Re: [Qemu-devel] [QEMU PATCH v5 4/6] migration: migrate QTAILQ

["Dr. David Alan Gilbert" <dgilbert@redhat.com>]

* Paolo Bonzini (pbonzini@redhat.com) wrote:
> 
> 
> On 05/10/2016 18:56, Dr. David Alan Gilbert wrote:
> > * Jianjun Duan (duanj@linux.vnet.ibm.com) wrote:
> >> Currently we cannot directly transfer a QTAILQ instance because of the
> >> limitation in the migration code. Here we introduce an approach to
> >> transfer such structures. In our approach such a structure is tagged
> >> with VMS_LINKED. We then modified vmstate_save_state and vmstate_load_state
> >> so that when VMS_LINKED is encountered, put and get from VMStateInfo are
> >> called respectively. We created VMStateInfo vmstate_info_qtailq for QTAILQ.
> >> Similar VMStateInfo can be created for other data structures such as list.
> >> This approach will be used to transfer pending_events and ccs_list in spapr
> >> state.
> >>
> >> We also create some macros in qemu/queue.h to access a QTAILQ using pointer
> >> arithmetic. This ensures that we do not depend on the implementation
> >> details about QTAILQ in the migration code.
> > 
> > I think we're going to need a way to have a more flexible
> > loops; and thus my choice here wouldn't be to use the .get/.put together
> > with the VMSD; but I think we'll end up needing a new
> > data structure, maybe a VMStateLoop *loop in VMStateField.
> 
> Or just realize that we already have a Turing-complete programming
> language at our disposal, and it's called C. :)

Yes, and it does seem that the virtio migration code has used it to its
full abilities.

> > Specifically, your format of QTAILQ is perfectly reasonable - a
> > byte before each entry which is 1 to indicate there's an entry or 0
> > to indicate termination, but there are lots of other variants, e.g.
> > 
> >    a) put_scsi_requests uses that byte to hold a flag, so it's 0,1,2
> >       0 still means terminate but 1 or 2 set a flag in the structure.
> 
> Seriously, the way I'd implement (a) is to make the QTAILQ reader peek
> for the next byte.  If 0, eat it and exit the loop.  Otherwise, go ahead
> with vmstate_load_state which will parse the byte again.
> 
> Likewise, on saving let the VMState write a non-zero byte at the
> beginning, and then write a zero at the end.
> 
> Yes, it's sickening but that's what you do to honor backwards compatibility.

Actually, that's not *that* bad an idea.

Lets go with Jianjun's structure for the moment; we can always expand on it.

It seems we have ~3 concepts that feel partially independent:

    a) The format of the loop on the wire (eg one byte per iteration, 0 terminates)
    b) The way the list is represented (QTAILQ, simple array, device specific linked-list)
    c) The data gathered in each iteration
    d) The allocation of (c)

This patch has a,b,d all wrapped up together in the get/put functions -
where I was hoping to find a way to separate them a bit so that we
could say; I want a loop, with this format, into this data structure, using this allocator.

