[Qemu-devel] [PATCH 0/2 v2] introduce visitor for parsing suffixed integer

[Qemu-devel] [PATCH 0/2 v2] introduce visitor for parsing suffixed integer

[Igor Mammedov]

v2:
 * Naming changes:
    - s/visit_type_uint_suffixed_int/visit_type_suffixed_int/
    - use 'suffix_factor' instead of 'unit'
 * Added  documentation to visit_type_suffixed_int()
 * Fixed errp check.
 * Style fixes

Reference to previous version dicussion:
    http://lists.gnu.org/archive/html/qemu-devel/2012-12/msg00758.html

Git tree for testing:
    https://github.com/imammedo/qemu/tree/type_suffixed_int_v2

Igor Mammedov (2):
  add visitor for parsing int[KMGT] input string
  target-i386: use visit_type_suffixed_int() to parse tsc_freq property
    value

 qapi/qapi-dealloc-visitor.c |  8 ++++++++
 qapi/qapi-visit-core.c      | 35 +++++++++++++++++++++++++++++++++++
 qapi/qapi-visit-core.h      |  4 ++++
 qapi/string-input-visitor.c | 25 +++++++++++++++++++++++++
 target-i386/cpu.c           |  3 ++-
 5 files changed, 74 insertions(+), 1 deletion(-)

-- 
1.7.11.7

[Qemu-devel] [PATCH 1/2] add visitor for parsing int[KMGT] input string

[Igor Mammedov]

Caller of visit_type_suffixed_int() have to specify
value of 'K' suffix using suffix_factor argument.
Example of selecting suffix_factor value:
 * Kbytes: 1024
 * Khz: 1000

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
---
 v3:
  - Fix errp check. Spotted-By: Andreas Färber <afaerber@suse.de>
  - s/type_unit_suffixed_int/type_suffixed_int/
  - use 'suffix_factor' instead of 'unit'
  - document visit_type_suffixed_int()
  - add comment on current impl. limitation
 v2:
  - convert type_freq to type_unit_suffixed_int.
  - provide qapi_dealloc_type_unit_suffixed_int() impl.
---
 qapi/qapi-dealloc-visitor.c |  8 ++++++++
 qapi/qapi-visit-core.c      | 35 +++++++++++++++++++++++++++++++++++
 qapi/qapi-visit-core.h      |  4 ++++
 qapi/string-input-visitor.c | 25 +++++++++++++++++++++++++
 4 files changed, 72 insertions(+)

diff --git a/qapi/qapi-dealloc-visitor.c b/qapi/qapi-dealloc-visitor.c
index 75214e7..09c3b01 100644
--- a/qapi/qapi-dealloc-visitor.c
+++ b/qapi/qapi-dealloc-visitor.c
@@ -143,6 +143,13 @@ static void qapi_dealloc_type_enum(Visitor *v, int *obj, const char *strings[],
 {
 }
 
+static void qapi_dealloc_type_suffixed_int(Visitor *v, int64_t *obj,
+                                           const char *name,
+                                           const int suffix_factor,
+                                           Error **errp)
+{
+}
+
 Visitor *qapi_dealloc_get_visitor(QapiDeallocVisitor *v)
 {
     return &v->visitor;
@@ -170,6 +177,7 @@ QapiDeallocVisitor *qapi_dealloc_visitor_new(void)
     v->visitor.type_str = qapi_dealloc_type_str;
     v->visitor.type_number = qapi_dealloc_type_number;
     v->visitor.type_size = qapi_dealloc_type_size;
+    v->visitor.type_suffixed_int = qapi_dealloc_type_suffixed_int;
 
