr_handle *intr_handle)
 }
 
 int
-rte_intr_callback_register(struct rte_intr_handle *intr_handle,
+rte_intr_callback_register(const struct rte_intr_handle *intr_handle,
 			rte_intr_callback_fn cb, void *cb_arg)
 {
 	int ret, wake_thread;
@@ -490,7 +491,7 @@ rte_intr_callback_register(struct rte_intr_handle *intr_handle,
 }
 
 int
-rte_intr_callback_unregister(struct rte_intr_handle *intr_handle,
+rte_intr_callback_unregister(const struct rte_intr_handle *intr_handle,
 			rte_intr_callback_fn cb_fn, void *cb_arg)
 {
 	int ret;
@@ -554,8 +555,11 @@ rte_intr_callback_unregister(struct rte_intr_handle *intr_handle,
 }
 
 int
-rte_intr_enable(struct rte_intr_handle *intr_handle)
+rte_intr_enable(const struct rte_intr_handle *intr_handle)
 {
+	if (intr_handle && intr_handle->type == RTE_INTR_HANDLE_VDEV)
+		return 0;
+
 	if (!intr_handle || intr_handle->fd < 0 || intr_handle->uio_cfg_fd < 0)
 		return -1;
 
@@ -598,8 +602,11 @@ rte_intr_enable(struct rte_intr_handle *intr_handle)
 }
 
 int
-rte_intr_disable(struct rte_intr_handle *intr_handle)
+rte_intr_disable(const struct rte_intr_handle *intr_handle)
 {
+	if (intr_handle && intr_handle->type == RTE_INTR_HANDLE_VDEV)
+		return 0;
+
 	if (!intr_handle || intr_handle->fd < 0 || intr_handle->uio_cfg_fd < 0)
 		return -1;
 
@@ -644,6 +651,7 @@ rte_intr_disable(struct rte_intr_handle *intr_handle)
 static int
 eal_intr_process_interrupts(struct epoll_event *events, int nfds)
 {
+	bool call = false;
 	int n, bytes_read;
 	struct rte_intr_source *src;
 	struct rte_intr_callback *cb;
@@ -692,13 +700,18 @@ eal_intr_process_interrupts(struct epoll_event *events, int nfds)
 			bytes_read = sizeof(buf.vfio_intr_count);
 			break;
 #endif
+		case RTE_INTR_HANDLE_VDEV:
 		case RTE_INTR_HANDLE_EXT:
+			bytes_read = 0;
+			call = true;
+			break;
+
 		default:
 			bytes_read = 1;
 			break;
 		}
 
-		if (src->intr_handle.type != RTE_INTR_HANDLE_EXT) {
+		if (bytes_read > 0) {
 			/**
 			 * read out to clear the ready-to-be-read flag
 			 * for epoll_wait.
@@ -715,12 +728,14 @@ eal_intr_process_interrupts(struct epoll_event *events, int nfds)
 			} else if (bytes_read == 0)
 				RTE_LOG(ERR, EAL, "Read nothing from file "
 					"descriptor %d\n", events[n].data.fd);
+			else
+				call = true;
 		}
 
 		/* grab a lock, again to call callbacks and update status. */
 		rte_spinlock_lock(&intr_lock);
 
-		if (bytes_read > 0) {
+		if (call) {
 
 			/* Finally, call all callbacks. */
 			TAILQ_FOREACH(cb, &src->callbacks, next) {
@@ -730,8 +745,7 @@ eal_intr_process_interrupts(struct epoll_event *events, int nfds)
 				rte_spinlock_unlock(&intr_lock);
 
 				/* call the actual callback */
-				active_cb.cb_fn(&src->intr_handle,
-					active_cb.cb_arg);
+				active_cb.cb_fn(active_cb.cb_arg);
 
 				/*get the lock back. */
 				rte_spinlock_lock(&intr_lock);
@@ -831,7 +845,7 @@ eal_intr_thread_main(__rte_unused void *arg)
 		TAILQ_FOREACH(src, &intr_sources, next) {
 			if (src->callbacks.tqh_first == NULL)
 				continue; /* skip those with no callbacks */
-			ev.events = EPOLLIN | EPOLLPRI;
+			ev.events = EPOLLIN | EPOLLPRI | EPOLLRDHUP | EPOLLHUP;
 			ev.data.fd = src->intr_handle.fd;
 
 			/**
@@ -871,22 +885,25 @@ rte_eal_intr_init(void)
 	 * create a pipe which will be waited by epoll and notified to
 	 * rebuild the wait list of epoll.
 	 */
-	if (pipe(intr_pipe.pipefd) < 0)
+	if (pipe(intr_pipe.pipefd) < 0) {
+		rte_errno = errno;
 		return -1;
+	}
 
 	/* create the host thread to wait/handle the interrupt */
 	ret = pthread_create(&intr_thread, NULL,
 			eal_intr_thread_main, NULL);
 	if (ret != 0) {
+		rte_errno = ret;
 		RTE_LOG(ERR, EAL,
 			"Failed to create thread for interrupt handling\n");
 	} else {
 		/* Set thread_name for aid in debugging. */
 		snprintf(thread_name, RTE_MAX_THREAD_NAME_LEN,
 			"eal-intr-thread");
-		ret_1 = pthread_setname_np(intr_thread, thread_name);
+		ret_1 = rte_thread_setname(intr_thread, thread_name);
 		if (ret_1 != 0)
-			RTE_LOG(ERR, EAL,
+			RTE_LOG(DEBUG, EAL,
 			"Failed to set thread name for interrupt handling\n");
 	}
 
@@ -898,6 +915,7 @@ eal_intr_proc_rxtx_intr(int fd, const struct rte_intr_handle *intr_handle)
 {
 	union rte_intr_read_buffer buf;
 	int bytes_read = 1;
+	int nbytes;
 
 	switch (intr_handle->type) {
 	case RTE_INTR_HANDLE_UIO:
@@ -911,6 +929,14 @@ eal_intr_proc_rxtx_intr(int fd, const struct rte_intr_handle *intr_handle)
 		bytes_read = sizeof(buf.vfio_intr_count);
 		break;
 #endif
+	case RTE_INTR_HANDLE_VDEV:
+		/* for vdev, fd points to:
+		 * a. eventfd which does not need to read out;
+		 * b. datapath fd which needs PMD to read out.
+		 */
+		return;
+	case RTE_INTR_HANDLE_EXT:
+		return;
 	default:
 		bytes_read = 1;
 		RTE_LOG(INFO, EAL, "unexpected intr type\n");
@@ -922,15 +948,15 @@ eal_intr_proc_rxtx_intr(int fd, const struct rte_intr_handle *intr_handle)
 	 * for epoll_wait.
 	 */
 	do {
-		bytes_read = read(fd, &buf, bytes_read);
-		if (bytes_read < 0) {
+		nbytes = read(fd, &buf, bytes_read);
+		if (nbytes < 0) {
 			if (errno == EINTR || errno == EWOULDBLOCK ||
 			    errno == EAGAIN)
 				continue;
 			RTE_LOG(ERR, EAL,
 				"Error reading from fd %d: %s\n",
 				fd, strerror(errno));
-		} else if (bytes_read == 0)
+		} else if (nbytes == 0)
 			RTE_LOG(ERR, EAL, "Read nothing from fd %d\n", fd);
 		return;
 	} while (1);
@@ -1083,10 +1109,14 @@ rte_intr_rx_ctl(struct rte_intr_handle *intr_handle, int epfd,
 	struct rte_epoll_event *rev;
 	struct rte_epoll_data *epdata;
 	int epfd_op;
+	unsigned int efd_idx;
 	int rc = 0;
 
+	efd_idx = (vec >= RTE_INTR_VEC_RXTX_OFFSET) ?
+		(vec - RTE_INTR_VEC_RXTX_OFFSET) : vec;
+
 	if (!intr_handle || intr_handle->nb_efd == 0 ||
-	    vec >= intr_handle->nb_efd) {
+	    efd_idx >= intr_handle->nb_efd) {
 		RTE_LOG(ERR, EAL, "Wrong intr vector number.\n");
 		return -EPERM;
 	}
@@ -1094,7 +1124,7 @@ rte_intr_rx_ctl(struct rte_intr_handle *intr_handle, int epfd,
 	switch (op) {
 	case RTE_INTR_EVENT_ADD:
 		epfd_op = EPOLL_CTL_ADD;
-		rev = &intr_handle->elist[vec];
+		rev = &intr_handle->elist[efd_idx];
 		if (rev->status != RTE_EPOLL_INVALID) {
 			RTE_LOG(INFO, EAL, "Event already been added.\n");
 			return -EEXIST;
@@ -1106,7 +1136,8 @@ rte_intr_rx_ctl(struct rte_intr_handle *intr_handle, int epfd,
 		epdata->data   = data;
 		epdata->cb_fun = (rte_intr_event_cb_t)eal_intr_proc_rxtx_intr;
 		epdata->cb_arg = (void *)intr_handle;
-		rc = rte_epoll_ctl(epfd, epfd_op, intr_handle->efds[vec], rev);
+		rc = rte_epoll_ctl(epfd, epfd_op,
+				   intr_handle->efds[efd_idx], rev);
 		if (!rc)
 			RTE_LOG(DEBUG, EAL,
 				"efd %d associated with vec %d added on epfd %d"
@@ -1116,7 +1147,7 @@ rte_intr_rx_ctl(struct rte_intr_handle *intr_handle, int epfd,
 		break;
 	case RTE_INTR_EVENT_DEL:
 		epfd_op = EPOLL_CTL_DEL;
-		rev = &intr_handle->elist[vec];
+		rev = &intr_handle->elist[efd_idx];
 		if (rev->status == RTE_EPOLL_INVALID) {
 			RTE_LOG(INFO, EAL, "Event does not exist.\n");
 			return -EPERM;
@@ -1134,6 +1165,24 @@ rte_intr_rx_ctl(struct rte_intr_handle *intr_handle, int epfd,
 	return rc;
 }
 
