[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Minios-devel] [UNIKRAFT PATCH v2 04/10] plat/xen: Add support for communication with Xenstore daemon



Hi Yuri,

Please see my comments inline.

On 08/29/2018 03:38 PM, Yuri Volchkov wrote:
> Hey Costin.
> 
> I found just a couple of important things. But both of them are really
> tiny.
> 
> The comments are inline
> 
> -Yuri.
> 
> Costin Lupu <costin.lupu@xxxxxxxxx> writes:
> 
>> Add support for communication with Xenstore daemon via the shared
>> page. In Unikraft, the incoming messages are processed by the
>> Xenstore thread.
>>
>> Signed-off-by: Costin Lupu <costin.lupu@xxxxxxxxx>
>> ---
>>  plat/xen/Makefile.uk       |   1 +
>>  plat/xen/xenbus/xenbus.c   |   8 +
>>  plat/xen/xenbus/xs_comms.c | 535 
>> +++++++++++++++++++++++++++++++++++++++++++++
>>  plat/xen/xenbus/xs_comms.h |  64 ++++++
>>  4 files changed, 608 insertions(+)
>>  create mode 100644 plat/xen/xenbus/xs_comms.c
>>  create mode 100644 plat/xen/xenbus/xs_comms.h
>>
>> diff --git a/plat/xen/Makefile.uk b/plat/xen/Makefile.uk
>> index b8c70e1..2703a54 100644
>> --- a/plat/xen/Makefile.uk
>> +++ b/plat/xen/Makefile.uk
>> @@ -79,4 +79,5 @@ LIBXENBUS_ASINCLUDES-y         += 
>> $(LIBXENPLAT_ASINCLUDES-y)
>>  LIBXENBUS_CFLAGS-y             += $(LIBXENPLAT_CFLAGS-y)
>>  LIBXENBUS_CINCLUDES-y          += $(LIBXENPLAT_CINCLUDES-y)
>>  LIBXENBUS_SRCS-y               += $(LIBXENPLAT_BASE)/xenbus/xenbus.c
>> +LIBXENBUS_SRCS-y               += $(LIBXENPLAT_BASE)/xenbus/xs_comms.c
>>  endif
>> diff --git a/plat/xen/xenbus/xenbus.c b/plat/xen/xenbus/xenbus.c
>> index 1bc57c3..a20546b 100644
>> --- a/plat/xen/xenbus/xenbus.c
>> +++ b/plat/xen/xenbus/xenbus.c
>> @@ -43,6 +43,7 @@
>>  #include <uk/errptr.h>
>>  #include <uk/assert.h>
>>  #include <xenbus/xenbus.h>
>> +#include "xs_comms.h"
>>  
>>  static struct xenbus_handler xbh;
>>  
>> @@ -86,6 +87,13 @@ static int xenbus_init(struct uk_alloc *a)
>>  
>>      xbh.a = a;
>>  
>> +    ret = xs_comms_init();
>> +    if (ret) {
>> +            uk_printd(DLVL_ERR,
>> +                    "Error initializing Xenstore communication.");
>> +            return ret;
>> +    }
>> +
>>      UK_TAILQ_FOREACH_SAFE(drv, &xbh.drv_list, next, drv_next) {
>>              if (drv->init) {
>>                      ret = drv->init(a);
>> diff --git a/plat/xen/xenbus/xs_comms.c b/plat/xen/xenbus/xs_comms.c
>> new file mode 100644
>> index 0000000..df3739d
>> --- /dev/null
>> +++ b/plat/xen/xenbus/xs_comms.c
>> @@ -0,0 +1,535 @@
>> +/* SPDX-License-Identifier: BSD-3-Clause */
>> +/*
>> + * Authors: Steven Smith (sos22@xxxxxxxxx)
>> + *          Grzegorz Milos (gm281@xxxxxxxxx)
>> + *          John D. Ramsdell
>> + *          Costin Lupu <costin.lupu@xxxxxxxxx>
>> + *
>> + * Copyright (c) 2006, Cambridge University
>> + *               2018, NEC Europe Ltd., NEC Corporation. All rights 
>> reserved.
>> + *
>> + * Redistribution and use in source and binary forms, with or without
>> + * modification, are permitted provided that the following conditions
>> + * are met:
>> + *
>> + * 1. Redistributions of source code must retain the above copyright
>> + *    notice, this list of conditions and the following disclaimer.
>> + * 2. Redistributions in binary form must reproduce the above copyright
>> + *    notice, this list of conditions and the following disclaimer in the
>> + *    documentation and/or other materials provided with the distribution.
>> + * 3. Neither the name of the copyright holder nor the names of its
>> + *    contributors may be used to endorse or promote products derived from
>> + *    this software without specific prior written permission.
>> + *
>> + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS 
>> IS"
>> + * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
>> + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 
>> PURPOSE
>> + * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
>> + * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
>> + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
>> + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
>> + * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
>> + * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
>> + * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF 
>> THE
>> + * POSSIBILITY OF SUCH DAMAGE.
>> + *
>> + * THIS HEADER MAY NOT BE EXTRACTED OR MODIFIED IN ANY WAY.
>> + */
>> +/*
>> + * Communication with Xenstore