     QTAILQ_INIT(&v->stack);
 
diff --git a/qapi/qapi-visit-core.c b/qapi/qapi-visit-core.c
index 7a82b63..9f6b8ae 100644
--- a/qapi/qapi-visit-core.c
+++ b/qapi/qapi-visit-core.c
@@ -311,3 +311,38 @@ void input_type_enum(Visitor *v, int *obj, const char *strings[],
     g_free(enum_str);
     *obj = value;
 }
+
+/**
+ * visit_type_suffixed_int:
+ * @v: visitor used for accesing external representation of value
+ * @obj: object that stores internal value
+ * @name: name of the option
+ * @suffix_factor: multiplication factor of suffix 'K'
+ * @errp: object to store error code
+ *
+ * Converts option value represented by @v taking in account multiplication
+ * factor of a suffix character that might follow number.
+ *   i.e. resulted *@obj = number * suffix.
+ *
+ * Where following suffixes are recognised:
+ *  no suffix = 1
+ *  K = @suffix_factor
+ *  M = K * @suffix_factor
+ *  G = M * @suffix_factor
+ *  T = G * @suffix_factor
+ *
+ * In case @v doesn't provide type_suffixed_int() parser, convertion
+ * fall-backs to suffix-less type_int64() parser.
+ */
+void visit_type_suffixed_int(Visitor *v, int64_t *obj, const char *name,
+                             const int suffix_factor, Error **errp)
+{
+    if (error_is_set(errp)) {
+        return;
+    }
+    if (v->type_suffixed_int) {
+        v->type_suffixed_int(v, obj, name, suffix_factor, errp);
+    } else {
+        visit_type_int64(v, obj, name, errp);
+    }
+}
diff --git a/qapi/qapi-visit-core.h b/qapi/qapi-visit-core.h
index 60aceda..05e9940 100644
--- a/qapi/qapi-visit-core.h
+++ b/qapi/qapi-visit-core.h
@@ -62,6 +62,8 @@ struct Visitor
     void (*type_int64)(Visitor *v, int64_t *obj, const char *name, Error **errp);
     /* visit_type_size() falls back to (*type_uint64)() if type_size is unset */
     void (*type_size)(Visitor *v, uint64_t *obj, const char *name, Error **errp);
+    void (*type_suffixed_int)(Visitor *v, int64_t *obj, const char *name,
+                              const int uffix_factor, Error **errp);
 };
 
 void visit_start_handle(Visitor *v, void **obj, const char *kind,
@@ -91,5 +93,7 @@ void visit_type_size(Visitor *v, uint64_t *obj, const char *name, Error **errp);
 void visit_type_bool(Visitor *v, bool *obj, const char *name, Error **errp);
 void visit_type_str(Visitor *v, char **obj, const char *name, Error **errp);
 void visit_type_number(Visitor *v, double *obj, const char *name, Error **errp);
+void visit_type_suffixed_int(Visitor *v, int64_t *obj, const char *name,
+                             const int suffix_factor, Error **errp);
 
 #endif
diff --git a/qapi/string-input-visitor.c b/qapi/string-input-visitor.c
index 497eb9a..9641bce 100644
--- a/qapi/string-input-visitor.c
+++ b/qapi/string-input-visitor.c
@@ -110,6 +110,30 @@ static void parse_start_optional(Visitor *v, bool *present,
     *present = true;
 }
 
+static void parse_type_suffixed_int(Visitor *v, int64_t *obj, const char *name,
+                                    const int suffix_factor, Error **errp)
+{
+    StringInputVisitor *siv = DO_UPCAST(StringInputVisitor, visitor, v);
+    char *endp = (char *) siv->string;
+    long long val = 0;
+
+    if (siv->string) {
+        /* TODO: presently strtosz_suffix_unit() is limited to [0..INT64_MAX]
+         * it should be fixed to handle [INT64_MIN..INT64_MAX],
+         * to cover whole int64_t range
+         */
+        val = strtosz_suffix_unit(siv->string, &endp,
+                                  STRTOSZ_DEFSUFFIX_B, suffix_factor);
+    }
+    if (!siv->string || val == -1 || *endp) {
+        error_set(errp, QERR_INVALID_PARAMETER_VALUE, name,
+                  "a value representable as a non-negative int64");
+        return;
+    }
+
+    *obj = val;
+}
+
 Visitor *string_input_get_visitor(StringInputVisitor *v)
 {
     return &v->visitor;
@@ -132,6 +156,7 @@ StringInputVisitor *string_input_visitor_new(const char *str)
     v->visitor.type_str = parse_type_str;
     v->visitor.type_number = parse_type_number;
     v->visitor.start_optional = parse_start_optional;
+    v->visitor.type_suffixed_int = parse_type_suffixed_int;
 
     v->string = str;
     return v;
-- 
1.7.11.7

Re: [Qemu-devel] [PATCH 1/2] add visitor for parsing int[KMGT] input string

[Eduardo Habkost]