Dave

> The other possibility is just to bump the version and make the SCSI
> request flag a separate byte after the "is there another entry" byte.
> 
> Paolo
> 
> >    b) slirp_state_load also uses a null byte termination but not off a QTAILQ
> >       (although I think it could be flipped for one) (it uses '42' for the
> >       non-0 value, but looks like it could become 1)
> > 
> >    c) virtio_blk also rolls it's own linked list but again with the 0/1 byte
> > 
> >   Now how would I modify your QTAILQ load/store to do (a) without copying the whole
> > thing?
> > 
> > Dave
> > 
> >>
> >> Signed-off-by: Jianjun Duan <duanj@linux.vnet.ibm.com>
> >> ---
> >>  include/migration/vmstate.h | 26 ++++++++++++++++++
> >>  include/qemu/queue.h        | 32 ++++++++++++++++++++++
> >>  migration/trace-events      |  4 +++
> >>  migration/vmstate.c         | 66 +++++++++++++++++++++++++++++++++++++++++++++
> >>  4 files changed, 128 insertions(+)
> >>
> >> diff --git a/include/migration/vmstate.h b/include/migration/vmstate.h
> >> index 459dd4a..e60c994 100644
> >> --- a/include/migration/vmstate.h
> >> +++ b/include/migration/vmstate.h
> >> @@ -186,6 +186,12 @@ enum VMStateFlags {
> >>       * to determine the number of entries in the array. Only valid in
> >>       * combination with one of VMS_VARRAY*. */
> >>      VMS_MULTIPLY_ELEMENTS = 0x4000,
> >> +    /* For fields which need customized handling, such as QTAILQ in queue.h.
> >> +     * When this flag is set in VMStateField, info->get/put will
> >> +     * be used in vmstate_load/save_state instead of recursive call.
> >> +     * User should implement set info to handle the concerned data structure.
> >> +     */
> >> +    VMS_LINKED            = 0x8000,
> >>  };
> >>  
> >>  struct VMStateField {
> >> @@ -246,6 +252,7 @@ extern const VMStateInfo vmstate_info_timer;
> >>  extern const VMStateInfo vmstate_info_buffer;
> >>  extern const VMStateInfo vmstate_info_unused_buffer;
> >>  extern const VMStateInfo vmstate_info_bitmap;
> >> +extern const VMStateInfo vmstate_info_qtailq;
> >>  
> >>  #define type_check_2darray(t1,t2,n,m) ((t1(*)[n][m])0 - (t2*)0)
> >>  #define type_check_array(t1,t2,n) ((t1(*)[n])0 - (t2*)0)
> >> @@ -657,6 +664,25 @@ extern const VMStateInfo vmstate_info_bitmap;
> >>      .offset       = offsetof(_state, _field),                        \
> >>  }
> >>  
> >> +/* For QTAILQ that need customized handling
> >> + * _type: type of QTAILQ element
> >> + * _next: name of QTAILQ entry field in QTAILQ element
> >> + * _vmsd: VMSD for QTAILQ element
> >> + * size: size of QTAILQ element
> >> + * start: offset of QTAILQ entry in QTAILQ element
> >> + */
> >> +#define VMSTATE_QTAILQ_V(_field, _state, _version, _vmsd, _type, _next)  \
> >> +{                                                                        \
> >> +    .name         = (stringify(_field)),                                 \
> >> +    .version_id   = (_version),                                          \
> >> +    .vmsd         = &(_vmsd),                                            \
> >> +    .size         = sizeof(_type),                                       \
> >> +    .info         = &vmstate_info_qtailq,                                \
> >> +    .flags        = VMS_LINKED,                                          \
> >> +    .offset       = offsetof(_state, _field),                            \
> >> +    .start        = offsetof(_type, _next),                              \
> >> +}
> >> +
> >>  /* _f : field name
> >>     _f_n : num of elements field_name
> >>     _n : num of elements
> >> diff --git a/include/qemu/queue.h b/include/qemu/queue.h
> >> index 342073f..12c3f80 100644
> >> --- a/include/qemu/queue.h
> >> +++ b/include/qemu/queue.h
> >> @@ -438,4 +438,36 @@ struct {                                                                \
> >>  #define QTAILQ_PREV(elm, headname, field) \
> >>          (*(((struct headname *)((elm)->field.tqe_prev))->tqh_last))
> >>  
> >> +/*
> >> + * Offsets of layout of a tail queue head.
> >> + */
> >> +#define QTAILQ_FIRST_OFFSET 0
> >> +#define QTAILQ_LAST_OFFSET (sizeof(void *))
> >> +
> >> +/*
> >> + * Offsets of layout of a tail queue element.
> >> + */
> >> +#define QTAILQ_NEXT_OFFSET 0
> >> +#define QTAILQ_PREV_OFFSET (sizeof(void *))
> >> +
> >> +/*
> >> + * Tail queue tranversal using pointer arithmetic.
> >> + */
> >> +#define QTAILQ_RAW_FOREACH(elm, head, entry)                                   \
> >> +        for ((elm) = *((void **) ((char *) (head) + QTAILQ_FIRST_OFFSET));     \
> >> +             (elm);                                                            \
> >> +             (elm) =                                                           \
> >> +                 *((void **) ((char *) (elm) + (entry) + QTAILQ_NEXT_OFFSET)))
> >> +/*
> >> + * Tail queue insertion using pointer arithmetic.
> >> + */
> >> +#define QTAILQ_RAW_INSERT_TAIL(head, elm, entry) do {                          \
> >> +        *((void **) ((char *) (elm) + (entry) + QTAILQ_NEXT_OFFSET)) = NULL;   \
> >> +        *((void **) ((char *) (elm) + (entry) + QTAILQ_PREV_OFFSET)) =         \
> >> +            *((void **) ((char *) (head) +QTAILQ_LAST_OFFSET));                \
> >> +        **((void ***)((char *) (head) +QTAILQ_LAST_OFFSET)) = (elm);           \
