Welcome to mirror list, hosted at ThFree Co, Russian Federation.

git.openwrt.org/project/libubox.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorFelix Fietkau <nbd@nbd.name>2023-11-27 17:44:33 +0300
committerFelix Fietkau <nbd@nbd.name>2023-11-27 20:30:01 +0300
commitd4c3066e7c5efa5f395c21db77609516d386cbf8 (patch)
tree405d0aaa23f7207eb9d2f0c891ac2fb54bca8659 /udebug.c
parentb3fa3d92e3ebae9b224a0111427d6bdd8b869399 (diff)
udebug: add udebug library code
Copied and adapted from udebug.git Signed-off-by: Felix Fietkau <nbd@nbd.name>
Diffstat (limited to 'udebug.c')
-rw-r--r--udebug.c758
1 files changed, 758 insertions, 0 deletions
diff --git a/udebug.c b/udebug.c
new file mode 100644
index 0000000..5be9b54
--- /dev/null
+++ b/udebug.c
@@ -0,0 +1,758 @@
+/*
+ * udebug - debug ring buffer library
+ *
+ * Copyright (C) 2023 Felix Fietkau <nbd@nbd.name>
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+#define _GNU_SOURCE
+#include <sys/types.h>
+#include <sys/mman.h>
+#include <sys/socket.h>
+#include <unistd.h>
+#include <string.h>
+#include <stdio.h>
+#include <fcntl.h>
+#include <errno.h>
+#include <poll.h>
+#include <time.h>
+#include "udebug-priv.h"
+#include "usock.h"
+
+#define ALIGN(i, sz) (((i) + (sz) - 1) & ~((sz) - 1))
+
+#ifndef MAP_ANONYMOUS
+#define MAP_ANONYMOUS MAP_ANON
+#endif
+
+#define UDEBUG_MIN_ALLOC_LEN 128
+static struct blob_buf b;
+
+static void __randname(char *template)
+{
+ int i;
+ struct timespec ts;
+ unsigned long r;
+
+ clock_gettime(CLOCK_REALTIME, &ts);
+ r = ts.tv_sec + ts.tv_nsec;
+ for (i=0; i<6; i++, r>>=5)
+ template[i] = 'A'+(r&15)+(r&16)*2;
+}
+
+int udebug_id_cmp(const void *k1, const void *k2, void *ptr)
+{
+ uint32_t id1 = (uint32_t)(uintptr_t)k1, id2 = (uint32_t)(uintptr_t)k2;
+ if (id1 < id2)
+ return -1;
+ else if (id1 > id2)
+ return 1;
+ return 0;
+}
+
+static inline int
+shm_open_anon(char *name)
+{
+ char *template = name + strlen(name) - 6;
+ int fd;
+
+ if (template < name || memcmp(template, "XXXXXX", 6) != 0)
+ return -1;
+
+ for (int i = 0; i < 100; i++) {
+ __randname(template);
+ fd = shm_open(name, O_RDWR | O_CREAT | O_EXCL, 0600);
+ if (fd >= 0) {
+ if (shm_unlink(name) < 0) {
+ close(fd);
+ continue;
+ }
+ return fd;
+ }
+
+ if (fd < 0 && errno != EEXIST)
+ return -1;
+ }
+
+ return -1;
+}
+
+static void __udebug_disconnect(struct udebug *ctx, bool reconnect)
+{
+ uloop_fd_delete(&ctx->fd);
+ close(ctx->fd.fd);
+ ctx->fd.fd = -1;
+ ctx->poll_handle = -1;
+ if (ctx->reconnect.cb && reconnect)
+ uloop_timeout_set(&ctx->reconnect, 1);
+}
+
+uint64_t udebug_timestamp(void)
+{
+ struct timespec ts;
+ uint64_t val;
+
+ clock_gettime(CLOCK_REALTIME, &ts);
+
+ val = ts.tv_sec;
+ val *= UDEBUG_TS_SEC;
+ val += ts.tv_nsec / 1000;
+
+ return val;
+}
+
+static int