>> + * Ported from Mini-OS xenbus.c
>> + */
>> +
>> +#include <string.h>
>> +#include <uk/errptr.h>
>> +#include <uk/bitmap.h>
>> +#include <uk/wait.h>
>> +#include <uk/arch/spinlock.h>
>> +#include <common/events.h>
>> +#include <xen-x86/mm.h>
>> +#include <xen-x86/setup.h>
>> +#include "xs_comms.h"
>> +
>> +
>> +/*
>> + * Xenstore handler structure
>> + */
>> +struct xs_handler {
>> +    /**< Communication: event channel */
>> +    evtchn_port_t evtchn;
>> +    /**< Communication: shared memory */
>> +    struct xenstore_domain_interface *buf;
>> +    /**< Thread processing incoming xs replies */
>> +    struct uk_thread *thread;
>> +    /**< Waiting queue for notifying incoming xs replies */
>> +    struct uk_waitq waitq;
>> +};
>> +
>> +static struct xs_handler xsh = {
>> +    .waitq = __WAIT_QUEUE_INITIALIZER(xsh.waitq),
>> +};
>> +
>> +static int xs_avail_space_for_read(unsigned int req_size)
>> +{
>> +    return (xsh.buf->rsp_prod - xsh.buf->rsp_cons >= req_size);
>> +}
>> +
>> +static int xs_avail_space_for_write(unsigned int req_size)
>> +{
>> +    return (xsh.buf->req_prod - xsh.buf->req_cons +
>> +            req_size <= XENSTORE_RING_SIZE);
>> +}
>> +
>> +/*
>> + * In-flight request structure.
>> + */
>> +struct xs_request {
>> +    /**< Waiting queue for incoming reply notification */
>> +    struct uk_waitq waitq;
>> +    /**< Request header */
>> +    struct xsd_sockmsg hdr;
>> +    /**< Received reply */
>> +    struct {
>> +            /**< Reply string + size */
>> +            struct xs_iovec iovec;
>> +            /**< Error number */
>> +            int errornum;
>> +            /**< Non-zero for incoming replies */
>> +            int recvd;
>> +    } reply;
>> +};
>> +
>> +/*
>> + * Pool of in-flight requests.
>> + * Request IDs are reused, hence the limited set of entries.
>> + *
>> + * TODO sequential IDs
>> + */
>> +struct xs_request_pool {
>> +    /**< Number of live requests */
>> +    __u32 num_live;
>> +    /**< Lock */
>> +    spinlock_t lock;
>> +    /**< Waiting queue for 'not-full' notifications */
>> +    struct uk_waitq waitq;
>> +
>> +    /* Map size is power of 2 */
>> +#define REQID_MAP_SHIFT  5
>> +#define REQID_MAP_SIZE   (1 << REQID_MAP_SHIFT)
>> +    unsigned long entries_bm[BITS_TO_LONGS(REQID_MAP_SIZE) * sizeof(long)];
>> +    /**< Entries */
>> +    struct xs_request entries[REQID_MAP_SIZE];
>> +};
>> +
>> +static struct xs_request_pool xs_req_pool;
>> +
>> +static void xs_request_pool_init(struct xs_request_pool *pool)
>> +{
>> +    struct xs_request *xs_req;
>> +
>> +    pool->num_live = 0;
>> +    ukarch_spin_lock_init(&pool->lock);
>> +    uk_waitq_init(&pool->waitq);
>> +    bitmap_zero(pool->entries_bm, REQID_MAP_SIZE);
>> +    for (int i = 0; i < REQID_MAP_SIZE; i++) {
>> +            xs_req = &pool->entries[i];
>> +            xs_req->hdr.req_id = i;
>> +            uk_waitq_init(&xs_req->waitq);
>> +    }
>> +}
>> +
>> +/*
>> + * Allocate an identifier for a Xenstore request.
>> + * Blocks if none are available.
>> + */
>> +static struct xs_request *xs_request_get(void)
>> +{
>> +    unsigned long entry_idx;
>> +
>> +    /* wait for an available entry */
>> +    while (1) {
>> +            ukarch_spin_lock(&xs_req_pool.lock);
>> +
>> +            if (xs_req_pool.num_live < REQID_MAP_SIZE)
>> +                    break;
>> +
>> +            ukarch_spin_unlock(&xs_req_pool.lock);
>> +
>> +            uk_waitq_wait_event(&xs_req_pool.waitq,
>> +                    (xs_req_pool.num_live < REQID_MAP_SIZE));
>> +    }
>> +
>> +    /* find an available entry */
>> +    entry_idx =
>> +            find_first_zero_bit(xs_req_pool.entries_bm, REQID_MAP_SIZE);
>> +
>> +    set_bit(entry_idx, xs_req_pool.entries_bm);
>> +    xs_req_pool.num_live++;
>> +
>> +    ukarch_spin_unlock(&xs_req_pool.lock);
>> +
>> +    return &xs_req_pool.entries[entry_idx];
>> +}
>> +
>> +/* Release a request identifier */
>> +static void xs_request_put(struct xs_request *xs_req)
>> +{
>> +    __u32 reqid = xs_req->hdr.req_id;
>> +
>> +    ukarch_spin_lock(&xs_req_pool.lock);
>> +
>> +    UK_ASSERT(test_bit(reqid, xs_req_pool.entries_bm) == 1);
>> +
>> +    clear_bit(reqid, xs_req_pool.entries_bm);
>> +    xs_req_pool.num_live--;
>> +
>> +    if (xs_req_pool.num_live == 0 ||
>> +            xs_req_pool.num_live == REQID_MAP_SIZE - 1)
> I understand the second condition, but why we need to wake up waiter if
> num_live == 0?