On Mon, Dec 10, 2012 at 10:33:06PM +0100, Igor Mammedov wrote:
> Caller of visit_type_suffixed_int() have to specify
> value of 'K' suffix using suffix_factor argument.
> Example of selecting suffix_factor value:
>  * Kbytes: 1024
>  * Khz: 1000
> 
> Signed-off-by: Igor Mammedov <imammedo@redhat.com>

Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>


I wonder if we could later introduce a visit_type_frequency() function
that simply calls visit_type_suffixed_int(). This would allow us to use
a 'frequency' type on QAPI, like the existing 'size' type we already
have.

I suggest having explicitly distinct types on QAPI because the 'size'
type probably won't abort (and maybe it _can't_ abort, to keep
compatibility) in case it finds a "100MB" string. Likewise, the
'frequency' type wouldn't abort in case it finds a "100MHz" string.

With separate types, we could also make the 'frequency' type _not_
accept "100B" as a valid string (strtosz_suffix_unit() accepts "B" as a
valid suffix, today).


> ---
>  v3:
>   - Fix errp check. Spotted-By: Andreas Färber <afaerber@suse.de>
>   - s/type_unit_suffixed_int/type_suffixed_int/
>   - use 'suffix_factor' instead of 'unit'
>   - document visit_type_suffixed_int()
>   - add comment on current impl. limitation
>  v2:
>   - convert type_freq to type_unit_suffixed_int.
>   - provide qapi_dealloc_type_unit_suffixed_int() impl.
> ---
>  qapi/qapi-dealloc-visitor.c |  8 ++++++++
>  qapi/qapi-visit-core.c      | 35 +++++++++++++++++++++++++++++++++++
>  qapi/qapi-visit-core.h      |  4 ++++
>  qapi/string-input-visitor.c | 25 +++++++++++++++++++++++++
>  4 files changed, 72 insertions(+)
> 
[...]

-- 
Eduardo

Re: [Qemu-devel] [PATCH 1/2] add visitor for parsing int[KMGT] input string

[Igor Mammedov]

On Wed, 12 Dec 2012 16:16:42 -0200
Eduardo Habkost <ehabkost@redhat.com> wrote:

> On Mon, Dec 10, 2012 at 10:33:06PM +0100, Igor Mammedov wrote:
> > Caller of visit_type_suffixed_int() have to specify
> > value of 'K' suffix using suffix_factor argument.
> > Example of selecting suffix_factor value:
> >  * Kbytes: 1024
> >  * Khz: 1000
> > 
> > Signed-off-by: Igor Mammedov <imammedo@redhat.com>
> 
> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
> 
> 
> I wonder if we could later introduce a visit_type_frequency() function
> that simply calls visit_type_suffixed_int(). This would allow us to use
> a 'frequency' type on QAPI, like the existing 'size' type we already
> have.
> 
> I suggest having explicitly distinct types on QAPI because the 'size'
> type probably won't abort (and maybe it _can't_ abort, to keep
> compatibility) in case it finds a "100MB" string. Likewise, the
It won't accept MB with current code, but we could probably pass
something like custom suffix table {KHz => 1000, MHz=>1000000, ...}  instead
of unit for variables that accept frequency, and a corresponding table for
sizes and whatever else if needed. Than we could use only
visit_type_suffixed_int() and avoid creating an extra boiler code for every
kind of units that might be needed in future.
  