+__udebug_buf_map(struct udebug_buf *buf)
+{
+ void *ptr, *ptr2;
+
+ ptr = mmap(NULL, buf->head_size + 2 * buf->data_size, PROT_NONE,
+ MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
+ if (ptr == MAP_FAILED)
+ return -1;
+
+ ptr2 = mmap(ptr, buf->head_size + buf->data_size,
+ PROT_READ | PROT_WRITE, MAP_FIXED | MAP_SHARED, buf->fd, 0);
+ if (ptr2 != ptr)
+ goto err_unmap;
+
+ ptr2 = mmap(ptr + buf->head_size + buf->data_size, buf->data_size,
+ PROT_READ | PROT_WRITE, MAP_FIXED | MAP_SHARED, buf->fd,
+ buf->head_size);
+ if (ptr2 != ptr + buf->head_size + buf->data_size)
+ goto err_unmap;
+
+ buf->hdr = ptr;
+ buf->data = ptr + buf->head_size;
+ return 0;
+
+err_unmap:
+ munmap(ptr, buf->head_size + 2 * buf->data_size);
+ return -1;
+}
+
+static int
+writev_retry(int fd, struct iovec *iov, int iov_len, int sock_fd)
+{
+ uint8_t fd_buf[CMSG_SPACE(sizeof(int))] = { 0 };
+ struct msghdr msghdr = { 0 };
+ struct cmsghdr *cmsg;
+ int len = 0;
+ int *pfd;
+
+ msghdr.msg_iov = iov,
+ msghdr.msg_iovlen = iov_len,
+ msghdr.msg_control = fd_buf;
+ msghdr.msg_controllen = sizeof(fd_buf);
+
+ cmsg = CMSG_FIRSTHDR(&msghdr);
+ cmsg->cmsg_type = SCM_RIGHTS;
+ cmsg->cmsg_level = SOL_SOCKET;
+ cmsg->cmsg_len = CMSG_LEN(sizeof(int));
+
+ pfd = (int *) CMSG_DATA(cmsg);
+ msghdr.msg_controllen = cmsg->cmsg_len;
+
+ do {
+ ssize_t cur_len;
+
+ if (sock_fd < 0) {
+ msghdr.msg_control = NULL;
+ msghdr.msg_controllen = 0;
+ } else {
+ *pfd = sock_fd;
+ }
+
+ cur_len = sendmsg(fd, &msghdr, 0);
+ if (cur_len < 0) {
+ struct pollfd pfd = {
+ .fd = fd,
+ .events = POLLOUT
+ };
+
+ switch(errno) {
+ case EAGAIN:
+ poll(&pfd, 1, -1);
+ break;
+ case EINTR:
+ break;
+ default:
+ return -1;
+ }
+ continue;
+ }
+
+ if (len > 0)
+ sock_fd = -1;
+
+ len += cur_len;
+ while (cur_len >= (ssize_t) iov->iov_len) {
+ cur_len -= iov->iov_len;
+ iov_len--;
+ iov++;
+ if (!iov_len)
+ return len;
+ }
+ iov->iov_base += cur_len;
+ iov->iov_len -= cur_len;
+ msghdr.msg_iov = iov;
+ msghdr.msg_iovlen = iov_len;
+ } while (1);
+
+ /* Should never reach here */
+ return -1;
+}
+
+static int
+recv_retry(int fd, struct iovec *iov, bool wait, int *recv_fd)
+{
+ uint8_t fd_buf[CMSG_SPACE(sizeof(int))] = { 0 };
+ struct msghdr msghdr = { 0 };
+ struct cmsghdr *cmsg;
+ int total = 0;
+ int bytes;
+ int *pfd;
+
+ msghdr.msg_iov = iov,
+ msghdr.msg_iovlen = 1,
+ msghdr.msg_control = fd_buf;
+ msghdr.msg_controllen = sizeof(fd_buf);
+
+ cmsg = CMSG_FIRSTHDR(&msghdr);
+ cmsg->cmsg_type = SCM_RIGHTS;
+ cmsg->cmsg_level = SOL_SOCKET;
+ cmsg->cmsg_len = CMSG_LEN(sizeof(int));
+
+ pfd = (int *) CMSG_DATA(cmsg);
+
+ while (iov->iov_len > 0) {
+ if (recv_fd) {
+ msghdr.msg_control = fd_buf;
+ msghdr.msg_controllen = cmsg->cmsg_len;
+ } else {
+ msghdr.msg_control = NULL;
