Interface Protection Using Parity

The AHB parity feature adds check signals that protect groups of AHB interface signals. It is intended for systems that need detection of transient or functional wire errors across component boundaries.

The Check_Type property controls whether check signals exist:

Check_Type	Meaning
False	No check signals. If the property is not declared, it is treated as False.
Odd_Parity_Byte_All	Odd parity checking is included for all defined signal groups.

Odd parity means the protected payload bits plus the check bit are driven so the group contains an odd number of asserted bits.

Parity signal rules

Rule	Meaning
Most parity bits cover up to 8 payload bits	Keeps generation logic small enough for timing.
Critical control signals can use one check bit	For a single protected control bit, odd parity is the inversion of that bit.
Wider check signals map by byte groups	Check bit n covers payload bits [(8n+7):8n]; the most significant group can cover fewer than 8 bits.
Check signals are driven when their Check Enable term is true	Validity depends on the protected signal group.
Data parity covers all payload bits	Even inactive byte lanes still need correct HWDATACHK/HRDATACHK generation.
Missing protected signals are treated as LOW	If some covered optional signals are absent, parity assumes zero for those bits.
Omit unused check signals	If none of the covered signals exist on an interface, the corresponding check signal is absent.

Examples of check signal groups include:

Check signal	Protects	Check enable idea
HTRANSCHK	HTRANS	HRESETn
HADDRCHK	HADDR	HRESETn
HCTRLCHK1	HBURST, HMASTLOCK, HWRITE, HSIZE, HNONSEC	HTRANS != IDLE
HCTRLCHK2	HEXCL, HMASTER	HTRANS != IDLE
HPROTCHK	HPROT	HTRANS != IDLE
HWSTRBCHK	HWSTRB	Write data phase
HWDATACHK	HWDATA	Write data phase
HRDATACHK	HRDATA	Read data phase and HREADY
HREADYCHK, HREADYOUTCHK	Ready signaling	HRESETn
HRESPCHK	HRESP, HEXOKAY	Data phase
HSELxCHK	HSELx	HRESETn
User check signals	HAUSER, HWUSER, HRUSER, HBUSER	Matching User signal phase

Error detection behavior

The AHB parity scheme defines the signals, not a single mandatory recovery action. When parity mismatch is detected, the receiver can terminate or propagate the transaction, correct or propagate the parity error, update memory or leave it untouched, or signal error through another system mechanism such as an interrupt. The transaction response is independent of parity detection unless the system explicitly connects the two.

Parity protection checklist

Check	Reason
Check signals are generated from the correct cycle	Parity must match protected data.
Check enable conditions match the protected phase	Prevents false parity failures on invalid cycles.
Data parity includes inactive byte lanes	The parity rule covers the payload bits, not only active strobes.
Parity error behavior is documented	Software and safety monitors need a policy.
Parity configurations are included in compatibility tests	Configurable IP fails easily at optional-feature boundaries.

AHB Protocol