166 	 * Bit width field in the GAS is only one byte long, 255 max.  in acpi_tb_init_generic_address()
178 				    "%s - 32-bit FADT register is too long (%u bytes, %u bits) "  in acpi_tb_init_generic_address()
188 	 * The 64-bit Address field is non-aligned in the byte packed  in acpi_tb_init_generic_address()
206  *              address32           - 32-bit address of the register
207  *              address64           - 64-bit address of the register
209  * RETURN:      The resolved 64-bit address
211  * DESCRIPTION: Select between 32-bit and 64-bit versions of addresses within
217  * the 32-bit and 64-bit address fields (FIRMWARE_CTRL/X_FIRMWARE_CTRL and
222  * By default, as per the ACPICA specification, a valid 64-bit address is
223  * used regardless of the value of the 32-bit address. However, this
234 		/* 64-bit address is zero, use 32-bit address */  in acpi_tb_select_address()
241 		/* Address mismatch between 32-bit and 64-bit versions */  in acpi_tb_select_address()
244 				   "32/64X %s address mismatch in FADT: "  in acpi_tb_select_address()
245 				   "0x%8.8X/0x%8.8X%8.8X, using %u-bit address",  in acpi_tb_select_address()
249 				   64));  in acpi_tb_select_address()
251 		/* 32-bit address override */  in acpi_tb_select_address()
258 	/* Default is to use the 64-bit address */  in acpi_tb_select_address()
400  *              Expand 32-bit addresses to 64-bit as necessary. Also validate
406  * Notes on 64-bit register addresses:
408  * After this FADT conversion, later ACPICA code will only use the 64-bit "X"
411  * The 64-bit X fields are optional extensions to the original 32-bit FADT
414  * 32-bit V1.0 fields to the 64-bit X fields if the 64-bit X field is originally
417  * For ACPI 1.0 FADTs (that contain no 64-bit addresses), all 32-bit address
418  * fields are expanded to the corresponding 64-bit X fields in the internal
421  * For ACPI 2.0+ FADTs, all valid (non-zero) 32-bit address fields are expanded
422  * to the corresponding 64-bit X fields, if the 64-bit field is originally
423  * zero. Adhering to the ACPI specification, we completely ignore the 32-bit
424  * field if the 64-bit field is valid, regardless of whether the host OS is
425  * 32-bit or 64-bit.
470 	 * Expand the 32-bit DSDT addresses to 64-bit as necessary.  in acpi_tb_convert_fadt()
471 	 * Later ACPICA code will always use the X 64-bit field.  in acpi_tb_convert_fadt()
483 	/* Examine all of the 64-bit extended address fields (X fields) */  in acpi_tb_convert_fadt()
487 		 * Get the 32-bit and 64-bit addresses, as well as the register  in acpi_tb_convert_fadt()
506 		 * Expand the ACPI 1.0 32-bit addresses to the ACPI 2.0 64-bit "X"  in acpi_tb_convert_fadt()
508 		 * the 64-bit address structures.  in acpi_tb_convert_fadt()
511 		 * Now always use the 64-bit address if it is valid (non-zero), in  in acpi_tb_convert_fadt()
512 		 * accordance with the ACPI specification which states that a 64-bit  in acpi_tb_convert_fadt()
513 		 * address supersedes the 32-bit version. This behavior can be  in acpi_tb_convert_fadt()
516 		 * During 64-bit address construction and verification,  in acpi_tb_convert_fadt()
531 		 * Note: space_id is always I/O for 32-bit legacy address fields  in acpi_tb_convert_fadt()
540 							   "32/64X address mismatch in FADT/%s: "  in acpi_tb_convert_fadt()
541 							   "0x%8.8X/0x%8.8X%8.8X, using %u-bit address",  in acpi_tb_convert_fadt()
546 							   ? 32 : 64));  in acpi_tb_convert_fadt()
552 				 * 64-bit X length field.  in acpi_tb_convert_fadt()
555 				 * 64-bit GAS structure can accommodate, 0xFF bits).  in acpi_tb_convert_fadt()
561 							   "32/64X length mismatch in FADT/%s: %u/%u",  in acpi_tb_convert_fadt()
570 			 * Hardware register access code always uses the 64-bit fields.  in acpi_tb_convert_fadt()
571 			 * So if the 64-bit field is zero or is to be overridden,  in acpi_tb_convert_fadt()
572 			 * initialize it with the 32-bit fields.  in acpi_tb_convert_fadt()
573 			 * Note that when the 32-bit address favor is specified, the  in acpi_tb_convert_fadt()
574 			 * 64-bit fields are always re-initialized so that  in acpi_tb_convert_fadt()
576 			 * configured to the values to trigger a 32-bit compatible  in acpi_tb_convert_fadt()