Skip to contentSkip to conversation

scsi_transport_spi: fix domain validation failure from incorrect width setting

PATCH In review James Bottomley 2 participants 3 messages Sep 22, 2007 – Sep 24, 2007 SCSI

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 scsi_transport_spi: fix domain validation failure from incorrect width setting

The previous patch was a bit non-standard because it actually adjusted the user setting max_width. This would mean that if the user replaced the cable and then re-ran domain validation, it still wouldn't try wide transfers, which is wrong.

The attached patch fixes the problem by physically nailing the wide setting to what it deduces it should be for the whole of Domain Validation.

James

drivers/scsi/scsi_transport_spi.c +22 -6
index 6f56f87..bb8a726 100644
@@ -787,10 +787,12 @@ spi_dv_device_internal(struct scsi_device *sdev, u8 *buffer)
 	struct scsi_target *starget = sdev->sdev_target;
 	struct Scsi_Host *shost = sdev->host;
 	int len = sdev->inquiry_len;
+	int min_period = spi_min_period(starget);
+	int max_width = spi_max_width(starget);
 	/* first set us up for narrow async */
 	DV_SET(offset, 0);
 	DV_SET(width, 0);
-	
+
 	if (spi_dv_device_compare_inquiry(sdev, buffer, buffer, DV_LOOPS)
 	    != SPI_COMPARE_SUCCESS) {
 		starget_printk(KERN_ERR, starget, "Domain Validation Initial Inquiry Failed\n");
@@ -798,9 +800,13 @@ spi_dv_device_internal(struct scsi_device *sdev, u8 *buffer)
 		return;
 	}
 
+	if (!scsi_device_wide(sdev)) {
+		spi_max_width(starget) = 0;
+		max_width = 0;
+	}
+	
 	/* test width */
-	if (i->f->set_width && spi_max_width(starget) &&
-	    scsi_device_wide(sdev)) {
+	if (i->f->set_width && max_width) {
 		i->f->set_width(starget, 1);
 
 		if (spi_dv_device_compare_inquiry(sdev, buffer,
@@ -809,6 +815,11 @@ spi_dv_device_internal(struct scsi_device *sdev, u8 *buffer)
 		    != SPI_COMPARE_SUCCESS) {
 			starget_printk(KERN_ERR, starget, "Wide Transfers Fail\n");
 			i->f->set_width(starget, 0);
+			/* Make sure we don't force wide back on by asking
+			 * for a transfer period that requires it */
+			max_width = 0;
+			if (min_period < 10)
+				min_period = 10;
 		}
 	}
 
@@ -828,7 +839,8 @@ spi_dv_device_internal(struct scsi_device *sdev, u8 *buffer)
 
 	/* now set up to the maximum */
 	DV_SET(offset, spi_max_offset(starget));
-	DV_SET(period, spi_min_period(starget));
+	DV_SET(period, min_period);
+
 	/* try QAS requests; this should be harmless to set if the
 	 * target supports it */
 	if (scsi_device_qas(sdev)) {
@@ -837,14 +849,14 @@ spi_dv_device_internal(struct scsi_device *sdev, u8 *buffer)
 		DV_SET(qas, 0);
 	}
 
-	if (scsi_device_ius(sdev) && spi_min_period(starget) < 9) {
+	if (scsi_device_ius(sdev) && min_period < 9) {
 		/* This u320 (or u640). Set IU transfers */
 		DV_SET(iu, 1);
 		/* Then set the optional parameters */
 		DV_SET(rd_strm, 1);
 		DV_SET(wr_flow, 1);
 		DV_SET(rti, 1);
-		if (spi_min_period(starget) == 8)
+		if (min_period == 8)
 			DV_SET(pcomp_en, 1);
 	} else {
 		DV_SET(iu, 0);
@@ -862,6 +874,10 @@ spi_dv_device_internal(struct scsi_device *sdev, u8 *buffer)
 	} else {
 		DV_SET(dt, 1);
 	}
+	/* set width last because it will pull all the other
+	 * parameters down to required values */
+	DV_SET(width, max_width);
+
 	/* Do the read only INQUIRY tests */
 	spi_dv_retrain(sdev, buffer, buffer + sdev->inquiry_len,
 		       spi_dv_device_compare_inquiry);
#1 lore raw

Dear James!

Sorry for replying so late!

drivers/scsi/scsi_transport_spi.c

Quoted 12 lines · James Bottomley

The previous patch was a bit non-standard because it actually adjusted the user setting max_width. This would mean that if the user replaced the cable and then re-ran domain validation, it still wouldn't try wide transfers, which is wrong.

The attached patch fixes the problem by physically nailing the wide setting to what it deduces it should be for the whole of Domain Validation.

James

drivers/scsi/scsi_transport_spi.c +0 -0

[snip]

A self-compiled kernel based on the Ubuntu "gutsy" kernel sources (linux-source-2.6.22_2.6.22-12.36_all.deb, for whatever that is worth) boots successfully, after applying your patch (which applied cleanly)

Thanks so much for your help,

Stefan Boresch

PS: BTW, your original post (diagnosis of hardware issues) has also convinced my vendor to do some work to honor his warranty. So, my next job is to dig into the big box which just arrived by fed-ex containing all possible replacements (including a new motherboard ..). Maybe, I'll eventually get U320 running on this machine ;-)

Signature · 4 lines
Stefan Boresch
Institute for Computational Biological Chemistry
University of Vienna, Waehringerstr. 17       A-1090 Vienna, Austria
Phone: -43-1-427752715                        Fax:   -43-1-427752790
#2 lore raw
Quoted 16 lines · Stefan Boresch

Dear James!

Sorry for replying so late!

Quoted 11 lines · James Bottomley

The previous patch was a bit non-standard because it actually adjusted the user setting max_width. This would mean that if the user replaced the cable and then re-ran domain validation, it still wouldn't try wide transfers, which is wrong.

The attached patch fixes the problem by physically nailing the wide setting to what it deduces it should be for the whole of Domain Validation.

James

Quoted 8 lines

drivers/scsi/scsi_transport_spi.c

Quoted 1 line
drivers/scsi/scsi_transport_spi.c +0 -0

[snip]

A self-compiled kernel based on the Ubuntu "gutsy" kernel sources (linux-source-2.6.22_2.6.22-12.36_all.deb, for whatever that is worth) boots successfully, after applying your patch (which applied cleanly)

Thanks so much for your help,

You're very welcome.

Quoted 7 lines

Stefan Boresch

PS: BTW, your original post (diagnosis of hardware issues) has also convinced my vendor to do some work to honor his warranty. So, my next job is to dig into the big box which just arrived by fed-ex containing all possible replacements (including a new motherboard ..). Maybe, I'll eventually get U320 running on this machine ;-)

OK ... I'd replace the cable first ... and check for bent pins in the connector. What Domain Validation is saying is that there's something wrong with the upper 8 bits of the wide bus. Those are the last few wires on the 68 wire ribbon cable (wire one is usually marked in a different colour).

James