glthread: pack "size" in Pointer calls as 16 bits

The only legal values are {1, 2, 3, 4, GL_BGRA}.
We need GLpacked16i to be unsigned, not signed, because GL_BGRA is
greater than 0x8000.

This decreases the size of 1 frame by 10% in Viewperf2020/Catia1.
It decreases the size of many Pointer calls by 8 bytes.

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27350>

src/mapi/glapi/gen/gl_marshal.py

@@ -229,6 +229,8 @@ class PrintCode(gl_XML.gl_print_base):
                     out('cmd->{0} = MIN2({0}, 0xffff); /* clamped to 0xffff (invalid enum) */'.format(p.name))
                 elif type == 'GLclamped16i':
                     out('cmd->{0} = CLAMP({0}, INT16_MIN, INT16_MAX);'.format(p.name))
+                elif type == 'GLpacked16i':
+                    out('cmd->{0} = {0} < 0 ? UINT16_MAX : MIN2({0}, UINT16_MAX);'.format(p.name))
                 else:
                     out('cmd->{0} = {0};'.format(p.name))
             if variable_params:

src/mapi/glapi/gen/marshal_XML.py

@@ -43,6 +43,9 @@ def get_marshal_type(func_name, param):
         if (type, param.name) == ('GLsizei', 'stride'):
             return 'GLclamped16i'
 
+        if (type, param.name) == ('GLint', 'size'):
+            return 'GLpacked16i'
+
     return type
 
 def get_type_size(func_name, param):
@@ -60,6 +63,7 @@ def get_type_size(func_name, param):
         'GLushort': 2,
         'GLhalfNV': 2,
         'GLclamped16i': 2, # clamped by glthread
+        'GLpacked16i': 2, # clamped by glthread
         'GLint': 4,
         'GLuint': 4,
         'GLbitfield': 4,

src/mesa/main/glthread_marshal.h

@@ -35,6 +35,12 @@
 #include "main/macros.h"
 #include "main/matrix.h"
 
+/* 32-bit signed integer clamped to 0..UINT16_MAX to compress parameters
+ * for glthread. All values < 0 and >= UINT16_MAX are expected to throw
+ * GL_INVALID_VALUE. Negative values are mapped to UINT16_MAX.
+ */
+typedef uint16_t GLpacked16i;
+
 /* 32-bit signed integer clamped to 16 bits. */
 typedef int16_t GLclamped16i;