+ msghdr.msg_controllen = 0;
+ }
+
+ *pfd = -1;
+ bytes = recvmsg(fd, &msghdr, 0);
+ if (!bytes)
+ return -2;
+ if (bytes < 0) {
+ bytes = 0;
+ if (errno == EINTR)
+ continue;
+
+ if (errno != EAGAIN)
+ return -2;
+ }
+ if (!wait && !bytes)
+ return 0;
+
+ if (recv_fd)
+ *recv_fd = *pfd;
+ else if (*pfd >= 0)
+ close(*pfd);
+
+ if (bytes > 0)
+ recv_fd = NULL;
+
+ wait = true;
+ iov->iov_len -= bytes;
+ iov->iov_base += bytes;
+ total += bytes;
+
+ if (iov->iov_len > 0) {
+ struct pollfd pfd = {
+ .fd = fd,
+ .events = POLLIN
+ };
+ int ret;
+ do {
+ ret = poll(&pfd, 1, UDEBUG_TIMEOUT);
+ } while (ret < 0 && errno == EINTR);
+
+ if (!(pfd.revents & POLLIN))
+ return -1;
+ }
+ }
+
+ return total;
+}
+
+void udebug_send_msg(struct udebug *ctx, struct udebug_client_msg *msg,
+ struct blob_attr *meta, int fd)
+{
+ struct iovec iov[2] = {
+ { .iov_base = msg, .iov_len = sizeof(*msg) },
+ {}
+ };
+
+ if (!meta) {
+ blob_buf_init(&b, 0);
+ meta = b.head;
+ }
+
+ iov[1].iov_base = meta;
+ iov[1].iov_len = blob_pad_len(meta);
+ writev_retry(ctx->fd.fd, iov, ARRAY_SIZE(iov), fd);
+}
+
+static void
+udebug_buf_msg(struct udebug_buf *buf, enum udebug_client_msg_type type)
+{
+ struct udebug_client_msg msg = {
+ .type = type,
+ .id = buf->id,
+ };
+
+ udebug_send_msg(buf->ctx, &msg, NULL, -1);
+}
+
+static size_t __udebug_headsize(unsigned int ring_size, unsigned int page_size)
+{
+ ring_size *= sizeof(struct udebug_ptr);
+ return ALIGN(sizeof(struct udebug_hdr) + ring_size, page_size);
+}
+
+int udebug_buf_open(struct udebug_buf *buf, int fd, uint32_t ring_size, uint32_t data_size)
+{
+ INIT_LIST_HEAD(&buf->list);
+ buf->fd = fd;
+ buf->ring_size = ring_size;
+ buf->head_size = __udebug_headsize(ring_size, sysconf(_SC_PAGESIZE));
+ buf->data_size = data_size;
+
+ if (buf->ring_size > (1U << 24) || buf->data_size > (1U << 29))
+ return -1;
+
+ if (__udebug_buf_map(buf))
+ return -1;
+
+ if (buf->ring_size != buf->hdr->ring_size ||
+ buf->data_size != buf->hdr->data_size) {
+ munmap(buf->hdr, buf->head_size + 2 * buf->data_size);
+ buf->hdr = NULL;
+ return -1;
+ }
+
+ return 0;
+}
+
+int udebug_buf_init(struct udebug_buf *buf, size_t entries, size_t size)
+{
+ uint32_t pagesz = sysconf(_SC_PAGESIZE);
+ char filename[] = "/udebug.XXXXXX";
+ unsigned int order = 12;
+ uint8_t ring_order = 5;
+ size_t head_size;
+ int fd;
+
+ INIT_LIST_HEAD(&buf->list);
+ if (size < pagesz)
+ size = pagesz;
+ while(size > 1U << order)
+ order++;
+ size = 1 << order;
+ while (entries > 1U << ring_order)
+ ring_order++;
+ entries = 1 << ring_order;
+
+ if (size > (1U << 29) || entries > (1U << 24))
+ return -1;
+
+ head_size = __udebug_headsize(entries, pagesz);
+ while (ALIGN(sizeof(*buf->hdr) + (entries * 2) * sizeof(struct udebug_ptr), pagesz) == head_size)
