Re: Agilent 82357B repeatable hard failure

dave penkler <[email protected]> Mon, 2 Oct 2023 14:31:48 +0200

Newsgroups	gmane.linux.hardware.gpib.general
Message-ID	<CAL=kjP15-wuQW761wHEayjD7fafRz_0VF5_3PMr8q8mqt-vDEA@mail.gmail.com>

Hi Jim,
I refactored the take control function to avoid reading the status and
checking ATN when called from read.
See the attached patch.
Would this work for you ?
-Dave

On Sun, 1 Oct 2023 at 14:48, Jim Houston <[email protected]> wrote:

> Hi Dave, Everyone,
>
> The loop reading the status and checking for ATN in
> agilent_82357a_take_control
> was helpful in finding the problems with the ADSR but maybe it isn't
> needed anymore.
> I did a couple tests.  I tried setting timeout to 1 to check that ATN is
> there on the first read.
> Then I took out the loop and the timeout check.  I ran the
> hp3478_read_calibration script
> in a loop for a few minutes.  It works.
>
> I checked the times for a hp3478_read_calibration run and removing the
> check
> for the ATN reduced the time from 7.6 seconds to 7.3 seconds.
>
> JIm
>
>
>

_______________________________________________
Linux-gpib-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-gpib-general

agilent_82357a.patch (text/x-patch, 1.8 KB)

Index: drivers/gpib/agilent_82357a/agilent_82357a.c
===================================================================
--- drivers/gpib/agilent_82357a/agilent_82357a.c	(revision 2063)
+++ drivers/gpib/agilent_82357a/agilent_82357a.c	(working copy)
@@ -34,7 +34,7 @@
 
 unsigned int agilent_82357a_update_status( gpib_board_t *board, unsigned int clear_mask );
 
-int agilent_82357a_take_control(gpib_board_t *board, int synchronous);
+int agilent_82357a_take_control_internal(gpib_board_t *board, int synchronous);
 
 static void agilent_82357a_bulk_complete(struct urb *urb PT_REGS_ARG)
 {
@@ -577,7 +577,7 @@
 	/* Fix for a bug in 9914A that does not return the contents of ADSR
            when the board is in listener active state and ATN is not asserted.
            Set ATN here to obtain a valid board level ibsta  */
-	agilent_82357a_take_control(board,0);
+	agilent_82357a_take_control_internal(board,0);
 
 	//FIXME check trailing flags for error
 	return retval;
@@ -737,13 +737,11 @@
 	return agilent_82357a_generic_write(board, buffer, length, 1, 0, bytes_written);
 }
 
-int agilent_82357a_take_control(gpib_board_t *board, int synchronous)
+int agilent_82357a_take_control_internal(gpib_board_t *board, int synchronous)
 {
 	agilent_82357a_private_t *a_priv = board->private_data;
-	const int timeout = 10;
 	struct agilent_82357a_register_pairlet write;
 	int retval;
-	int i;
 
 /* It looks like the 9914 does not handle tcs properly.
    See comment above tms9914_take_control_workaround() in
@@ -762,6 +760,14 @@
 	{
 		printk("%s: agilent_82357a_write_registers() returned error\n", __FUNCTION__);
 	}
+	return retval;
+}
+int agilent_82357a_take_control(gpib_board_t *board, int synchronous)
+{
+	const int timeout = 10;
+	int i;
+
+	agilent_82357a_take_control_internal( board, synchronous );
 	// busy wait until ATN is asserted
 	for(i = 0; i < timeout; ++i)
 	{