> 'frequency' type wouldn't abort in case it finds a "100MHz" string.
> 
> With separate types, we could also make the 'frequency' type _not_
> accept "100B" as a valid string (strtosz_suffix_unit() accepts "B" as a
> valid suffix, today).
> 
> 
> > ---
> >  v3:
> >   - Fix errp check. Spotted-By: Andreas Färber <afaerber@suse.de>
> >   - s/type_unit_suffixed_int/type_suffixed_int/
> >   - use 'suffix_factor' instead of 'unit'
> >   - document visit_type_suffixed_int()
> >   - add comment on current impl. limitation
> >  v2:
> >   - convert type_freq to type_unit_suffixed_int.
> >   - provide qapi_dealloc_type_unit_suffixed_int() impl.
> > ---
> >  qapi/qapi-dealloc-visitor.c |  8 ++++++++
> >  qapi/qapi-visit-core.c      | 35 +++++++++++++++++++++++++++++++++++
> >  qapi/qapi-visit-core.h      |  4 ++++
> >  qapi/string-input-visitor.c | 25 +++++++++++++++++++++++++
> >  4 files changed, 72 insertions(+)
> > 
> [...]
> 
> -- 
> Eduardo
> 


-- 
Regards,
  Igor

Re: [Qemu-devel] [PATCH 1/2] add visitor for parsing int[KMGT] input string

[Eduardo Habkost]

On Fri, Dec 14, 2012 at 02:20:48PM +0100, Igor Mammedov wrote:
> On Wed, 12 Dec 2012 16:16:42 -0200
> Eduardo Habkost <ehabkost@redhat.com> wrote:
> 
> > On Mon, Dec 10, 2012 at 10:33:06PM +0100, Igor Mammedov wrote:
> > > Caller of visit_type_suffixed_int() have to specify
> > > value of 'K' suffix using suffix_factor argument.
> > > Example of selecting suffix_factor value:
> > >  * Kbytes: 1024
> > >  * Khz: 1000
> > > 
> > > Signed-off-by: Igor Mammedov <imammedo@redhat.com>
> > 
> > Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
> > 
> > 
> > I wonder if we could later introduce a visit_type_frequency() function
> > that simply calls visit_type_suffixed_int(). This would allow us to use
> > a 'frequency' type on QAPI, like the existing 'size' type we already
> > have.
> > 
> > I suggest having explicitly distinct types on QAPI because the 'size'
> > type probably won't abort (and maybe it _can't_ abort, to keep
> > compatibility) in case it finds a "100MB" string. Likewise, the
> It won't accept MB with current code, but we could probably pass
> something like custom suffix table {KHz => 1000, MHz=>1000000, ...}  instead
> of unit for variables that accept frequency, and a corresponding table for
> sizes and whatever else if needed. Than we could use only
> visit_type_suffixed_int() and avoid creating an extra boiler code for every
> kind of units that might be needed in future.

The above make sense, but I wasn't worrying about how it's done
internally when calling/configuring the parsing code. My only point is
about the type definitions in QAPI. I believe that having this in QAPI:
  'type':'frequency'
  'type':'size'
looks better than:
  'type': 'suffixed_int', 'suffix_factor': 1000, 'unit': 'Hz'
  'type': 'suffixed_int', 'suffix_factor': 1024, 'unit': 'B'
or, even worse:
  'type': 'suffixed_int', 'suffix_table': { 'K': 1000, 'KHz': 1000, 'M': 1000000, 'MHz': 10000000, ... }
  'type': 'suffixed_int', 'suffix_table': { 'K': 1024, 'KB': 1024, 'M': 1048576, 'MB': 1048576, ... }
or:
  'type': 'suffixed_int', 'suffix_table': { 'K': 1000, 'M': 1000000, ... }, 'unit': 'Hz'
  'type': 'suffixed_int', 'suffix_table': { 'K': 1024, 'M': 1048576, ... }, 'unit': 'B'


