[OpenVMS documentation]
[Site home] [Send comments] [Help with this site] [How to order documentation] [OpenVMS site] [Compaq site]
Updated: 11 December 1998

OpenVMS Calling Standard


Previous Contents Index

3.4.7 Procedure Descriptor for Null Frame Procedures

The null frame procedure descriptor built by a compiler provides information about a procedure with no frame. The size of the descriptor is 16 bytes (defined by PDSC$K_NULL_SIZE).

The fields defined in the null frame descriptor are illustrated in Figure 3-7 and described in Table 3-5.

Figure 3-7 Null Frame Procedure Descriptor (PDSC) Format


Table 3-5 Contents of Null Frame Procedure Descriptor (PDSC)
Field Name Contents
PDSC$W_FLAGS The PDSC descriptor flag bits <15:0> are defined as follows:
PDSC$V_KIND A 4-bit field <3:0> that identifies the type of procedure descriptor. For a null frame procedure, this field must specify a value 8 (defined by constant PDSC$K_KIND_NULL).
Bits 4--7 Must be 0.
PDSC$V_REI_RETURN Bit 8. If set to 1, the procedure expects the stack at entry to be set, so an REI instruction correctly returns from the procedure. Also, if set, the contents of the PDSC$B_SAVE_RA field are unpredictable and the return address is found on the stack.
Bit 9 Must be 0 (reserved).
PDSC$V_BASE_FRAME For compiled code, this bit must be 0. If set to 1, indicates the logical base frame of a stack that precedes all frames corresponding to user code. The interpretation and use of this frame and whether there are any predecessor frames is system software defined (and subject to change).
Bit 11 Must be 0 (reserved).
PDSC$V_NATIVE For compiled code, this bit must be set to 1.
PDSC$V_NO_JACKET For compiled code, this bit must be set to 1.
PDSC$V_TIE_FRAME For compiled code, this bit must be 0. Reserved for use by system software.
Bit 15 Must be 0 (reserved).
PDSC$V_FUNC_RETURN A 4-bit field <11:8> that describes which registers are used for the function value return (if there is one) and what format is used for those registers.

Table 3-7 lists and describes the possible encoded values of PDSC$V_FUNC_RETURN.

PDSC$W_SIGNATURE_OFFSET A 16-bit signed byte offset from the start of the procedure descriptor. This offset designates the start of the procedure signature block (if any). A 0 in this field indicates that no signature information is present. Note that in a bound procedure descriptor (as described in Section 3.7.4), signature information might be present in the related procedure descriptor. A 1 in this field indicates a standard default signature. An offset value of 1 is not otherwise a valid offset because both procedure descriptors and signature blocks must be quadword aligned.
PDSC$Q_ENTRY The absolute address of the first instruction of the entry code sequence for the procedure.

3.5 Procedure Signatures

As a means of enhancing certain aspects of program interoperation between images built from native Alpha code and images translated from VAX code, native Alpha compilers can optionally generate information that describes the parameters of a procedure. This auxiliary information is called procedure signature information, or sometimes just signature information.

Signature information is used when a call from a native procedure passes control to a translated procedure and vice versa. Translated VAX code on Alpha processors uses a VAX argument list and function return conventions as described in Sections 2.4 and 2.5. Here, the signature information is used to control how passed and returned arguments according to Alpha conventions are manipulated and placed for use by translated VAX code and vice versa.

If a procedure is compiled with signature information, PDSC$W_SIGNATURE_OFFSET contains a byte offset from the procedure descriptor to the start of a procedure signature control block. The maximum size of the procedure signature control block is 72 bytes (defined by constant PSIG$K_MAX_SIZE). The fields defined in the procedure signature information block are illustrated in Figure 3-8 and described in Table 3-6.

Figure 3-8 Procedure Signature Information Block (PSIG)


Table 3-6 Contents of the Procedure Signature Information Block (PSIG)
Field Name Contents
PSIG$V_FUNC_RETURN A 4-bit field <3:0> that describes which registers are used for the function value return (if there is one) and what format is used for those registers.