The first condition is used for suspending. We should wait until all
requests were answered, that's why we're waiting for num_live to get to
zero. Maybe it would be better to remove the first condition and add it
back when we'll have the suspend/resume functionality ready.

> And, in the second condition, you probably meant "REQID_MAP_SIZE - 2",
> not "- 1". We want to wake up waiters if all request from the pull _WAS_
> in-flight.

All requests are in-flight iff num_live == REQID_MAP_SIZE (which btw I
have to rename).

> So, if we entered xs_request_put at the point of time, when there was no
> free requests in the pull, num_live will be "REQID_MAP_SIZE - 1". After
> the decrementing, it becomes "REQID_MAP_SIZE - 2".

I'm afraid here it should be REQID_MAP_SIZE - 1, in other words if we
have at least on slot free then we can signal that it's "non-full".
> 
> 
>> +            uk_waitq_wake_up(&xs_req_pool.waitq);
>> +
>> +    ukarch_spin_unlock(&xs_req_pool.lock);
>> +}
>> +
>> +/*
>> + * Send request to Xenstore. A request is made of multiple iovecs which are
>> + * preceded by a single iovec referencing the request header. The iovecs are
>> + * seen by Xenstore as if sent atomically. This can block.
>> + *
>> + * TODO A different thread (e.g. Xenbus thread) should write the messages if
>> + * we expect high loads of requests.
>> + */
>> +static int xs_msg_write(struct xsd_sockmsg *xsd_req,
>> +    const struct xs_iovec *iovec)
>> +{
>> +    XENSTORE_RING_IDX prod;
>> +    const struct xs_iovec *crnt_iovec;
>> +    struct xs_iovec hdr_iovec;
>> +    unsigned int req_size, req_off;
>> +    unsigned int buf_off;
>> +    unsigned int this_chunk_len;
>> +    int rc;
>> +
>> +    req_size = sizeof(*xsd_req) + xsd_req->len;
>> +    if (req_size > XENSTORE_RING_SIZE)
>> +            return -ENOSPC;
>> +
>> +    hdr_iovec.data = xsd_req;
>> +    hdr_iovec.len  = sizeof(*xsd_req);
>> +
>> +    /* The batched iovecs are preceded by a single header. */
>> +    crnt_iovec = &hdr_iovec;
>> +
>> +    /*
>> +     * Wait for the ring to drain to the point where
>> +     * we can send the message.
>> +     */
>> +    while (!xs_avail_space_for_write(req_size)) {
>> +            /* Wait for there to be space on the ring */
>> +            uk_printd(DLVL_EXTRA,
>> +                    "prod %d, len %d, cons %d, size %d; waiting.\n",
>> +                    xsh.buf->req_prod, req_size,
>> +                    xsh.buf->req_cons, XENSTORE_RING_SIZE);
>> +
>> +            uk_waitq_wait_event(&xsh.waitq,
>> +                    xs_avail_space_for_write(req_size));
>> +            uk_printd(DLVL_EXTRA, "Back from wait.\n");
>> +    }
> This will work with cooperative scheduler for sure. But I guess that
> here will be a problem in case of preemptive one. If we was preempted at
> this point, and another thread wants to write to the ring buffer too, we
> are screwed. For example, the other thread could use the free space in
> the ring buffer, and when we get cpu time back, we will continue with
> confidence that there is enough space for us.
> 
> So far we was ignoring places like this. How about we start to leave at
> least notes. For example /* TODO: disable preemption here */.
> 
> Or, a dedicated thread would solve this problem, as discussed in v1 and
> as you mentioned in the comment above this function.

In deed, for preemptive scheduling we would have a problem here. I guess
