For the TRUSTED RECEIVE commands of the ATA8-ACS command, in the ATA8-ACS spec, the total data length shall be 512 bytes. Pad bytes are appended as needed to meet this requirement. Pad bytes shall have a value of 00h.
For the SECURITY PROTOCOL IN commands of the SPC-4 command, In the SPC-4 spec, when INC_512 is 0, the ALLOCATION LENGTH field expresses the number of bytes to be transferred. It means any value. If the length is larger than 8 bytes, the byte 6-7 indicate the SUPPORTED SECURITY PROTOCOL LIST LENGTH. If the length is larger than (SECURITY PROTOCOL LIST LENGTH + 8), all are returned and plus the pad data. Best Regards Eric -----Original Message----- From: Tian, Feng Sent: Thursday, September 1, 2016 10:42 AM To: Ramesh R. <[email protected]>; edk2-devel <[email protected]>; Jin, Eric <[email protected]> Cc: Tian, Feng <[email protected]> Subject: RE: BootableImageSupportTest\StorageSecurityCommandProtocolTest I checked the ATA spec, it says the transfer length of "Trust-Send" ATA cmd should be 512. But for NVMe and other SCSI device, I didn't see any length limitation on "Security Protocol In" cmd with security protocol field 0 and security protocol specific field 0. It seems user could pass in any length value to get security protocol information. And last, user could get the whole one by passing down "supported security protocol list length" + 8. Ramesh, do you meet real failure case? Eric, what's your opinion on this? Thanks Feng -----Original Message----- From: Ramesh R. [mailto:[email protected]] Sent: Wednesday, August 31, 2016 1:20 AM To: Tian, Feng <[email protected]>; edk2-devel <[email protected]>; Jin, Eric <[email protected]> Subject: RE: BootableImageSupportTest\StorageSecurityCommandProtocolTest Hi Feng, Any update or suggestion on this? Can we consider this as SCT tool issue and would be fixed in next version ? Thanks, Ramesh -----Original Message----- From: Tian, Feng [mailto:[email protected]] Sent: 26 August 2016 12:54 To: Ramesh R.; edk2-devel; Jin, Eric Cc: Tian, Feng Subject: RE: BootableImageSupportTest\StorageSecurityCommandProtocolTest Yes, I agree it's weird. We are looking at this and will get back to you if we have findings. Thanks Feng -----Original Message----- From: edk2-devel [mailto:[email protected]] On Behalf Of Ramesh R. Sent: Thursday, August 25, 2016 4:44 PM To: edk2-devel <[email protected]> Subject: [edk2] BootableImageSupportTest\StorageSecurityCommandProtocolTest Hi, When the we run the "BootableImageSupportTest\StorageSecurityCommandProtocolTest" test on the NVME devices we are getting into error because of the below testing code. // // According to TCG definition, when the Security Protocol field is set to 00h, and SP // Specific is set to 0000h in a TRUSTED RECEIVE command, return security protocol // information. This Command is not associated with a security send command // Status = StorageSecurityCommand->ReceiveData ( StorageSecurityCommand, BlockIo->Media->MediaId, 100000000, // Timeout 10-sec 0, // SecurityProtocol 0, // SecurityProtocolSpecifcData 10, // PayloadBufferSize, DataBuffer, // PayloadBuffer &RcvDataSize ); // // for ATA8-ACS SecurityProtocol, 512 byte is a request // if (IsAtaDevice) { if((Status == EFI_DEVICE_ERROR) || (Status == EFI_WARN_BUFFER_TOO_SMALL)){ AssertionType = EFI_TEST_ASSERTION_PASSED; } else { AssertionType = EFI_TEST_ASSERTION_FAILED; } } else { if((!EFI_ERROR(Status)) || (Status == EFI_WARN_BUFFER_TOO_SMALL)){ AssertionType = EFI_TEST_ASSERTION_PASSED; } else { AssertionType = EFI_TEST_ASSERTION_FAILED; } } For Ata devices, EFI_DEVICE_ERROR considered as valid error case and for the Nvme ( Non ATA) device it's considered as error. Could you please let us know why there is difference in this case ?. Thanks, Ramesh _______________________________________________ edk2-devel mailing list [email protected] https://lists.01.org/mailman/listinfo/edk2-devel _______________________________________________ edk2-devel mailing list [email protected] https://lists.01.org/mailman/listinfo/edk2-devel