> >> +        *((void **) ((char *) (head) + QTAILQ_LAST_OFFSET)) =                  \
> >> +            (void *) ((char *) (elm) + (entry) + QTAILQ_NEXT_OFFSET);          \
> >> +} while (/*CONSTCOND*/0)
> >> +
> >>  #endif /* QEMU_SYS_QUEUE_H */
> >> diff --git a/migration/trace-events b/migration/trace-events
> >> index dfee75a..9a6ec59 100644
> >> --- a/migration/trace-events
> >> +++ b/migration/trace-events
> >> @@ -52,6 +52,10 @@ vmstate_n_elems(const char *name, int n_elems) "%s: %d"
> >>  vmstate_subsection_load(const char *parent) "%s"
> >>  vmstate_subsection_load_bad(const char *parent,  const char *sub, const char *sub2) "%s: %s/%s"
> >>  vmstate_subsection_load_good(const char *parent) "%s"
> >> +get_qtailq(const char *name, int version_id) "%s v%d"
> >> +get_qtailq_end(const char *name, const char *reason, int val) "%s %s/%d"
> >> +put_qtailq(const char *name, int version_id) "%s v%d"
> >> +put_qtailq_end(const char *name, const char *reason) "%s %s"
> >>  
> >>  # migration/qemu-file.c
> >>  qemu_file_fclose(void) ""
> >> diff --git a/migration/vmstate.c b/migration/vmstate.c
> >> index 66802cb..192db8a 100644
> >> --- a/migration/vmstate.c
> >> +++ b/migration/vmstate.c
> >> @@ -5,7 +5,9 @@
> >>  #include "migration/vmstate.h"
> >>  #include "qemu/bitops.h"
> >>  #include "qemu/error-report.h"
> >> +#include "qemu/queue.h"
> >>  #include "trace.h"
> >> +#include "migration/qjson.h"
> >>  
> >>  static void vmstate_subsection_save(QEMUFile *f, const VMStateDescription *vmsd,
> >>                                      void *opaque, QJSON *vmdesc);
> >> @@ -121,6 +123,8 @@ int vmstate_load_state(QEMUFile *f, const VMStateDescription *vmsd,
> >>                  if (field->flags & VMS_STRUCT) {
> >>                      ret = vmstate_load_state(f, field->vmsd, addr,
> >>                                               field->vmsd->version_id);
> >> +                } else if (field->flags & VMS_LINKED) {
> >> +                    ret = field->info->get(f, addr, size, field);
> >>                  } else {
> >>                      ret = field->info->get(f, addr, size, NULL);
> >>  
> >> @@ -193,6 +197,8 @@ static const char *vmfield_get_type_name(VMStateField *field)
> >>  
> >>      if (field->flags & VMS_STRUCT) {
> >>          type = "struct";
> >> +    } else if (field->flags & VMS_LINKED) {
> >> +        type = "linked";
> >>      } else if (field->info->name) {
> >>          type = field->info->name;
> >>      }
> >> @@ -327,6 +333,8 @@ void vmstate_save_state(QEMUFile *f, const VMStateDescription *vmsd,
> >>                  }
> >>                  if (field->flags & VMS_STRUCT) {
> >>                      vmstate_save_state(f, field->vmsd, addr, vmdesc_loop);
> >> +                } else if  (field->flags & VMS_LINKED) {
> >> +                    field->info->put(f, addr, size, field, vmdesc_loop);
> >>                  } else {
> >>                      field->info->put(f, addr, size, NULL, NULL);
> >>                  }
> >> @@ -939,3 +947,61 @@ const VMStateInfo vmstate_info_bitmap = {
> >>      .get = get_bitmap,
> >>      .put = put_bitmap,
> >>  };
> >> +
> >> +/*get for QTAILQ */
> >> +static int get_qtailq(QEMUFile *f, void *pv, size_t unused_size,
> >> +                      VMStateField *field)
> >> +{
> >> +    int ret = 0;
> >> +    const VMStateDescription *vmsd = field->vmsd;
> >> +    size_t size = field->size;
> >> +    size_t entry = field->start;
> >> +    int version_id = field->version_id;
> >> +    void *elm;
> >> +
> >> +    trace_get_qtailq(vmsd->name, version_id);
> >> +    if (version_id > vmsd->version_id) {
> >> +        trace_get_qtailq_end(vmsd->name, "too new", -EINVAL);
> > 
> > Can you make those error_report's please - if it fails we want to
> > see why in the log.
> > 
> > Dave
> > 
> >> +        return -EINVAL;
> >> +    }
> >> +    if (version_id < vmsd->minimum_version_id) {
> >> +        trace_get_qtailq_end(vmsd->name, "too old", -EINVAL);
> >> +        return -EINVAL;
> >> +    }
> >> +
> >> +    while (qemu_get_byte(f)) {
> >> +        elm =  g_malloc(size);
> >> +        ret = vmstate_load_state(f, vmsd, elm, version_id);
> >> +        if (ret) {
> >> +            return ret;
> >> +        }
> >> +        QTAILQ_RAW_INSERT_TAIL(pv, elm, entry);
> >> +    }
> >> +
> >> +    trace_get_qtailq_end(vmsd->name, "end", ret);
> >> +    return ret;
> >> +}
> >> +
> >> +/* put for QTAILQ */
> >> +static void put_qtailq(QEMUFile *f, void *pv, size_t unused_size,
> >> +                       VMStateField *field, QJSON *vmdesc)
> >> +{
> >> +    const VMStateDescription *vmsd = field->vmsd;
> >> +    size_t entry = field->start;
> >> +    void *elm;
> >> +
> >> +    trace_put_qtailq(vmsd->name, vmsd->version_id);
> >> +
> >> +    QTAILQ_RAW_FOREACH(elm, pv, entry) {
> >> +        qemu_put_byte(f, true);
> >> +        vmstate_save_state(f, vmsd, elm, vmdesc);
> >> +    }
> >> +    qemu_put_byte(f, false);
> >> +
> >> +    trace_put_qtailq_end(vmsd->name, "end");
> >> +}
> >> +const VMStateInfo vmstate_info_qtailq = {
> >> +    .name = "qtailq",
> >> +    .get  = get_qtailq,
> >> +    .put  = put_qtailq,
> >> +};
> >> -- 
> >> 1.9.1
> >>
> > --
> > Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
> > 
> > 
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK