kernel_optimize_test/drivers/hv/hv_utils_transport.c

351 lines
7.6 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* Kernel/userspace transport abstraction for Hyper-V util driver.
*
* Copyright (C) 2015, Vitaly Kuznetsov <vkuznets@redhat.com>
*/
#include <linux/slab.h>
#include <linux/fs.h>
#include <linux/poll.h>
#include "hyperv_vmbus.h"
#include "hv_utils_transport.h"
static DEFINE_SPINLOCK(hvt_list_lock);
static struct list_head hvt_list = LIST_HEAD_INIT(hvt_list);
static void hvt_reset(struct hvutil_transport *hvt)
{
kfree(hvt->outmsg);
hvt->outmsg = NULL;
hvt->outmsg_len = 0;
if (hvt->on_reset)
hvt->on_reset();
}
static ssize_t hvt_op_read(struct file *file, char __user *buf,
size_t count, loff_t *ppos)
{
struct hvutil_transport *hvt;
int ret;
hvt = container_of(file->f_op, struct hvutil_transport, fops);
if (wait_event_interruptible(hvt->outmsg_q, hvt->outmsg_len > 0 ||
hvt->mode != HVUTIL_TRANSPORT_CHARDEV))
return -EINTR;
mutex_lock(&hvt->lock);
if (hvt->mode == HVUTIL_TRANSPORT_DESTROY) {
ret = -EBADF;
goto out_unlock;
}
if (!hvt->outmsg) {
ret = -EAGAIN;
goto out_unlock;
}
if (count < hvt->outmsg_len) {
ret = -EINVAL;
goto out_unlock;
}
if (!copy_to_user(buf, hvt->outmsg, hvt->outmsg_len))
ret = hvt->outmsg_len;
else
ret = -EFAULT;
kfree(hvt->outmsg);
hvt->outmsg = NULL;
hvt->outmsg_len = 0;
Drivers: hv: utils: fix a race on userspace daemons registration Background: userspace daemons registration protocol for Hyper-V utilities drivers has two steps: 1) daemon writes its own version to kernel 2) kernel reads it and replies with module version at this point we consider the handshake procedure being completed and we do hv_poll_channel() transitioning the utility device to HVUTIL_READY state. At this point we're ready to handle messages from kernel. When hvutil_transport is in HVUTIL_TRANSPORT_CHARDEV mode we have a single buffer for outgoing message. hvutil_transport_send() puts to this buffer and till the buffer is cleared with hvt_op_read() returns -EFAULT to all consequent calls. Host<->guest protocol guarantees there is no more than one request at a time and we will not get new requests till we reply to the previous one so this single message buffer is enough. Now to the race. When we finish negotiation procedure and send kernel module version to userspace with hvutil_transport_send() it goes into the above mentioned buffer and if the daemon is slow enough to read it from there we can get a collision when a request from the host comes, we won't be able to put anything to the buffer so the request will be lost. To solve the issue we need to know when the negotiation is really done (when the version message is read by the daemon) and transition to HVUTIL_READY state after this happens. Implement a callback on read to support this. Old style netlink communication is not affected by the change, we don't really know when these messages are delivered but we don't have a single message buffer there. Reported-by: Barry Davis <barry_davis@stormagic.com> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> Signed-off-by: K. Y. Srinivasan <kys@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-06-10 08:08:57 +08:00
if (hvt->on_read)
hvt->on_read();
hvt->on_read = NULL;
out_unlock:
mutex_unlock(&hvt->lock);
return ret;
}
static ssize_t hvt_op_write(struct file *file, const char __user *buf,
size_t count, loff_t *ppos)
{
struct hvutil_transport *hvt;
u8 *inmsg;
int ret;
hvt = container_of(file->f_op, struct hvutil_transport, fops);
inmsg = memdup_user(buf, count);
if (IS_ERR(inmsg))
return PTR_ERR(inmsg);
if (hvt->mode == HVUTIL_TRANSPORT_DESTROY)
ret = -EBADF;
else
ret = hvt->on_msg(inmsg, count);
kfree(inmsg);
return ret ? ret : count;
}
static __poll_t hvt_op_poll(struct file *file, poll_table *wait)
{
struct hvutil_transport *hvt;
hvt = container_of(file->f_op, struct hvutil_transport, fops);
poll_wait(file, &hvt->outmsg_q, wait);
if (hvt->mode == HVUTIL_TRANSPORT_DESTROY)
return EPOLLERR | EPOLLHUP;
if (hvt->outmsg_len > 0)
return EPOLLIN | EPOLLRDNORM;
return 0;
}
static int hvt_op_open(struct inode *inode, struct file *file)
{
struct hvutil_transport *hvt;
int ret = 0;
bool issue_reset = false;
hvt = container_of(file->f_op, struct hvutil_transport, fops);
mutex_lock(&hvt->lock);
if (hvt->mode == HVUTIL_TRANSPORT_DESTROY) {
ret = -EBADF;
} else if (hvt->mode == HVUTIL_TRANSPORT_INIT) {
/*
* Switching to CHARDEV mode. We switch bach to INIT when
* device gets released.