the best solution would be to delegate the writes to the xenstore thread.

>> +
>> +    /* We must write requests after reading the consumer index. */
>> +    mb();
> Maybe rmb() here is enough? Or do we actually need a full barrier?

Before the barrier we have a read operation (reading the consumer
index), while after we have write operations (writing the requests). So
we need to order the writes after reads with full barrier.

>> +
>> +    /*
>> +     * We're now guaranteed to be able to send the message
>> +     * without overflowing the ring. Do so.
>> +     */
>> +
>> +    prod = xsh.buf->req_prod;
>> +    req_off = 0;
>> +    buf_off = 0;
>> +    while (req_off < req_size) {
>> +            this_chunk_len = MIN(crnt_iovec->len - buf_off,
>> +                    XENSTORE_RING_SIZE - MASK_XENSTORE_IDX(prod));
>> +
>> +            memcpy(
>> +                    (char *) xsh.buf->req + MASK_XENSTORE_IDX(prod),
>> +                    (char *) crnt_iovec->data + buf_off,
>> +                    this_chunk_len
>> +            );
>> +
>> +            prod += this_chunk_len;
>> +            req_off += this_chunk_len;
>> +            buf_off += this_chunk_len;
>> +
>> +            if (buf_off == crnt_iovec->len) {
>> +                    buf_off = 0;
>> +                    if (crnt_iovec == &hdr_iovec)
>> +                            crnt_iovec = iovec;
>> +                    else
>> +                            crnt_iovec++;
>> +            }
>> +    }
>> +
>> +    uk_printd(DLVL_EXTRA, "Complete main loop of %s.\n", __func__);
>> +    UK_ASSERT(buf_off == 0);
>> +    UK_ASSERT(req_off == req_size);
>> +    UK_ASSERT(prod <= xsh.buf->req_cons + XENSTORE_RING_SIZE);
>> +
>> +    /* Remote must see entire message before updating indexes */
>> +    wmb();
>> +
>> +    xsh.buf->req_prod += req_size;
>> +
>> +    /* Send evtchn to notify remote */
>> +    rc = notify_remote_via_evtchn(xsh.evtchn);
>> +    UK_ASSERT(rc == 0);
>> +
>> +    return 0;
>> +}
>> +
>> +int xs_msg_reply(enum xsd_sockmsg_type msg_type, xenbus_transaction_t xbt,
>> +    const struct xs_iovec *req_iovecs, int req_iovecs_num,
>> +    struct xs_iovec *rep_iovec)
>> +{
>> +    struct xs_request *xs_req;
>> +    int err;
>> +
>> +    if (req_iovecs == NULL)
>> +            return -EINVAL;
>> +
>> +    xs_req = xs_request_get();
>> +    xs_req->hdr.type = msg_type;
>> +    /* req_id was on pool init  */
>> +    xs_req->hdr.tx_id = xbt;
>> +    xs_req->hdr.len = 0;
>> +    for (int i = 0; i < req_iovecs_num; i++)
>> +            xs_req->hdr.len += req_iovecs[i].len;
>> +
>> +    xs_req->reply.recvd = 0;
>> +
>> +    /* send the request */
>> +    err = xs_msg_write(&xs_req->hdr, req_iovecs);
>> +    if (err)
>> +            goto out;
>> +
>> +    /* wait reply */
>> +    uk_waitq_wait_event(&xs_req->waitq,
>> +            xs_req->reply.recvd != 0);
>> +
>> +    err = xs_req->reply.errornum;
>> +    if (err == 0) {
>> +            if (rep_iovec)
>> +                    *rep_iovec = xs_req->reply.iovec;
>> +            else
>> +                    free(xs_req->reply.iovec.data);
>> +    }
>> +
>> +out:
>> +    xs_request_put(xs_req);
>> +
>> +    return (-err);
> xs_msg_write already returns a negative error code. This line will make
> it positive. I guess you want to invert it when reading from
> xs_req->reply.errornum, a few lines above, instead of doing it here

