Re: [Qemu-devel] [PATCH 2/4] QAPI: Introduce memchar-write QMP command

[Luiz Capitulino]

On Mon, 29 Oct 2012 12:10:24 +0800
Lei Li <address@hidden> wrote:

> On 10/27/2012 01:17 AM, Luiz Capitulino wrote:
> > On Fri, 26 Oct 2012 19:21:50 +0800
> > Lei Li <address@hidden> wrote:
> >
> >> Signed-off-by: Lei Li <address@hidden>
> >> ---
> >>   hmp-commands.hx  |   17 +++++++++++++++++
> >>   hmp.c            |   15 +++++++++++++++
> >>   hmp.h            |    1 +
> >>   qapi-schema.json |   47 +++++++++++++++++++++++++++++++++++++++++++++++
> >>   qemu-char.c      |   44 ++++++++++++++++++++++++++++++++++++++++++++
> >>   qmp-commands.hx  |   34 ++++++++++++++++++++++++++++++++++
> >>   6 files changed, 158 insertions(+), 0 deletions(-)
> >>
> >> diff --git a/hmp-commands.hx b/hmp-commands.hx
> >> index e0b537d..a37b8e9 100644
> >> --- a/hmp-commands.hx
> >> +++ b/hmp-commands.hx
> >> @@ -825,6 +825,23 @@ Inject an NMI on the given CPU (x86 only).
> >>   ETEXI
> >>   
> >>       {
> >> +        .name       = "memchar_write",
> >> +        .args_type  = "chardev:s,data:s",
> >> +        .params     = "chardev data",
> >> +        .mhandler.cmd = hmp_memchar_write,
> >> +    },
> >> +
> >> +STEXI
> >> address@hidden memchar_write @var{chardev} @var{data}
> >> address@hidden memchar_write
> >> +Provide writing interface for CirMemCharDriver. Write @var{data}
> >> +to cirmemchr char device. Note that we  will add 'control' options
> >> +for read and write command that specifies behavior when the queue
> >> +is full/empty, for now just assume a drop behaver in these two commands.
> > You can drop everything after "Note".
> >
> ok
> 
> >> +
> >> +ETEXI
> >> +
> >> +    {
> >>           .name       = "migrate",
> >>           .args_type  = "detach:-d,blk:-b,inc:-i,uri:s",
> >>           .params     = "[-d] [-b] [-i] uri",
> >> diff --git a/hmp.c b/hmp.c
> >> index 2b97982..082985b 100644
> >> --- a/hmp.c
> >> +++ b/hmp.c
> >> @@ -683,6 +683,21 @@ void hmp_pmemsave(Monitor *mon, const QDict *qdict)
> >>       hmp_handle_error(mon, &errp);
> >>   }
> >>   
> >> +void hmp_memchar_write(Monitor *mon, const QDict *qdict)
> >> +{
> >> +    uint32_t size;
> >> +    const char *chardev = qdict_get_str(qdict, "chardev");
> >> +    const char *data = qdict_get_str(qdict, "data");
> >> +    enum DataFormat format;
> > Why do you need this variable?
> 
> Sure, I will drop this.
> 
> >> +    Error *errp = NULL;
> >> +
> >> +    size = strlen(data);
> >> +    format = DATA_FORMAT_UTF8;
> >> +    qmp_memchar_write(chardev, size, data, true, format, &errp);
> >> +
> >> +    hmp_handle_error(mon, &errp);
> >> +}
> >> +
> >>   static void hmp_cont_cb(void *opaque, int err)
> >>   {
> >>       if (!err) {
> >> diff --git a/hmp.h b/hmp.h
> >> index 71ea384..406ebb1 100644
> >> --- a/hmp.h
> >> +++ b/hmp.h
> >> @@ -43,6 +43,7 @@ void hmp_system_powerdown(Monitor *mon, const QDict 
> >> *qdict);
> >>   void hmp_cpu(Monitor *mon, const QDict *qdict);
> >>   void hmp_memsave(Monitor *mon, const QDict *qdict);
> >>   void hmp_pmemsave(Monitor *mon, const QDict *qdict);
> >> +void hmp_memchar_write(Monitor *mon, const QDict *qdict);
> >>   void hmp_cont(Monitor *mon, const QDict *qdict);
> >>   void hmp_system_wakeup(Monitor *mon, const QDict *qdict);
> >>   void hmp_inject_nmi(Monitor *mon, const QDict *qdict);
> >> diff --git a/qapi-schema.json b/qapi-schema.json
> >> index c615ee2..43ef6bc 100644
> >> --- a/qapi-schema.json
> >> +++ b/qapi-schema.json
> >> @@ -325,6 +325,53 @@
> >>   { 'command': 'query-chardev', 'returns': ['ChardevInfo'] }
> >>   
> >>   ##
> >> +# @DataFormat:
> >> +#
> >> +# An enumeration of data format. The default value would
> >> +# be utf8.
> > Please, remove the "default value" part. This is decided by the command
> > using this type.
> 
> Now the option format is optional, if it's not set then default by 'utf8'.
> I think it's a reasonable behaver. :)

What I meant is that the right place to say what the value is is in
the command documentation, not in @DataFormat doc. Here, only having
"An enumeration of data encodings" is fine.

> >> +#
> >> +# @utf8: The data format is 'utf8'.
> >> +#
> >> +# @base64: The data format is 'base64'.
> >> +#
> >> +# Note: The data format start with 'utf8' and 'base64',
> >> +#       will support other data format as well.
> > Please, drop this note. It's not needed.
> >
> ok
> 
> >> +#
> >> +# Since: 1.3
> >> +##
> >> +{ 'enum': 'DataFormat'
> >> +  'data': [ 'utf8', 'base64' ] }
> >> +
> >> +##
> >> +# @memchar-write:
> >> +#
> >> +# Provide writing interface for memchardev. Write data to memchar
> >> +# char device.
> >> +#
> >> +# @chardev: the name of the memchar char device.
> >> +#
> >> +# @size: the size to write in bytes. Should be power of 2.
> >> +#
> >> +# @data: the source data write to memchar.
> >> +#
> >> +# @format: #optional the format of the data write to memchardev, by
> >> +#          default is 'utf8'.
> >> +#
> >> +# Returns: Nothing on success
> >> +#          If @chardev is not a valid memchr device, DeviceNotFound
> >> +#
> >> +# Notes: The option 'block' is not supported now due to the miss
> >> +#        feature in qmp. Will add it later when we gain the necessary
> >> +#        infrastructure enhancement. For now just assume 'drop' behaver
> >> +#        for this command.
> > Please, replace this note with an explanation of the current behavior. No
> > need to talk about the future.
> 
> Sure.
> 
> >> +#
> >> +# Since: 1.3
> >> +##
> >> +{ 'command': 'memchar-write',
> >> +  'data': {'chardev': 'str', 'size': 'int', 'data': 'str',
> >> +           '*format': 'DataFormat'} }
> >> +
> >> +##
> >>   # @CommandInfo:
> >>   #
> >>   # Information about a QMP command
> >> diff --git a/qemu-char.c b/qemu-char.c
> >> index c3ec43d..6114e29 100644
> >> --- a/qemu-char.c
> >> +++ b/qemu-char.c
> >> @@ -2717,6 +2717,50 @@ fail:
> >>       return NULL;
> >>   }
> >>   
> >> +void qmp_memchar_write(const char *chardev, int64_t size,
> >> +                       const char *data, bool has_format,
> >> +                       enum DataFormat format,
> >> +                       Error **errp)
> >> +{
> >> +    CharDriverState *chr;
> >> +    guchar *write_data;
> >> +    int ret;
> >> +    gsize write_count;
> >> +
> >> +    chr = qemu_chr_find(chardev);
> >> +    if (!chr) {
> >> +        error_set(errp, QERR_DEVICE_NOT_FOUND, chardev);
> >> +        return;
> >> +    }
> >> +
> >> +    if (strcmp(chr->filename, "memchr") != 0) {
> >> +        error_setg(errp,"%s is not memory char device\n", chardev);
> >> +        return;
> >> +    }
> > This is wrong, as it's unreliable. A hackish (but hopefully correct) way
> > of doing this would be to check chr->init against the circmem init function.
> > But please, put this behind a function because it's ugly.
> 
> Hmm, I guess it's because the chr->filename could be NULL. but we can still
> get it since CharDriverState keeps the inited type of backend. I am not sure
> that check chr->init against the cirmem init funciton could work, I'll have
> a try.
> 
> >
> >> +
> >> +    /* XXX: For the sync command as 'block', waiting for the qmp
> >> +     * to support necessary feature. Now just act as 'drop' */
> > Please, replace this note with an explanation of the current behavior.
> 
> ok.
> 
> >> +    if (cirmem_chr_is_full(chr)) {
> >> +        error_setg(errp, "Failed to write to memchr %s", chardev);
> >> +        return;
> >> +    }
> >> +
> >> +    write_count = (gsize)size;
> >> +
> >> +    if (has_format && (format == DATA_FORMAT_BASE64)) {
> >> +        write_data = g_base64_decode(data, &write_count);
> >> +    } else {
> >> +        write_data = (uint8_t *)data;
> >> +    }
> >> +
> >> +    ret = cirmem_chr_write(chr, write_data, write_count);
> > IIRC circmem_chr_write() doesn't check if write_count if a power of two,
> > what happens if it isn't?
> 
> I have considered your suggestion on this in v4. The reason why I did not
> add the assert is that I think we have already check with this when cirmem
> init, so there is no need to check it again. Correct me if I am wrong. :)

