AT697E datasheet(91/154 Pages) ATMEL

91 / 154 page

AT697E

4226G–AERO–05/09

Trap Types

The ftt field can be read by the STFSR instruction. An LDFSR instruction does not affect

ftt field.

Table 39. Trap Type Definition

Floating Point Condition Code

Table 40. FCC Field Definition

11..10

fcc

Contain the FPU condition codes. These bits are updated by floating-point compare instructions

(FCMP and FCMPE). They are read and written by the STFSR and LDFSR instructions,

respectively. FBfcc bases its control transfer on this field.

9..5

aexc

Accumulate IEEE floating-point exceptions while fp_exception traps are disabled using the

TEM field. After an FPop completes, the TEM and cexc fields are logically anded together. If the

result is nonzero, an fp_exception trap is generated; otherwise, the new cexc field is or’d into

the aexc field. Thus, while traps are masked, exceptions are accumulated in the aexc field.

4..0

cexc

Indicate that one or more IEEE floating-point exceptions were generated by the most recently

executed FPop instruction. The absence of an exception causes the corresponding bit to be

cleared.

Bit Number

Mnemonic

Description

Name

Description

none

No trap.

IEEE_exception

An IEEE_754_exception floating-point trap type indicates that a floating-point exception occurred that conforms

to the ANSI/IEEE Standard 754-1985. The exception type is encoded in the cexc field.

Unfinished_FPop

An unfinished_FPop indicates that an implementation’s FPU was unable to generate correct results or

exceptions

unimplemented_FPop

An unimplemented_FPop indicates that an implementation’s FPU decoded an FPop that it does not implement.

In this case, the cexc field is unchanged

sequence_error

A sequence_error indicates one of three abnormal error conditions in the FPU, all caused by erroneous

supervisor software:

- An attempt was made to execute a floating-point instruction when the FPU was not able to accept one. This

type of sequence_error arises from a logic error in supervisor software that has caused a previous floating-point

trap to be incompletely serviced (for example, the floating-point queue was not emptied after a previous floating-

point exception).

- An attempt was made to execute a STDFQ instruction when the floatingpoint deferred-trap queue (FQ) was

empty, that is, when FSR.qne = 0. (Note that generation of sequence_error is recommended, but not required in

this case)

hardware error

A hardware_error indicates that the FPU detected a catastrophic internal error, such as an illegal state or a parity

error on an f register access. If a hardware_error occurs during execution of user code, it may not be possible to

recover sufficient state to continue execution of the user application.

invalid register

An invalid_fp_register trap type indicates that one (or more) operands of an FPop are misaligned, that is, a

double-precision register number is not 0 mod 2, or a quadruple-precision register number is not 0 mod 4. It is

recommended that implementations generate an fp_exception trap with FSR.ftt = invalid_fp_register in this case,

but an implementation may choose not to generate a trap.

FCC

Description

f rs1 = f rs2

f rs1 < f rs2

f rs1 > f rs2

AT697E Datasheet(PDF) 91 Page - ATMEL Corporation

AT697E Datasheet(HTML) 91 Page - ATMEL Corporation

Html Pages

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 ...More

Datasheet Download

Link URL