INVEXCEPTN, Exception while above ASTDEL

[John Travell]

The code from TCPIP$INTERNET_SERVICES+20860: looks like the entry into a subroutine, and the cause of this crash as far as the evidence goes is that the calling routine supplied address FFFFFFFF.82430640 in R16.

The relevant instructions in the called routine...
TCPIP$INTERNET_SERVICES+20884: ADDL R31,R16,R1 copied this supposedly bogus value to R1, then
TCPIP$INTERNET_SERVICES+20890: SUBL R1,#X40,R0 does the next step.
TCPIP$INTERNET_SERVICES+208A0: LDL R2,#X0030(R0) loads what is actually the low l/w of WCB$Q_OPENTIME into R2, we don't know yet what it should have been. Finally,
TCPIP$INTERNET_SERVICES+208AC: LDQ R26,#X0008(R2) tries to access an unaligned address, and suffers the accvio...

The calling routine probably did a JSR R26,(R26) from FFFFFFFF.80669E3C.
You could do an ex/i 80669E3C-20;30 to check what is there.

We will NOT solve this without the listings for the relevant version of TCPIP, and getting access to those is like pulling hens teeth...
As Volker says, please (if you have not yet done so) log a call with HP, with the expectation of this being raised to engineering, and report the results back when they give you a solution.

[Vladimir Fabecic]

Hello
I got a response from HP support.
It says "This crash is caused by a known FTP related TCPIP 5.5 problem. The fix will be in the upcoming TCPIP 5.5 ECO2 kit."
They sent me attachment that contains 2 images to fix this problem.
So, looks like VMS support is doing the job well.

[Volker Halle]

Vladimir,

thanks for reporting the solution. Which images did you get ?

Volker.

[STEPHEN CONWAY_1]

> getting access to those is like pulling hens teeth...

That's job security man :)

> So, looks like VMS support is doing the job well.

Your welcome.

> Which images did you get ?

It was a bug in TCPIP$INTERNET_SERVICES, and we always ship the ACP (TCPIP$INETACP) along with a new driver.

Note: TCP/IP V5.5-EC2 release is near. Will contain five fixes to NET.

Steve

[Vladimir Fabecic]

Backup list shows tho images:
TCPIP$INETACP.EXE_ECO_E_A_V55ECO1;1
TCPIP$INTERNET_SERVICES.EXE_ECO_E_A_V55ECO1;1
So that means that TCPIP$INETACP.EXE and TCPIP$INTERNET_SERVICES.EXE should be replaced by next ECO.