Cortex-A9 Floating-Point Unit Technical Reference Manual: 2.5.3. Floating-Point Exception Register

The FPEXC Register characteristics are:

Purpose: Provides global enable control of the Advanced SIMD and VFP extensions.
Usage constraints: Accessible in all FPU configurations, with restrictions. See Processor modes for accessing the FPU system registers.
Configurations: Available in all FPU configurations.
Attributes: See the register summary in Table 2.2.

Figure 2.3 shows the FPEXC Register bit assignments.

Figure 2.3. FPEXC Register bit assignments

Table 2.6 shows the FPEXC Register bit assignments.

Table 2.6. FPEXC Register bit assignments

Bits	Name	Function
[31]	EX	Exception bit: This bit reads-as-zero and ignores writes. The Cortex-A9 NEON MPE never requires asynchronous exception handling.
[30]	EN	Enable bit: b0 = VFP extension is disabled. b1 = VFP extension is enabled and operates normally. The EN bit is cleared to 0 at reset.
[29]	DEX	Defined synchronous instruction exceptional flag: b0 = no exception has occurred b1 = attempt to perform a VFP vector operation has been trapped^[a] The DEX bit is cleared to 0 at reset.
[28:26]	-	RAZ/WI.
[25:0]	-	UNK/SBZP.
^[a]The Cortex-A9 FPU hardware does not support the deprecated VFP short vector feature. Attempts to execute VFP data-processing instructions when the FPSCR.LEN field is non-zero result in the FPSCR.DEX bit being set and a synchronous Undefined instruction exception being taken. You can use software to emulate the short vector feature, if required.

You can access the FPEXC Register with the following VMSR instructions:

VMRS <Rd>, FPEXC ; Read Floating-Point Status and Control Register

VMSR FPEXC, <Rt> ; Write Floating-Point Status and Control Register

Copyright © 2008-2010 ARM. All rights reserved.	ARM DDI 0408F
Non-Confidential	ID050110
PDF version