From: Tom Coughlan <coughlan@redhat.com>
Date: Thu, 23 Apr 2009 16:26:52 -0400
Subject: [scsi] st: option to use SILI in variable block reads
Message-id: 1240518412.7392.240.camel@p670.boston.redhat.com
O-Subject: [RHEL 5.4 PATCH] st: add option to use SILI in variable block reads
Bugzilla: 457970
RH-Acked-by: Tomas Henzl <thenzl@redhat.com>
RH-Acked-by: Rob Evers <revers@redhat.com>
RH-Acked-by: Mike Christie <mchristi@redhat.com>

The SCSI Stream Commands Standard (SSC, for tapes) includes an option
called Suppress Incorrect-Length Indicator (SILI). From the upstream
commit: "Add new option MT_ST_SILI to enable setting the SILI bit in
reads in variable block mode. If SILI is set, reading a block shorter
than the byte count does not result in CHECK CONDITION. The length of
the block is determined using the residual count from the HBA. Avoiding
the REQUEST SENSE command for every block speeds up some real
applications considerably."

This is apparently most noticeable for customers migrating from Solaris
to RHEL.

The patch is upstream:

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=40f6b36c6243462fb95d0343237331c423494b03

The risk of regression is low, because the feature is off by default.
Just the same, I would appreciate ACKs from the CC'd individuals at a
minimum. This will also give us some incentive to do some tape testing
during 5.4 QE.

This addresses Bug 457970.

Finally a scuzzy patch explicitly labeled as silly... ;)

Tom

diff --git a/Documentation/scsi/st.txt b/Documentation/scsi/st.txt
index 20e30cf..00815c0 100644
--- a/Documentation/scsi/st.txt
+++ b/Documentation/scsi/st.txt
@@ -2,7 +2,7 @@ This file contains brief information about the SCSI tape driver.
 The driver is currently maintained by Kai Mäkisara (email
 Kai.Makisara@kolumbus.fi)
 
-Last modified: Mon Mar  7 21:14:44 2005 by kai.makisara
+Last modified: Thu Feb 21 21:54:16 2008 by kai.makisara
 
 
 BASICS
@@ -372,6 +372,11 @@ MTSETDRVBUFFER
 	     MT_ST_SYSV sets the SYSV sematics (mode)
 	     MT_ST_NOWAIT enables immediate mode (i.e., don't wait for
 	        the command to finish) for some commands (e.g., rewind)
+	     MT_ST_SILI enables setting the SILI bit in SCSI commands when
+		reading in variable block mode to enhance performance when
+		reading blocks shorter than the byte count; set this only
+		if you are sure that the drive supports SILI and the HBA
+		correctly returns transfer residuals
 	     MT_ST_DEBUGGING debugging (global; debugging must be
 		compiled into the driver)
 	MT_ST_SETBOOLEANS
diff --git a/drivers/scsi/st.c b/drivers/scsi/st.c
index cb40dee..08c2ee0 100644
--- a/drivers/scsi/st.c
+++ b/drivers/scsi/st.c
@@ -182,6 +182,7 @@ static int modes_defined;
 
 static struct st_buffer *new_tape_buffer(int, int, int);
 static int enlarge_buffer(struct st_buffer *, int, int);
+static void clear_buffer(struct st_buffer *);
 static void normalize_buffer(struct st_buffer *);
 static int append_to_buffer(const char __user *, struct st_buffer *, int);
 static int from_buffer(struct st_buffer *, char __user *, int);
@@ -441,6 +442,7 @@ static void st_sleep_done(void *data, char *sense, int result, int resid)
 
 	memcpy(SRpnt->sense, sense, SCSI_SENSE_BUFFERSIZE);
 	(STp->buffer)->cmdstat.midlevel_result = SRpnt->result = result;
+	(STp->buffer)->cmdstat.residual = resid;
 	DEB( STp->write_pending = 0; )
 
 	if (SRpnt->waiting)
@@ -1158,6 +1160,7 @@ static int st_open(struct inode *inode, struct file *filp)
 		goto err_out;
 	}
 