+ entries *= 2;
+
+ fd = shm_open_anon(filename);
+ if (fd < 0)
+ return -1;
+
+ if (ftruncate(fd, head_size + size) < 0)
+ goto err_close;
+
+ buf->head_size = head_size;
+ buf->data_size = size;
+ buf->ring_size = entries;
+ buf->fd = fd;
+
+ if (__udebug_buf_map(buf))
+ goto err_close;
+
+ buf->hdr->ring_size = entries;
+ buf->hdr->data_size = size;
+
+ /* ensure hdr changes are visible */
+ __sync_synchronize();
+
+ return 0;
+
+err_close:
+ close(fd);
+ return -1;
+}
+
+static void *udebug_buf_alloc(struct udebug_buf *buf, uint32_t ofs, uint32_t len)
+{
+ struct udebug_hdr *hdr = buf->hdr;
+
+ hdr->data_used = u32_max(hdr->data_used, ofs + len + 1);
+
+ /* ensure that data_used update is visible before clobbering data */
+ __sync_synchronize();
+
+ return udebug_buf_ptr(buf, ofs);
+}
+
+uint64_t udebug_buf_flags(struct udebug_buf *buf)
+{
+ struct udebug_hdr *hdr = buf->hdr;
+ uint64_t flags;
+
+ if (!hdr)
+ return 0;
+
+ flags = hdr->flags[0];
+ if (sizeof(flags) != sizeof(uintptr_t))
+ flags |= ((uint64_t)hdr->flags[1]) << 32;
+
+ return flags;
+}
+
+void udebug_entry_init_ts(struct udebug_buf *buf, uint64_t timestamp)
+{
+ struct udebug_hdr *hdr = buf->hdr;
+ struct udebug_ptr *ptr;
+
+ if (!hdr)
+ return;
+
+ ptr = udebug_ring_ptr(hdr, hdr->head);
+ ptr->start = hdr->data_head;
+ ptr->len = 0;
+ ptr->timestamp = timestamp;
+}
+
+void *udebug_entry_append(struct udebug_buf *buf, const void *data, uint32_t len)
+{
+ struct udebug_hdr *hdr = buf->hdr;
+ struct udebug_ptr *ptr;
+ uint32_t ofs;
+ void *ret;
+
+ if (!hdr)
+ return NULL;
+
+ ptr = udebug_ring_ptr(hdr, hdr->head);
+ ofs = ptr->start + ptr->len;
+ if (ptr->len + len > buf->data_size / 2)
+ return NULL;
+
+ ret = udebug_buf_alloc(buf, ofs, len);
+ if (data)
+ memcpy(ret, data, len);
+ ptr->len += len;
+
+ return ret;
+}
+
+int udebug_entry_printf(struct udebug_buf *buf, const char *fmt, ...)
+{
+ va_list ap;
+ size_t ret;
+
+ va_start(ap, fmt);
+ ret = udebug_entry_vprintf(buf, fmt, ap);
+ va_end(ap);
+
+ return ret;
+}
+
+int udebug_entry_vprintf(struct udebug_buf *buf, const char *fmt, va_list ap)
+{
+ struct udebug_hdr *hdr = buf->hdr;
+ struct udebug_ptr *ptr;
+ uint32_t ofs;
+ uint32_t len;
+ char *str;
+
+ if (!hdr)
+ return -1;
+
+ ptr = udebug_ring_ptr(hdr, hdr->head);
+ ofs = ptr->start + ptr->len;
+ if (ptr->len > buf->data_size / 2)
+ return -1;
+
+ str = udebug_buf_alloc(buf, ofs, UDEBUG_MIN_ALLOC_LEN);
+ len = vsnprintf(str, UDEBUG_MIN_ALLOC_LEN, fmt, ap);
+ if (len <= UDEBUG_MIN_ALLOC_LEN)
+ goto out;
+
+ if (ptr->len + len > buf->data_size / 2)
+ return -1;
+
+ udebug_buf_alloc(buf, ofs, len + 1);
+ len = vsnprintf(str, len, fmt, ap);
+
+out:
+ ptr->len += len;
+ return 0;
+}
+
+void udebug_entry_add(struct udebug_buf *buf)
+{