Yes, good catch!

>> +}
>> +
>> +/*
>> + * Converts a Xenstore reply error to a positive error number.
>> + * Returns 0 if the reply is successful.
>> + */
>> +static int reply_to_errno(const char *reply)
>> +{
>> +    int err = 0;
>> +
>> +    for (int i = 0; i < (int) ARRAY_SIZE(xsd_errors); i++) {
>> +            if (!strcmp(reply, xsd_errors[i].errstring)) {
>> +                    err = xsd_errors[i].errnum;
>> +                    goto out;
>> +            }
>> +    }
>> +
>> +    uk_printd(DLVL_WARN, "Unknown Xenstore error: %s\n", reply);
>> +    err = EINVAL;
>> +
>> +out:
>> +    return err;
>> +}
>> +
>> +/* Process an incoming xs reply */
>> +static void process_reply(struct xsd_sockmsg *hdr, char *payload)
>> +{
>> +    struct xs_request *xs_req;
>> +
>> +    if (!test_bit(hdr->req_id, xs_req_pool.entries_bm)) {
>> +            uk_printd(DLVL_WARN, "Invalid reply id=%d", hdr->req_id);
>> +            free(payload);
>> +            return;
>> +    }
>> +
>> +    xs_req = &xs_req_pool.entries[hdr->req_id];
>> +
>> +    if (hdr->type == XS_ERROR) {
>> +            xs_req->reply.errornum = reply_to_errno(payload);
>> +            free(payload);
>> +
>> +    } else if (hdr->type != xs_req->hdr.type) {
>> +            uk_printd(DLVL_WARN, "Mismatching message type: %d", hdr->type);
>> +            free(payload);
>> +            return;
>> +
>> +    } else {
>> +            /* set reply */
>> +            xs_req->reply.iovec.data = payload;
>> +            xs_req->reply.iovec.len = hdr->len;
>> +            xs_req->reply.errornum = 0;
>> +    }
>> +
>> +    xs_req->reply.recvd = 1;
>> +
>> +    /* notify waiting requester */
>> +    uk_waitq_wake_up(&xs_req->waitq);
>> +}
>> +
>> +/* Process an incoming xs watch event */
>> +static void process_watch_event(char *watch_msg)
>> +{
>> +    /* TODO */
>> +}
>> +
>> +static void memcpy_from_ring(const char *ring, char *dest, int off, int len)
>> +{
>> +    int c1, c2;
>> +
>> +    c1 = MIN(len, XENSTORE_RING_SIZE - off);
>> +    c2 = len - c1;
>> +
>> +    memcpy(dest, ring + off, c1);
>> +    if (c2)
>> +            memcpy(dest + c1, ring, c2);
>> +}
>> +
>> +static void xs_msg_read(struct xsd_sockmsg *hdr)
>> +{
>> +    XENSTORE_RING_IDX cons;
>> +    char *payload;
>> +
>> +    payload = malloc(hdr->len + 1);
>> +    if (payload == NULL) {
>> +            uk_printd(DLVL_WARN,
>> +                    "No memory available for saving Xenstore message!");
>> +            return;
>> +    }
>> +
>> +    cons = xsh.buf->rsp_cons;
>> +
>> +    /* copy payload */
>> +    memcpy_from_ring(
>> +            xsh.buf->rsp,
>> +            payload,
>> +            MASK_XENSTORE_IDX(cons + sizeof(*hdr)),
>> +            hdr->len
>> +    );
>> +    payload[hdr->len] = '\0';
>> +
>> +    /* Remote must not see available space until we've copied the reply */
>> +    mb();
> Maybe wmb() here is enough? Or do we actually need a full barrier?