Table 3-7 lists and describes the possible encoded values of PSIG$V_FUNC_RETURN.

PSIG$V_REG_ARG_INFO A 24-bit field <27:4> that is divided into six groups of four bits that correspond to the six arguments that can be passed in registers. These groups describe how each of the first six arguments are to be passed in registers of the first group (bits <7:4>) describing the first argument.

Each register argument signature group is encoded as follows:
Value Name Meaning1,2
0 RASE$K_RA_NOARG Argument is not present
1 RASE$K_RA_Q 64-bit argument passed in an integer register
2 RASE$K_RA_I32 32-bit argument sign extended to 64 bits passed in an integer register
3 RASE$K_RA_U32 32-bit unsigned argument zero extended to 64 bits passed in an integer register
4 RASE$K_RA_FF F_floating argument passed in a floating-point register
5 RASE$K_RA_FD D_floating argument passed in a floating-point register
6 RASE$K_RA_FG G_floating argument passed in a floating-point register
7 RASE$K_RA_FS S_floating argument passed in a floating-point register
8 RASE$K_RA_FT T_floating argument passed in a floating-point register
9--15   Reserved for future use

PSIG$V_SUMMARY A 4-bit field <31:28> that contains coded argument signature information as follows:
Bit Name Meaning
0, 1 PSIG$M_SU_ASM Summary of arguments 7 through PSIG$B_ARG_COUNT:
00 = All arguments are 64-bit or not used
01 = All arguments are 32-bit sign extended or not used
10 = Reserved
11 = Other (not 00 or 01)
2 PSIG$M_SU_VLIST VAX formatted argument list expected
3   Must be 0 (reserved)

PSIG$M_SU_ASM values of 00 and 01 (binary) allow a quick test for the occurrence of either an all 32-bit or an all 64-bit argument list. The values for the PSIG$V_MEMORY_ARG_INFO field must be valid even when these occurrences apply.

PSIG$B_ARG_COUNT Unsigned byte (bits 0--7) that specifies the number of 64-bit argument items described in the argument signature information. This count includes the first six arguments.
PSIG$V_MEMORY_ARG_INFO Array of 2-bit values that describe each of arguments 7 through PSIG$B_ARG_COUNT. PSIG$S_MEMORY_ARG_INFO data is only defined for the arguments described by PSIG$B_ARG_COUNT. These memory argument signature bits are defined as follows:
Value Name Meaning1
0 MASE$K_MA_Q 64-bit argument
1   Reserved
2 MASE$K_MA_I32 32-bit sign-extended argument
3   Reserved


1For more specific impact on the converted field value, see Section 3.5.1.
2The X_floating and X_floating complex data types do not appear in this table because these types are not passed using the by value mechanism (see Section 3.8.5.1).

Table 3-7 Function Return Signature Encodings
Value Name Meaning1,2
0 PSIG$K_FR_I64 64-bit result in R0
or No function result provided
or First parameter mechanism used
1 PSIG$K_FR_D64 64-bit result with low 32 bits sign extended in R0 and high 32 bits sign extended in R1
2 PSIG$K_FR_I32 32-bit sign extended to 64-bit result in R0
3 PSIG$K_FR_U32 32-bit unsigned result (zero extended) in R0
4 PSIG$K_FR_FF F_floating result in F0
5 PSIG$K_FR_FD D_floating result in F0
6 PSIG$K_FR_FG G_floating result in F0
7 PSIG$K_FR_FS S_floating result in F0
8 PSIG$K_FR_FT T_floating result in F0
9, 10   Reserved for future use
11 PSIG$K_FR_FFC F_floating complex result in F0 and F1
12 PSIG$K_FR_FDC D_floating complex result in F0 and F1
13 PSIG$K_FR_FGC G_floating complex result in F0 and F1
14 PSIG$K_FR_FSC S_floating complex result in F0 and F1
15 PSIG$K_FR_FTC T_floating complex result in F0 and F1


1For more specific impact on the converted field value, see Section 3.5.1.
2The X_floating and X_floating complex data types do not appear in this table because these types are not passed using the by value mechanism (see Section 3.8.5.1).