+	(STp->buffer)->cleared = 0;
 	(STp->buffer)->writing = 0;
 	(STp->buffer)->syscall_result = 0;
 
@@ -1431,8 +1434,14 @@ static int setup_buffering(struct scsi_tape *STp, const char __user *buf,
 		if (STp->block_size)
 			bufsize = STp->block_size > st_fixed_buffer_size ?
 				STp->block_size : st_fixed_buffer_size;
-		else
+		else {
 			bufsize = count;
+			/* Make sure that data from previous user is not leaked even if
+			   HBA does not return correct residual */
+			if (is_read && STp->sili && !STbp->cleared)
+				clear_buffer(STbp);
+		}
+
 		if (bufsize > STbp->buffer_size &&
 		    !enlarge_buffer(STbp, bufsize, STp->restr_dma)) {
 			printk(KERN_WARNING "%s: Can't allocate %d byte tape buffer.\n",
@@ -1782,6 +1791,8 @@ static long read_tape(struct scsi_tape *STp, long count,
 	memset(cmd, 0, MAX_COMMAND_SIZE);
 	cmd[0] = READ_6;
 	cmd[1] = (STp->block_size != 0);
+	if (!cmd[1] && STp->sili)
+		cmd[1] |= 2;
 	cmd[2] = blks >> 16;
 	cmd[3] = blks >> 8;
 	cmd[4] = blks;
@@ -1910,8 +1921,11 @@ static long read_tape(struct scsi_tape *STp, long count,
 
 	}
 	/* End of error handling */ 
-	else			/* Read successful */
+	else {			/* Read successful */
 		STbp->buffer_bytes = bytes;
+		if (STp->sili) /* In fixed block mode residual is always zero here */
+			STbp->buffer_bytes -= STp->buffer->cmdstat.residual;
+	}
 
 	if (STps->drv_block >= 0) {
 		if (STp->block_size == 0)
@@ -2089,7 +2103,8 @@ static void st_log_options(struct scsi_tape * STp, struct st_modedef * STm, char
 		       name, STm->defaults_for_writes, STp->omit_blklims, STp->can_partitions,
 		       STp->scsi2_logical);
 		printk(KERN_INFO
-		       "%s:    sysv: %d nowait: %d\n", name, STm->sysv, STp->immediate);
+		       "%s:    sysv: %d nowait: %d sili: %d\n", name, STm->sysv, STp->immediate,
+			STp->sili);
 		printk(KERN_INFO "%s:    debugging: %d\n",
 		       name, debugging);
 	}
@@ -2132,6 +2147,7 @@ static int st_set_options(struct scsi_tape *STp, long options)
 		STp->scsi2_logical = (options & MT_ST_SCSI2LOGICAL) != 0;
 		STp->immediate = (options & MT_ST_NOWAIT) != 0;
 		STm->sysv = (options & MT_ST_SYSV) != 0;
+		STp->sili = (options & MT_ST_SILI) != 0;
 		DEB( debugging = (options & MT_ST_DEBUGGING) != 0;
 		     st_log_options(STp, STm, name); )
 	} else if (code == MT_ST_SETBOOLEANS || code == MT_ST_CLEARBOOLEANS) {
@@ -2163,6 +2179,8 @@ static int st_set_options(struct scsi_tape *STp, long options)
 			STp->immediate = value;
 		if ((options & MT_ST_SYSV) != 0)
 			STm->sysv = value;
+		if ((options & MT_ST_SILI) != 0)
+			STp->sili = value;
                 DEB(
 		if ((options & MT_ST_DEBUGGING) != 0)
 			debugging = value;
@@ -3650,6 +3668,8 @@ static int enlarge_buffer(struct st_buffer * STbuffer, int new_size, int need_dm
 		STbuffer->frp_segs += 1;
 		got += b_size;
 		STbuffer->buffer_size = got;
+		if (STbuffer->cleared)
+			memset(page_address(STbuffer->frp[segs].page), 0, b_size);
 		segs++;
 	}
 	STbuffer->b_data = page_address(STbuffer->frp[0].page);
@@ -3658,6 +3678,17 @@ static int enlarge_buffer(struct st_buffer * STbuffer, int new_size, int need_dm
 }
 
