Title: ATA/ATAPI-7 Power-On Reset Diagram

Mark,

Yes, this was the intention. If you check the HP1 state in figure 39 (page 78) you see in the text description that the host may continue sending COMRESET, or send and cease sending COMRESET. If a device is hot plugged, and the host is sending continuous COMRESET, then the device cannot transition states until COMRESET is de-asserted. This can be seen in state DP1 of figure 40. After the host de-asserts COMRESET, the device Phy layer continues the COMRESET Sequence by responding with COMINIT. This last bit is handled in the PHY Initialization state machine(s). Figure 93, which you refer to is at the Command level, and the command level notifies the Transport level to send the Register DH FIS. The transport level will only send the FIS if the Phy reports it is PHYRDY and the Link is ready to transmit. In most cases, the Phy will be ready and the Link will be ready to transmit long before the command layer is completed with diagnostics and initialization. It is true that the device cannot send the register FIS during COMRESET, as both the host and the device Phys are not PHYRDY  (the host is not in HP11:HP_HR_Ready the device is not in DP10:DR_Ready).

John

Western Digital

 


From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Mark Overby
Sent: Tuesday, August 03, 2004 10:26 PM
To: [EMAIL PROTECTED]
Subject: [t13] ATA/ATAPI-7 Power-On Reset Diagram

 

Section 17.2 of Vol 3.

The section implies that this is the power-on reset as well as the COMRESET protocol diagrams. In the case of a drive being "hot-plugged", the host will have already sent COMRESET, but should still advance it's state machines when the device sends COMINIT. However, by a strict reading of this, the device cannot send the initial register FIS until such time as the host has stopped sending COMRESET.

Is this the intention?

Reply via email to