3.5.1 Call Parameter PSIG Conversions

Where a VAX image is translated to an Alpha image, the VAX registers R0--15 are represented using the lower half of the corresponding Alpha registers R0--15 at call interface boundaries. No "type conversion" is performed in making parameters from either native or translated code available to each other. However, it is important to understand the effects of the PSIG field values when interfacing between native and translated environments.

Note that an address under OpenVMS Alpha is described using RASE$K_RA_I32 or MASE$K_MA_I32 as appropriate.

3.5.1.1 Native-to-Translated Code PSIG Conversions

The specific impact of the native-to-translated call conversions of the PSIG$V_REG_ARG_INFO and the PSIG$V_FUNC_RETURN field values are listed in Table 3-8.

Table 3-8 Native-to-Translated Conversion of the PSIG Field Values
Name Impact
PSIG$V_REG_ARG_INFO Field Conversions
RASE$K_RA_Q The low-order 32 bits of the integer register contents are used to fill the first of two longword entries in the VAX formatted argument list, while the high-order 32 bits are used to fill the second longword entry. This counts as two arguments in the VAX formatted argument list.
RASE$K_RA_I32
RASE$K_RA_U32
The low-order 32 bits of the integer register contents are used to fill one longword entry in the VAX formatted argument list passed to the translated procedure. The high-order 32 bits are ignored. This counts as one argument in the VAX formatted argument list.
RASE$K_RA_FF The single-precision contents of a floating-point register are used to fill one longword entry in the VAX formatted argument list passed to the translated procedure. This counts as one argument in the VAX formatted argument list. The Alpha store instruction STF is used to place the register contents into memory.
RASE$K_RA_FD
RASE$K_RA_FG
The double-precision contents of a floating-point register are used to fill two longword entries in the VAX formatted argument list passed to the translated procedure. This counts as two arguments in the VAX formatted argument list. The Alpha store instruction STG is used to place the register contents into memory.
RASE$K_RA_FS
RASE$K_RA_FT
Undefined.
PSIG$V_MEMORY_ARG_INFO Field Conversions
MASE$K_MA_Q
MASE$K_MA_I32
These convert like the RASE$K_RA_Q and RASE$K_RA_I32 field conversions, except that the Alpha argument list entry is stored in memory (rather than a register).
PSIG$V_FUNC_RETURN Field Conversions
PSIG$K_FR_I64 The translated code is returning a 64-bit result split between R0 and R1. The low-order 32 bits of R1 are shifted left and combined with the low-order 32 bits of R0 to form the 64-bit result that is returned to the native caller.
PSIG$K_FR_D64 The translated code is returning a 64-bit result split between R0 and R1. Both R0 and R1 are sign extended from 32 to 64 bits and returned to the native caller in place.
PSIG$K_FR_I32
PSIG$K_FR_U32
The translated code is returning a 32-bit result in R0. R0 is sign extended from 32 to 64 bits and returned to the native caller in place.
PSIG$K_FR_FF The single-precision contents of the result in R0 is loaded into Alpha register F0.
PSIG$K_FR_FD
PSIG$K_FR_FG
The double-precision contents in registers R0 and R1 are combined and loaded into Alpha register F0.
PSIG$K_FR_FS
PSIG$K_FR_FT
Undefined.
PSIG$K_FR_FFC The single-precision complex contents in registers R0 and R1 are loaded into Alpha registers F0 and F1.
PSIG$K_FR_FDC
PSIG$K_FR_FGC
The translated code is returning a double-precision complex result using the hidden first parameter method (by reference). The storage for the result is allocated prior to the call and the address is passed as the extra parameter. Upon return, the result is copied from the temporary storage into the Alpha floating-point registers and returned to the native caller.
PSIG$K_FR_FSC
PSIG$K_FR_FTC
Undefined.