I was actually asking an honest question: does cirmem_chr_write() assume
that len is a power of two? If yes, what if the user doesn't pass a
power of two value to qmp_memchar_write()?

> 
> >> +
> >> +    if (ret < 0) {
> >> +        error_setg(errp, "Failed to write to memchr %s", chardev);
> >> +        return;
> >> +    }
> >> +}
> >> +
> >>   QemuOpts *qemu_chr_parse_compat(const char *label, const char *filename)
> >>   {
> >>       char host[65], port[33], width[8], height[8];
> >> diff --git a/qmp-commands.hx b/qmp-commands.hx
> >> index 5ba8c48..7548b9b 100644
> >> --- a/qmp-commands.hx
> >> +++ b/qmp-commands.hx
> >> @@ -466,6 +466,40 @@ Note: inject-nmi fails when the guest doesn't support 
> >> injecting.
> >>   EQMP
> >>   
> >>       {
> >> +        .name       = "memchar-write",
> >> +        .args_type  = "chardev:s,size:i,data:s,format:s?,control:s?",
> > You've dropped 'control', haven't you?
> 
> Yes, sorry I forgot to drop here.. :(
> 
> >
> >> +        .help       = "write size of data to memchar chardev",
> >> +        .mhandler.cmd_new = qmp_marshal_input_memchar_write,
> >> +    },
> >> +
> >> +SQMP
> >> +memchar-write
> >> +-------------
> >> +
> >> +Provide writing interface for memchardev. Write data to memchar
> >> +char device.
> >> +
> >> +Arguments:
> >> +
> >> +- "chardev": the name of the char device, must be unique (json-string)
> >> +- "size": the memory size, in bytes, should be power of 2 (json-int)
> >> +- "data": the source data writed to memchar (json-string)
> >> +- "format": the data format write to memchardev, default is
> >> +            utf8. (json-string, optional)
> >> +          - Possible values: "utf8", "base64"
> >> +
> >> +Example:
> >> +
> >> +-> { "execute": "memchar-write",
> >> +                "arguments": { "chardev": foo,
> >> +                               "size": 8,
> >> +                               "data": "abcdefgh",
> >> +                               "format": "utf8" } }
> >> +<- { "return": {} }
> >> +
> >> +EQMP
> >> +
> >> +    {
> >>           .name       = "xen-save-devices-state",
> >>           .args_type  = "filename:F",
> >>       .mhandler.cmd_new = qmp_marshal_input_xen_save_devices_state,
> >
> 
>