+void
+rte_intr_free_epoll_fd(struct rte_intr_handle *intr_handle)
+{
+	uint32_t i;
+	struct rte_epoll_event *rev;
+
+	for (i = 0; i < intr_handle->nb_efd; i++) {
+		rev = &intr_handle->elist[i];
+		if (rev->status == RTE_EPOLL_INVALID)
+			continue;
+		if (rte_epoll_ctl(rev->epfd, EPOLL_CTL_DEL, rev->fd, rev)) {
+			/* force free if the entry valid */
+			eal_epoll_data_safe_free(rev);
+			rev->status = RTE_EPOLL_INVALID;
+		}
+	}
+}
+
 int
 rte_intr_efd_enable(struct rte_intr_handle *intr_handle, uint32_t nb_efd)
 {
@@ -1141,6 +1190,8 @@ rte_intr_efd_enable(struct rte_intr_handle *intr_handle, uint32_t nb_efd)
 	int fd;
 	uint32_t n = RTE_MIN(nb_efd, (uint32_t)RTE_MAX_RXTX_INTR_VEC_ID);
 
+	assert(nb_efd != 0);
+
 	if (intr_handle->type == RTE_INTR_HANDLE_VFIO_MSIX) {
 		for (i = 0; i < n; i++) {
 			fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
@@ -1148,12 +1199,14 @@ rte_intr_efd_enable(struct rte_intr_handle *intr_handle, uint32_t nb_efd)
 				RTE_LOG(ERR, EAL,
 					"can't setup eventfd, error %i (%s)\n",
 					errno, strerror(errno));
-				return -1;
+				return -errno;
 			}
 			intr_handle->efds[i] = fd;
 		}
 		intr_handle->nb_efd   = n;
 		intr_handle->max_intr = NB_OTHER_INTR + n;
+	} else if (intr_handle->type == RTE_INTR_HANDLE_VDEV) {
+		/* do nothing, and let vdev driver to initialize this struct */
 	} else {
 		intr_handle->efds[0]  = intr_handle->fd;
 		intr_handle->nb_efd   = RTE_MIN(nb_efd, 1U);
@@ -1167,19 +1220,8 @@ void
 rte_intr_efd_disable(struct rte_intr_handle *intr_handle)
 {
 	uint32_t i;
-	struct rte_epoll_event *rev;
-
-	for (i = 0; i < intr_handle->nb_efd; i++) {
-		rev = &intr_handle->elist[i];
-		if (rev->status == RTE_EPOLL_INVALID)
-			continue;
-		if (rte_epoll_ctl(rev->epfd, EPOLL_CTL_DEL, rev->fd, rev)) {
-			/* force free if the entry valid */
-			eal_epoll_data_safe_free(rev);
-			rev->status = RTE_EPOLL_INVALID;
-		}
-	}
 
+	rte_intr_free_epoll_fd(intr_handle);
 	if (intr_handle->max_intr > intr_handle->nb_efd) {
 		for (i = 0; i < intr_handle->nb_efd; i++)
 			close(intr_handle->efds[i]);
@@ -1197,5 +1239,20 @@ rte_intr_dp_is_en(struct rte_intr_handle *intr_handle)
 int
 rte_intr_allow_others(struct rte_intr_handle *intr_handle)
 {
-	return !!(intr_handle->max_intr - intr_handle->nb_efd);
+	if (!rte_intr_dp_is_en(intr_handle))
+		return 1;
+	else
+		return !!(intr_handle->max_intr - intr_handle->nb_efd);
+}
+
+int
+rte_intr_cap_multiple(struct rte_intr_handle *intr_handle)
+{
+	if (intr_handle->type == RTE_INTR_HANDLE_VFIO_MSIX)
+		return 1;
+
+	if (intr_handle->type == RTE_INTR_HANDLE_VDEV)
+		return 1;
+
+	return 0;
 }