In all 64-bit cases, the longword at the lower memory address forms the earlier argument in the VAX formatted argument list. Also, for single-precision floating-point types, the unused 32 bits of an Alpha 64-bit argument list entry are undefined.

3.5.1.2 Translated-to-Native Code PSIG Conversions

The specific impact of the translated-to-native call conversions of the PSIG$V_REG_ARG_INFO and the PSIG$V_FUNC_RETURN field values are listed in Table 3-9.

Table 3-9 Translated-to-Native Conversion of the PSIG Field Values
Name Impact
PSIG$V_REG_ARG_INFO Field Conversions
RASE$K_RA_Q The contents of two successive longwords from the VAX formatted argument list are combined to form a single quadword value that is placed in an integer register. This counts as one argument in the Alpha argument list.
RASE$K_RA_I32
RASE$K_RA_U32
The contents of one longword entry from the VAX formatted argument list is sign extended and placed in the integer register. This counts as one argument in the Alpha argument list.
RASE$K_RA_FF A single longword entry from the VAX formatted argument list is used to form a floating-point value in a floating-point register. This counts as one argument in the Alpha argument list. The Alpha load instruction LDF is used to place the argument in the floating-point register.
RASE$K_RA_FD
RASE$K_RA_FG
Two longword entries from the VAX formatted argument list are combined to form a single floating-point value in a floating-point register. This counts as one argument in the Alpha argument list. The Alpha load instruction LDG is used to place the argument in the floating-point register.
RASE$K_RA_FS
RASE$K_RA_FT
Undefined.
PSIG$V_MEMORY_ARG_INFO Field Conversions
MASE$K_MA_Q
MASE$K_MA_I32
These convert like RASE$K_RA_Q and RASE$K_RA_I32 field conversions, except that the Alpha argument list entry is stored in memory (rather than a register). 1
PSIG$V_FUNC_RETURN Field Conversions
PSIG$K_FR_I64 The native code is returning a 64-bit result in R0. The high 32 bits of R0 are moved to the low half of R1 and sign extended, and then R0 is sign extended from 32 to 64 bits. The 64-bit result is then returned to the translated caller in R0 and R1.
PSIG$K_FR_D64 The native code is returning a 64-bit result split between R0 and R1. Both R0 and R1 are sign extended from 32 to 64 bits and returned to the translated caller in place.
PSIG$K_FR_I32
PSIG$K_FR_U32
The native code is returning a 32-bit result in R0. R0 is sign extended from 32 to 64 bits and the result is then returned in place to the translated caller.
PSIG$K_FR_FF The single-precision result in Alpha register F0 is stored in the low-order half of register R0. 1
PSIG$K_FR_FD
PSIG$K_FR_FG
The double-precision result in Alpha register F0 is stored in the low-order halves of registers R0 and R1.
PSIG$K_FR_FS
PSIG$K_FR_FT
Undefined.
PSIG$K_FR_FFC The single-precision complex result in Alpha registers F0 and F1 is stored in the low-order halves of registers R0 and R1. 1
PSIG$K_FR_FDC
PSIG$K_FR_FGC
The native code is returning a double-precision complex result in the Alpha floating-point registers. The result is copied into the storage given by the hidden first parameter passed by the translated caller.
PSIG$K_FR_FSC
PSIG$K_FR_FTC
Undefined.


1Note that for single-precision floating-point types, the unused 32 bits of an Alpha 64-bit argument list entry are undefined.


Previous Next Contents Index

[Site home] [Send comments] [Help with this site] [How to order documentation] [OpenVMS site] [Compaq site]
[OpenVMS documentation]

Copyright © Compaq Computer Corporation 1998. All rights reserved.

Legal
5973PRO_004.HTML