*/
hvt->mode = HVUTIL_TRANSPORT_CHARDEV;
}
else if (hvt->mode == HVUTIL_TRANSPORT_NETLINK) {
/*
* We're switching from netlink communication to using char
* device. Issue the reset first.
*/
issue_reset = true;
hvt->mode = HVUTIL_TRANSPORT_CHARDEV;
} else {
ret = -EBUSY;
}
if (issue_reset)
hvt_reset(hvt);
mutex_unlock(&hvt->lock);
return ret;
}
static void hvt_transport_free(struct hvutil_transport *hvt)
{
misc_deregister(&hvt->mdev);
kfree(hvt->outmsg);
kfree(hvt);
}
static int hvt_op_release(struct inode *inode, struct file *file)
{
struct hvutil_transport *hvt;
int mode_old;
hvt = container_of(file->f_op, struct hvutil_transport, fops);
mutex_lock(&hvt->lock);
mode_old = hvt->mode;
if (hvt->mode != HVUTIL_TRANSPORT_DESTROY)
hvt->mode = HVUTIL_TRANSPORT_INIT;
/*
* Cleanup message buffers to avoid spurious messages when the daemon
* connects back.
*/
hvt_reset(hvt);
if (mode_old == HVUTIL_TRANSPORT_DESTROY)
complete(&hvt->release);
mutex_unlock(&hvt->lock);
return 0;
}
static void hvt_cn_callback(struct cn_msg *msg, struct netlink_skb_parms *nsp)
{
struct hvutil_transport *hvt, *hvt_found = NULL;
spin_lock(&hvt_list_lock);
list_for_each_entry(hvt, &hvt_list, list) {
if (hvt->cn_id.idx == msg->id.idx &&
hvt->cn_id.val == msg->id.val) {
hvt_found = hvt;
break;
}
}
spin_unlock(&hvt_list_lock);
if (!hvt_found) {
pr_warn("hvt_cn_callback: spurious message received!\n");
return;
}
/*
* Switching to NETLINK mode. Switching to CHARDEV happens when someone
* opens the device.