Similarly to the previous situation, before the barrier we have a read
operation (reading the request data), while after the barrier we have a
write operation (writing the consumer index). So we need to order write
after read with full barrier.

>> +    xsh.buf->rsp_cons += sizeof(*hdr) + hdr->len;
>> +
>> +    if (xsh.buf->rsp_prod - cons >= XENSTORE_RING_SIZE)
>> +            notify_remote_via_evtchn(xsh.evtchn);
>> +
>> +    if (hdr->type == XS_WATCH_EVENT)
>> +            process_watch_event(payload);
>> +    else
>> +            process_reply(hdr, payload);
>> +}
>> +
>> +static void xs_thread_func(void *ign __unused)
>> +{
>> +    struct xsd_sockmsg msg;
>> +    XENSTORE_RING_IDX prod = xsh.buf->rsp_prod;
>> +
>> +    for (;;) {
>> +            /* wait for incoming xs response */
>> +            uk_waitq_wait_event(&xsh.waitq, prod != xsh.buf->rsp_prod);
>> +
>> +            while (1) {
>> +                    prod = xsh.buf->rsp_prod;
>> +
>> +                    uk_printd(DLVL_EXTRA, "Rsp_cons %d, rsp_prod %d.\n",
>> +                            xsh.buf->rsp_cons, xsh.buf->rsp_prod);
>> +
>> +                    if (!xs_avail_space_for_read(sizeof(msg)))
>> +                            break;
>> +
>> +                    /* Make sure data is read after reading the indexes */
>> +                    rmb();
>> +
>> +                    /* copy the message header */
>> +                    memcpy_from_ring(
>> +                            xsh.buf->rsp,
>> +                            (char *) &msg,
>> +                            MASK_XENSTORE_IDX(xsh.buf->rsp_cons),
>> +                            sizeof(msg)
>> +                    );
>> +
>> +                    uk_printd(DLVL_EXTRA, "Msg len %lu, %u avail, id %u.\n",
>> +                            msg.len + sizeof(msg),
>> +                            xsh.buf->rsp_prod - xsh.buf->rsp_cons,
>> +                            msg.req_id);
>> +
>> +                    if (!xs_avail_space_for_read(sizeof(msg) + msg.len))
>> +                            break;
>> +
>> +                    /* Make sure data is read after reading the indexes */
>> +                    rmb();
>> +
>> +                    uk_printd(DLVL_EXTRA, "Message is good.\n");
>> +                    xs_msg_read(&msg);
>> +            }
>> +    }
>> +}
>> +
>> +static void xs_evtchn_handler(evtchn_port_t port,
>> +            struct __regs *regs __unused, void *ign __unused)
>> +{
>> +    UK_ASSERT(xsh.evtchn == port);
>> +    uk_waitq_wake_up(&xsh.waitq);
>> +}
>> +
>> +int xs_comms_init(void)
>> +{
>> +    struct uk_thread *thread;
>> +    evtchn_port_t port;
>> +
>> +    xs_request_pool_init(&xs_req_pool);
>> +
>> +    uk_waitq_init(&xsh.waitq);
>> +
>> +    thread = uk_thread_create("xenstore", xs_thread_func, NULL);
>> +    if (PTRISERR(thread))
>> +            return PTR2ERR(thread);
>> +
>> +    xsh.thread = thread;
>> +
>> +    xsh.evtchn = HYPERVISOR_start_info->store_evtchn;
>> +    xsh.buf = mfn_to_virt(HYPERVISOR_start_info->store_mfn);
>> +
>> +    port = bind_evtchn(xsh.evtchn, xs_evtchn_handler, NULL);
>> +    UK_ASSERT(port == xsh.evtchn);
>> +    unmask_evtchn(xsh.evtchn);
>> +
>> +    uk_printd(DLVL_INFO,
>> +            "Xenstore connection initialised on port %d, buf %p (mfn 
>> %#lx)\n",
>> +            port, xsh.buf, HYPERVISOR_start_info->store_mfn);
>> +
>> +    return 0;