>   
> > 'frequency' type wouldn't abort in case it finds a "100MHz" string.
> > 
> > With separate types, we could also make the 'frequency' type _not_
> > accept "100B" as a valid string (strtosz_suffix_unit() accepts "B" as a
> > valid suffix, today).
> > 
> > 
> > > ---
> > >  v3:
> > >   - Fix errp check. Spotted-By: Andreas Färber <afaerber@suse.de>
> > >   - s/type_unit_suffixed_int/type_suffixed_int/
> > >   - use 'suffix_factor' instead of 'unit'
> > >   - document visit_type_suffixed_int()
> > >   - add comment on current impl. limitation
> > >  v2:
> > >   - convert type_freq to type_unit_suffixed_int.
> > >   - provide qapi_dealloc_type_unit_suffixed_int() impl.
> > > ---
> > >  qapi/qapi-dealloc-visitor.c |  8 ++++++++
> > >  qapi/qapi-visit-core.c      | 35 +++++++++++++++++++++++++++++++++++
> > >  qapi/qapi-visit-core.h      |  4 ++++
> > >  qapi/string-input-visitor.c | 25 +++++++++++++++++++++++++
> > >  4 files changed, 72 insertions(+)
> > > 
> > [...]
> > 
> > -- 
> > Eduardo
> > 
> 
> 
> -- 
> Regards,
>   Igor

-- 
Eduardo

[Qemu-devel] [PATCH 2/2] target-i386: use visit_type_suffixed_int() to parse tsc_freq property value

[Igor Mammedov]

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
---
  v3:
   - s/visit_type_unit_suffixed_int/visit_type_suffixed_int/
  v2:
   - replace visit_type_freq() with visit_type_unit_suffixed_int()
     in x86_cpuid_set_tsc_freq()
---
 target-i386/cpu.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/target-i386/cpu.c b/target-i386/cpu.c
index 7be3ad8..18adff4 100644
--- a/target-i386/cpu.c
+++ b/target-i386/cpu.c
@@ -1194,8 +1194,9 @@ static void x86_cpuid_set_tsc_freq(Object *obj, Visitor *v, void *opaque,
     const int64_t min = 0;
     const int64_t max = INT64_MAX;
     int64_t value;
+    const int kHz_factor = 1000;
 
-    visit_type_int(v, &value, name, errp);
+    visit_type_suffixed_int(v, &value, name, kHz_factor, errp);
     if (error_is_set(errp)) {
         return;
     }
-- 
1.7.11.7

Re: [Qemu-devel] [PATCH 2/2] target-i386: use visit_type_suffixed_int() to parse tsc_freq property value

[Eduardo Habkost]

On Mon, Dec 10, 2012 at 10:33:07PM +0100, Igor Mammedov wrote:
> Signed-off-by: Igor Mammedov <imammedo@redhat.com>

Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>

Like I said in the previous patch, maybe it would be nice to have a
visit_type_frequency() helper, to hide the specific details/parameters
required to parse frequency values.


> ---
>   v3:
>    - s/visit_type_unit_suffixed_int/visit_type_suffixed_int/
>   v2:
>    - replace visit_type_freq() with visit_type_unit_suffixed_int()
>      in x86_cpuid_set_tsc_freq()
> ---
>  target-i386/cpu.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/target-i386/cpu.c b/target-i386/cpu.c
> index 7be3ad8..18adff4 100644
> --- a/target-i386/cpu.c
> +++ b/target-i386/cpu.c
> @@ -1194,8 +1194,9 @@ static void x86_cpuid_set_tsc_freq(Object *obj, Visitor *v, void *opaque,
>      const int64_t min = 0;
>      const int64_t max = INT64_MAX;
>      int64_t value;
> +    const int kHz_factor = 1000;
>  
> -    visit_type_int(v, &value, name, errp);
> +    visit_type_suffixed_int(v, &value, name, kHz_factor, errp);
>      if (error_is_set(errp)) {
>          return;
>      }
> -- 
> 1.7.11.7
> 
> 

-- 
Eduardo

Re: [Qemu-devel] [PATCH 0/2 v2] introduce visitor for parsing suffixed integer

[Anthony Liguori]

Igor Mammedov <imammedo@redhat.com> writes:

> v2:
>  * Naming changes:
>     - s/visit_type_uint_suffixed_int/visit_type_suffixed_int/
>     - use 'suffix_factor' instead of 'unit'
>  * Added  documentation to visit_type_suffixed_int()
>  * Fixed errp check.
>  * Style fixes

This is not how visitors are supposed to be used.