*/
mutex_lock(&hvt->lock);
if (hvt->mode == HVUTIL_TRANSPORT_INIT)
hvt->mode = HVUTIL_TRANSPORT_NETLINK;
if (hvt->mode == HVUTIL_TRANSPORT_NETLINK)
hvt_found->on_msg(msg->data, msg->len);
else
pr_warn("hvt_cn_callback: unexpected netlink message!\n");
mutex_unlock(&hvt->lock);
}
Drivers: hv: utils: fix a race on userspace daemons registration Background: userspace daemons registration protocol for Hyper-V utilities drivers has two steps: 1) daemon writes its own version to kernel 2) kernel reads it and replies with module version at this point we consider the handshake procedure being completed and we do hv_poll_channel() transitioning the utility device to HVUTIL_READY state. At this point we're ready to handle messages from kernel. When hvutil_transport is in HVUTIL_TRANSPORT_CHARDEV mode we have a single buffer for outgoing message. hvutil_transport_send() puts to this buffer and till the buffer is cleared with hvt_op_read() returns -EFAULT to all consequent calls. Host<->guest protocol guarantees there is no more than one request at a time and we will not get new requests till we reply to the previous one so this single message buffer is enough. Now to the race. When we finish negotiation procedure and send kernel module version to userspace with hvutil_transport_send() it goes into the above mentioned buffer and if the daemon is slow enough to read it from there we can get a collision when a request from the host comes, we won't be able to put anything to the buffer so the request will be lost. To solve the issue we need to know when the negotiation is really done (when the version message is read by the daemon) and transition to HVUTIL_READY state after this happens. Implement a callback on read to support this. Old style netlink communication is not affected by the change, we don't really know when these messages are delivered but we don't have a single message buffer there. Reported-by: Barry Davis <barry_davis@stormagic.com> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> Signed-off-by: K. Y. Srinivasan <kys@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-06-10 08:08:57 +08:00
int hvutil_transport_send(struct hvutil_transport *hvt, void *msg, int len,
void (*on_read_cb)(void))
{
struct cn_msg *cn_msg;
int ret = 0;
if (hvt->mode == HVUTIL_TRANSPORT_INIT ||
hvt->mode == HVUTIL_TRANSPORT_DESTROY) {
return -EINVAL;
} else if (hvt->mode == HVUTIL_TRANSPORT_NETLINK) {
cn_msg = kzalloc(sizeof(*cn_msg) + len, GFP_ATOMIC);
if (!cn_msg)
return -ENOMEM;
cn_msg->id.idx = hvt->cn_id.idx;
cn_msg->id.val = hvt->cn_id.val;
cn_msg->len = len;
memcpy(cn_msg->data, msg, len);
ret = cn_netlink_send(cn_msg, 0, 0, GFP_ATOMIC);
kfree(cn_msg);
Drivers: hv: utils: fix a race on userspace daemons registration Background: userspace daemons registration protocol for Hyper-V utilities drivers has two steps: 1) daemon writes its own version to kernel 2) kernel reads it and replies with module version at this point we consider the handshake procedure being completed and we do hv_poll_channel() transitioning the utility device to HVUTIL_READY state. At this point we're ready to handle messages from kernel. When hvutil_transport is in HVUTIL_TRANSPORT_CHARDEV mode we have a single buffer for outgoing message. hvutil_transport_send() puts to this buffer and till the buffer is cleared with hvt_op_read() returns -EFAULT to all consequent calls. Host<->guest protocol guarantees there is no more than one request at a time and we will not get new requests till we reply to the previous one so this single message buffer is enough. Now to the race. When we finish negotiation procedure and send kernel module version to userspace with hvutil_transport_send() it goes into the above mentioned buffer and if the daemon is slow enough to read it from there we can get a collision when a request from the host comes, we won't be able to put anything to the buffer so the request will be lost. To solve the issue we need to know when the negotiation is really done (when the version message is read by the daemon) and transition to HVUTIL_READY state after this happens. Implement a callback on read to support this. Old style netlink communication is not affected by the change, we don't really know when these messages are delivered but we don't have a single message buffer there. Reported-by: Barry Davis <barry_davis@stormagic.com> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> Signed-off-by: K. Y. Srinivasan <kys@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-06-10 08:08:57 +08:00
/*
* We don't know when netlink messages are delivered but unlike
* in CHARDEV mode we're not blocked and we can send next
* messages right away.