 
+/* Make sure that no data from previous user is in the internal buffer */
+static void clear_buffer(struct st_buffer * st_bp)
+{
+	int i;
+
+	for (i=0; i < st_bp->frp_segs; i++)
+		memset(page_address(st_bp->frp[i].page), 0, st_bp->frp[i].length);
+	st_bp->cleared = 1;
+}
+
+
 /* Release the extra buffer */
 static void normalize_buffer(struct st_buffer * STbuffer)
 {
@@ -3984,6 +4015,7 @@ static int st_probe(struct device *dev)
 	tpnt->two_fm = ST_TWO_FM;
 	tpnt->fast_mteom = ST_FAST_MTEOM;
 	tpnt->scsi2_logical = ST_SCSI2LOGICAL;
+	tpnt->sili = ST_SILI;
 	tpnt->immediate = ST_NOWAIT;
 	tpnt->default_drvbuffer = 0xff;		/* No forced buffering */
 	tpnt->partition = 0;
diff --git a/drivers/scsi/st.h b/drivers/scsi/st.h
index 5857543..962e534 100644
--- a/drivers/scsi/st.h
+++ b/drivers/scsi/st.h
@@ -11,6 +11,7 @@ struct st_cmdstatus {
 	int midlevel_result;
 	struct scsi_sense_hdr sense_hdr;
 	int have_sense;
+	int residual;
 	u64 uremainder64;
 	u8 flags;
 	u8 remainder_valid;
@@ -33,6 +34,7 @@ struct st_request {
 struct st_buffer {
 	unsigned char dma;	/* DMA-able buffer */
 	unsigned char do_dio;   /* direct i/o set up? */
+	unsigned char cleared;  /* internal buffer cleared after open? */
 	int buffer_size;
 	int buffer_blocks;
 	int buffer_bytes;
@@ -120,6 +122,7 @@ struct scsi_tape {
 	unsigned char try_dio;			/* try direct i/o in general? */
 	unsigned char c_algo;			/* compression algorithm */
 	unsigned char pos_unknown;			/* after reset position unknown */
+	unsigned char sili;			/* use SILI when reading in variable b mode */
 	int tape_type;
 	int long_timeout;	/* timeout for commands known to take long time */
 
diff --git a/drivers/scsi/st_options.h b/drivers/scsi/st_options.h
index b6b5c9c..d2f9479 100644
--- a/drivers/scsi/st_options.h
+++ b/drivers/scsi/st_options.h
@@ -3,7 +3,7 @@
 
    Copyright 1995-2003 Kai Makisara.
 
-   Last modified: Mon Apr  7 22:49:18 2003 by makisara
+   Last modified: Thu Feb 21 21:47:07 2008 by kai.makisara
 */
 
 #ifndef _ST_OPTIONS_H
@@ -94,6 +94,10 @@
    The default is BSD semantics. */
 #define ST_SYSV 0
 
+/* If ST_SILI is non-zero, the SILI bit is set when reading in variable block
+   mode and the block size is determined using the residual returned by the HBA. */
+#define ST_SILI 0
+
 /* Time to wait for the drive to become ready if blocking open */
 #define ST_BLOCK_SECONDS     120
 
diff --git a/include/linux/mtio.h b/include/linux/mtio.h
index 8c66151..ae1cd6c 100644
--- a/include/linux/mtio.h
+++ b/include/linux/mtio.h
@@ -338,6 +338,7 @@ struct mtftcmd {
 #define MT_ST_SCSI2LOGICAL      0x800
 #define MT_ST_SYSV              0x1000
 #define MT_ST_NOWAIT            0x2000
+#define MT_ST_SILI		0x4000
 
 /* The mode parameters to be controlled. Parameter chosen with bits 20-28 */
 #define MT_ST_CLEAR_DEFAULT	0xfffff