+ struct udebug_hdr *hdr = buf->hdr;
+ struct udebug_ptr *ptr = udebug_ring_ptr(hdr, hdr->head);
+ uint32_t notify;
+ uint8_t *data;
+
+ /* ensure strings are always 0-terminated */
+ data = udebug_buf_ptr(buf, ptr->start + ptr->len);
+ *data = 0;
+ hdr->data_head = ptr->start + ptr->len + 1;
+
+ /* ensure that all data changes are visible before advancing head */
+ __sync_synchronize();
+
+ u32_set(&hdr->head, u32_get(&hdr->head) + 1);
+ if (!u32_get(&hdr->head))
+ u32_set(&hdr->head_hi, u32_get(&hdr->head_hi) + 1);
+
+ /* ensure that head change is visible */
+ __sync_synchronize();
+
+ notify = __atomic_exchange_n(&hdr->notify, 0, __ATOMIC_RELAXED);
+ if (notify) {
+ struct udebug_client_msg msg = {
+ .type = CL_MSG_RING_NOTIFY,
+ .id = buf->id,
+ .notify_mask = notify,
+ };
+ blob_buf_init(&b, 0);
+
+ udebug_send_msg(buf->ctx, &msg, b.head, -1);
+ }
+}
+void udebug_buf_free(struct udebug_buf *buf)
+{
+ struct udebug *ctx = buf->ctx;
+
+ if (!list_empty(&buf->list) && buf->list.prev)
+ list_del(&buf->list);
+
+ if (ctx && ctx->fd.fd >= 0)
+ udebug_buf_msg(buf, CL_MSG_RING_REMOVE);
+
+ munmap(buf->hdr, buf->head_size + 2 * buf->data_size);
+ close(buf->fd);
+ memset(buf, 0, sizeof(*buf));
+}
+
+static void
+__udebug_buf_add(struct udebug *ctx, struct udebug_buf *buf)
+{
+ struct udebug_client_msg msg = {
+ .type = CL_MSG_RING_ADD,
+ .id = buf->id,
+ .ring_size = buf->hdr->ring_size,
+ .data_size = buf->hdr->data_size,
+ };
+ const struct udebug_buf_meta *meta = buf->meta;
+ void *c;
+
+ blob_buf_init(&b, 0);
+ blobmsg_add_string(&b, "name", meta->name);
+ c = blobmsg_open_array(&b, "flags");
+ for (size_t i = 0; i < meta->n_flags; i++) {
+ const struct udebug_buf_flag *flag = &meta->flags[i];
+ void *e = blobmsg_open_array(&b, NULL);
+ blobmsg_add_string(&b, NULL, flag->name);
+ blobmsg_add_u64(&b, NULL, flag->mask);
+ blobmsg_close_array(&b, e);
+ }
+ blobmsg_close_array(&b, c);
+
+ udebug_send_msg(ctx, &msg, b.head, buf->fd);
+}
+
+int udebug_buf_add(struct udebug *ctx, struct udebug_buf *buf,
+ const struct udebug_buf_meta *meta)
+{
+ list_add_tail(&buf->list, &ctx->local_rings);
+ buf->ctx = ctx;
+ buf->meta = meta;
+ buf->id = ctx->next_id++;
+ buf->hdr->format = meta->format;
+ buf->hdr->sub_format = meta->sub_format;
+
+ if (ctx->fd.fd >= 0)
+ __udebug_buf_add(ctx, buf);
+
+ return 0;
+}
+
+void udebug_init(struct udebug *ctx)
+{
+ INIT_LIST_HEAD(&ctx->local_rings);
+ avl_init(&ctx->remote_rings, udebug_id_cmp, true, NULL);
+ ctx->fd.fd = -1;
+ ctx->poll_handle = -1;
+}
+
+static void udebug_reconnect_cb(struct uloop_timeout *t)
+{
+ struct udebug *ctx = container_of(t, struct udebug, reconnect);
+
+ if (udebug_connect(ctx, ctx->socket_path) < 0) {
+ uloop_timeout_set(&ctx->reconnect, 1000);
+ return;
+ }
+
+ udebug_add_uloop(ctx);