*/
if (on_read_cb)
on_read_cb();
return ret;
}
/* HVUTIL_TRANSPORT_CHARDEV */
mutex_lock(&hvt->lock);
if (hvt->mode != HVUTIL_TRANSPORT_CHARDEV) {
ret = -EINVAL;
goto out_unlock;
}
if (hvt->outmsg) {
/* Previous message wasn't received */
ret = -EFAULT;
goto out_unlock;
}
hvt->outmsg = kzalloc(len, GFP_KERNEL);
if (hvt->outmsg) {
memcpy(hvt->outmsg, msg, len);
hvt->outmsg_len = len;
Drivers: hv: utils: fix a race on userspace daemons registration Background: userspace daemons registration protocol for Hyper-V utilities drivers has two steps: 1) daemon writes its own version to kernel 2) kernel reads it and replies with module version at this point we consider the handshake procedure being completed and we do hv_poll_channel() transitioning the utility device to HVUTIL_READY state. At this point we're ready to handle messages from kernel. When hvutil_transport is in HVUTIL_TRANSPORT_CHARDEV mode we have a single buffer for outgoing message. hvutil_transport_send() puts to this buffer and till the buffer is cleared with hvt_op_read() returns -EFAULT to all consequent calls. Host<->guest protocol guarantees there is no more than one request at a time and we will not get new requests till we reply to the previous one so this single message buffer is enough. Now to the race. When we finish negotiation procedure and send kernel module version to userspace with hvutil_transport_send() it goes into the above mentioned buffer and if the daemon is slow enough to read it from there we can get a collision when a request from the host comes, we won't be able to put anything to the buffer so the request will be lost. To solve the issue we need to know when the negotiation is really done (when the version message is read by the daemon) and transition to HVUTIL_READY state after this happens. Implement a callback on read to support this. Old style netlink communication is not affected by the change, we don't really know when these messages are delivered but we don't have a single message buffer there. Reported-by: Barry Davis <barry_davis@stormagic.com> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> Signed-off-by: K. Y. Srinivasan <kys@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-06-10 08:08:57 +08:00
hvt->on_read = on_read_cb;
wake_up_interruptible(&hvt->outmsg_q);
} else
ret = -ENOMEM;
out_unlock:
mutex_unlock(&hvt->lock);
return ret;
}
struct hvutil_transport *hvutil_transport_init(const char *name,
u32 cn_idx, u32 cn_val,
int (*on_msg)(void *, int),
void (*on_reset)(void))
{
struct hvutil_transport *hvt;
hvt = kzalloc(sizeof(*hvt), GFP_KERNEL);
if (!hvt)
return NULL;
hvt->cn_id.idx = cn_idx;
hvt->cn_id.val = cn_val;
hvt->mdev.minor = MISC_DYNAMIC_MINOR;
hvt->mdev.name = name;
hvt->fops.owner = THIS_MODULE;
hvt->fops.read = hvt_op_read;
hvt->fops.write = hvt_op_write;
hvt->fops.poll = hvt_op_poll;
hvt->fops.open = hvt_op_open;
hvt->fops.release = hvt_op_release;
hvt->mdev.fops = &hvt->fops;
init_waitqueue_head(&hvt->outmsg_q);
mutex_init(&hvt->lock);
init_completion(&hvt->release);
spin_lock(&hvt_list_lock);
list_add(&hvt->list, &hvt_list);
spin_unlock(&hvt_list_lock);
hvt->on_msg = on_msg;
hvt->on_reset = on_reset;
if (misc_register(&hvt->mdev))
goto err_free_hvt;
/* Use cn_id.idx/cn_id.val to determine if we need to setup netlink */
if (hvt->cn_id.idx > 0 && hvt->cn_id.val > 0 &&
cn_add_callback(&hvt->cn_id, name, hvt_cn_callback))
goto err_free_hvt;
return hvt;
err_free_hvt:
spin_lock(&hvt_list_lock);
list_del(&hvt->list);
spin_unlock(&hvt_list_lock);
kfree(hvt);
return NULL;
}
void hvutil_transport_destroy(struct hvutil_transport *hvt)
{
int mode_old;
mutex_lock(&hvt->lock);
mode_old = hvt->mode;
hvt->mode = HVUTIL_TRANSPORT_DESTROY;
wake_up_interruptible(&hvt->outmsg_q);
mutex_unlock(&hvt->lock);
/*
* In case we were in 'chardev' mode we still have an open fd so we
* have to defer freeing the device. Netlink interface can be freed
* now.
*/
spin_lock(&hvt_list_lock);
list_del(&hvt->list);
spin_unlock(&hvt_list_lock);
if (hvt->cn_id.idx > 0 && hvt->cn_id.val > 0)
cn_del_callback(&hvt->cn_id);
if (mode_old == HVUTIL_TRANSPORT_CHARDEV)
wait_for_completion(&hvt->release);
hvt_transport_free(hvt);
}