Just treat tsc_freq as a string property and parse it in the setter.

Regards,

Anthony Liguori

>
> Reference to previous version dicussion:
>     http://lists.gnu.org/archive/html/qemu-devel/2012-12/msg00758.html
>
> Git tree for testing:
>     https://github.com/imammedo/qemu/tree/type_suffixed_int_v2
>
> Igor Mammedov (2):
>   add visitor for parsing int[KMGT] input string
>   target-i386: use visit_type_suffixed_int() to parse tsc_freq property
>     value
>
>  qapi/qapi-dealloc-visitor.c |  8 ++++++++
>  qapi/qapi-visit-core.c      | 35 +++++++++++++++++++++++++++++++++++
>  qapi/qapi-visit-core.h      |  4 ++++
>  qapi/string-input-visitor.c | 25 +++++++++++++++++++++++++
>  target-i386/cpu.c           |  3 ++-
>  5 files changed, 74 insertions(+), 1 deletion(-)
>
> -- 
> 1.7.11.7

Re: [Qemu-devel] [PATCH 0/2 v2] introduce visitor for parsing suffixed integer

[Igor Mammedov]

On Sun, 23 Dec 2012 14:34:21 -0600
Anthony Liguori <anthony@codemonkey.ws> wrote:

> Igor Mammedov <imammedo@redhat.com> writes:
> 
> > v2:
> >  * Naming changes:
> >     - s/visit_type_uint_suffixed_int/visit_type_suffixed_int/
> >     - use 'suffix_factor' instead of 'unit'
> >  * Added  documentation to visit_type_suffixed_int()
> >  * Fixed errp check.
> >  * Style fixes
> 
> This is not how visitors are supposed to be used.
We were trying to generalize and make code more reusable.
Later could be used as a common base for specialized size and Hz visitors.

> Just treat tsc_freq as a string property and parse it in the setter.
Would http://lists.gnu.org/archive/html/qemu-devel/2012-10/msg03835.html
be better or you suggest just abandon generalization idea and parse it
locally in target-i386/cpu.c?

> 
> Regards,
> 
> Anthony Liguori
> 
> >
> > Reference to previous version dicussion:
> >     http://lists.gnu.org/archive/html/qemu-devel/2012-12/msg00758.html
> >
> > Git tree for testing:
> >     https://github.com/imammedo/qemu/tree/type_suffixed_int_v2
> >
> > Igor Mammedov (2):
> >   add visitor for parsing int[KMGT] input string
> >   target-i386: use visit_type_suffixed_int() to parse tsc_freq property
> >     value
> >
> >  qapi/qapi-dealloc-visitor.c |  8 ++++++++
> >  qapi/qapi-visit-core.c      | 35 +++++++++++++++++++++++++++++++++++
> >  qapi/qapi-visit-core.h      |  4 ++++
> >  qapi/string-input-visitor.c | 25 +++++++++++++++++++++++++
> >  target-i386/cpu.c           |  3 ++-
> >  5 files changed, 74 insertions(+), 1 deletion(-)
> >
> > -- 
> > 1.7.11.7


-- 
Regards,
  Igor

Re: [Qemu-devel] [PATCH 0/2 v2] introduce visitor for parsing suffixed integer

[Eduardo Habkost]

On Sun, Dec 23, 2012 at 02:34:21PM -0600, Anthony Liguori wrote:
> Igor Mammedov <imammedo@redhat.com> writes:
> 
> > v2:
> >  * Naming changes:
> >     - s/visit_type_uint_suffixed_int/visit_type_suffixed_int/
> >     - use 'suffix_factor' instead of 'unit'
> >  * Added  documentation to visit_type_suffixed_int()
> >  * Fixed errp check.
> >  * Style fixes
> 
> This is not how visitors are supposed to be used.
> 
> Just treat tsc_freq as a string property and parse it in the setter.

So, why visit_type_size() does exist? Should we work to eliminate it? If
not, why is it different from a "frequency" type?

-- 
Eduardo