+}
+
+void udebug_auto_connect(struct udebug *ctx, const char *path)
+{
+ free(ctx->socket_path);
+ ctx->reconnect.cb = udebug_reconnect_cb;
+ ctx->socket_path = path ? strdup(path) : NULL;
+ if (ctx->fd.fd >= 0)
+ return;
+
+ udebug_reconnect_cb(&ctx->reconnect);
+}
+
+int udebug_connect(struct udebug *ctx, const char *path)
+{
+ struct udebug_remote_buf *rb;
+ struct udebug_buf *buf;
+
+ if (ctx->fd.fd >= 0)
+ close(ctx->fd.fd);
+ ctx->fd.fd = -1;
+
+ if (!path)
+ path = UDEBUG_SOCK_NAME;
+
+ ctx->fd.fd = usock(USOCK_UNIX, path, NULL);
+ if (ctx->fd.fd < 0)
+ return -1;
+
+ list_for_each_entry(buf, &ctx->local_rings, list)
+ __udebug_buf_add(ctx, buf);
+
+ avl_for_each_element(&ctx->remote_rings, rb, node) {
+ if (!rb->poll)
+ continue;
+
+ rb->poll = false;
+ udebug_remote_buf_set_poll(ctx, rb, true);
+ }
+
+ return 0;
+}
+
+static bool
+udebug_recv_msg(struct udebug *ctx, struct udebug_client_msg *msg, int *fd,
+ bool wait)
+{
+ struct iovec iov = {
+ .iov_base = msg,
+ .iov_len = sizeof(*msg)
+ };
+ int ret;
+
+ ret = recv_retry(ctx->fd.fd, &iov, wait, fd);
+ if (ret == -2)
+ __udebug_disconnect(ctx, true);
+
+ return ret == sizeof(*msg);
+}
+
+struct udebug_client_msg *__udebug_poll(struct udebug *ctx, int *fd, bool wait)
+{
+ static struct udebug_client_msg msg = {};
+
+ while (udebug_recv_msg(ctx, &msg, fd, wait)) {
+ struct udebug_remote_buf *rb;
+ void *key;
+
+ if (msg.type != CL_MSG_RING_NOTIFY)
+ return &msg;
+
+ if (fd && *fd >= 0)
+ close(*fd);
+
+ if (!ctx->notify_cb)
+ continue;
+
+ key = (void *)(uintptr_t)msg.id;
+ rb = avl_find_element(&ctx->remote_rings, key, rb, node);
+ if (!rb || !rb->poll)
+ continue;
+
+ if (ctx->poll_handle >= 0)
+ __atomic_fetch_or(&rb->buf.hdr->notify,
+ 1UL << ctx->poll_handle,
+ __ATOMIC_RELAXED);
+ ctx->notify_cb(ctx, rb);
+ }
+
+ return NULL;
+}
+
+void udebug_poll(struct udebug *ctx)
+{
+ while (__udebug_poll(ctx, NULL, false));
+}
+
+static void udebug_fd_cb(struct uloop_fd *fd, unsigned int events)
+{
+ struct udebug *ctx = container_of(fd, struct udebug, fd);
+
+ if (fd->eof)
+ __udebug_disconnect(ctx, true);
+
+ udebug_poll(ctx);
+}
+
+void udebug_add_uloop(struct udebug *ctx)
+{
+ if (ctx->fd.registered)
+ return;
+
+ ctx->fd.cb = udebug_fd_cb;
+ uloop_fd_add(&ctx->fd, ULOOP_READ);
+}
+
+void udebug_free(struct udebug *ctx)
+{
+ struct udebug_remote_buf *rb, *tmp;
+ struct udebug_buf *buf;
+
+ free(ctx->socket_path);
+ ctx->socket_path = NULL;
+
+ __udebug_disconnect(ctx, false);
+ uloop_timeout_cancel(&ctx->reconnect);
+
+ while (!list_empty(&ctx->local_rings)) {
+ buf = list_first_entry(&ctx->local_rings, struct udebug_buf, list);
+ udebug_buf_free(buf);
+ }
+
+ avl_for_each_element_safe(&ctx->remote_rings, rb, node, tmp)
+ udebug_remote_buf_unmap(ctx, rb);
+}