>> +}
>> +
>> +void xs_comms_fini(void)
>> +{
>> +    mask_evtchn(xsh.evtchn);
>> +    unbind_evtchn(xsh.evtchn);
>> +
>> +    xsh.buf = NULL;
>> +
>> +    /* TODO stop thread, instead of killing it */
>> +    uk_thread_destroy(xsh.thread);
>> +    xsh.thread = NULL;
>> +}
>> diff --git a/plat/xen/xenbus/xs_comms.h b/plat/xen/xenbus/xs_comms.h
>> new file mode 100644
>> index 0000000..9b558d0
>> --- /dev/null
>> +++ b/plat/xen/xenbus/xs_comms.h
>> @@ -0,0 +1,64 @@
>> +/* SPDX-License-Identifier: BSD-3-Clause */
>> +/*
>> + * Authors: Costin Lupu <costin.lupu@xxxxxxxxx>
>> + *
>> + * Copyright (c) 2018, NEC Europe Ltd., NEC Corporation. All rights 
>> reserved.
>> + *
>> + * Redistribution and use in source and binary forms, with or without
>> + * modification, are permitted provided that the following conditions
>> + * are met:
>> + *
>> + * 1. Redistributions of source code must retain the above copyright
>> + *    notice, this list of conditions and the following disclaimer.
>> + * 2. Redistributions in binary form must reproduce the above copyright
>> + *    notice, this list of conditions and the following disclaimer in the
>> + *    documentation and/or other materials provided with the distribution.
>> + * 3. Neither the name of the copyright holder nor the names of its
>> + *    contributors may be used to endorse or promote products derived from
>> + *    this software without specific prior written permission.
>> + *
>> + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS 
>> IS"
>> + * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
>> + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 
>> PURPOSE
>> + * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
>> + * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
>> + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
>> + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
>> + * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
>> + * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
>> + * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF 
>> THE
>> + * POSSIBILITY OF SUCH DAMAGE.
>> + *
>> + * THIS HEADER MAY NOT BE EXTRACTED OR MODIFIED IN ANY WAY.
>> + */
>> +
>> +#ifndef __XS_COMMS_H__
>> +#define __XS_COMMS_H__
>> +
>> +#include <xen/io/xs_wire.h>
>> +#include <xenbus/xs.h>
>> +
>> +int  xs_comms_init(void);
>> +void xs_comms_fini(void);
>> +
>> +struct xs_iovec {
>> +    void *data;
>> +    unsigned int len;
>> +};
>> +
>> +/*
>> + * Sends a message to Xenstore and blocks waiting for a reply.
>> + * The reply is malloc'ed and should be freed by the caller.
>> + *
>> + * @param msg_type Xenstore message type
>> + * @param xbt Xenbus transaction id
>> + * @param req_iovecs Array of request strings buffers
>> + * @param req_iovecs_num Request strings buffers number
>> + * @param rep_iovec Incoming reply string buffer (optional)
>> + * @return 0 on success, a negative errno value on error.
>> + */
>> +int xs_msg_reply(enum xsd_sockmsg_type msg_type, xenbus_transaction_t xbt,
>> +    const struct xs_iovec *req_iovecs, int req_iovecs_num,
>> +    struct xs_iovec *rep_iovec);
>> +
>> +#endif /* __XS_COMMS_H__ */
>> -- 
>> 2.11.0
>>
> 

_______________________________________________
Minios-devel mailing list
Minios-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/minios-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.