rslib: Split rs control struct

The decoder library uses variable length arrays on stack. To get rid of
them it would be simple to allocate fixed length arrays on stack, but those
might become rather large. The other solution is to allocate the buffers in
the rs control structure, but this cannot be done as long as the structure
can be shared by several users. Sharing is desired because the RS polynom
tables are large and initialization is time consuming.

To solve this split the codec information out of the control structure and
have a pointer to a shared codec in it. Instantiate the control structure
for each user, create a new codec if no shareable is avaiable yet.  Adjust
all affected usage sites to the new scheme.

This allows to add per instance decoder buffers to the control structure
later on.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Segher Boessenkool <segher@kernel.crashing.org>
Cc: Kernel Hardening <kernel-hardening@lists.openwall.com>
Cc: Richard Weinberger <richard@nod.at>
Cc: Mike Snitzer <snitzer@redhat.com>
Cc: Anton Vorontsov <anton@enomsg.org>
Cc: Colin Cross <ccross@android.com>
Cc: Andrew Morton <akpm@linuxfoundation.org>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Alasdair Kergon <agk@redhat.com>
Signed-off-by: Kees Cook <keescook@chromium.org>

3 files changed, 79 insertions, 58 deletions

diff --git a/lib/reed_solomon/decode_rs.c b/lib/reed_solomon/decode_rs.c
index d61007ade4a08..794cced31c75f 100644
--- a/lib/reed_solomon/decode_rs.c
+++ b/lib/reed_solomon/decode_rs.c
@@ -10,6 +10,7 @@
  * Generic data width independent code which is included by the wrappers.
  */
 {
+	struct rs_codec *rs = rsc->codec;
 	int deg_lambda, el, deg_omega;
 	int i, j, r, k, pad;
 	int nn = rs->nn;
diff --git a/lib/reed_solomon/encode_rs.c b/lib/reed_solomon/encode_rs.c
index 10ca1ebb13dd0..9112d46e869ee 100644
--- a/lib/reed_solomon/encode_rs.c
+++ b/lib/reed_solomon/encode_rs.c
@@ -10,6 +10,7 @@
  * Generic data width independent code which is included by the wrappers.
  */
 {
+	struct rs_codec *rs = rsc->codec;
 	int i, j, pad;
 	int nn = rs->nn;
 	int nroots = rs->nroots;
diff --git a/lib/reed_solomon/reed_solomon.c b/lib/reed_solomon/reed_solomon.c
index 02c19ecffc280..cb21e8b5a4e0e 100644
--- a/lib/reed_solomon/reed_solomon.c
+++ b/lib/reed_solomon/reed_solomon.c
@@ -11,22 +11,23 @@
  *
  * The generic Reed Solomon library provides runtime configurable
  * encoding / decoding of RS codes.
- * Each user must call init_rs to get a pointer to a rs_control
- * structure for the given rs parameters. This structure is either
- * generated or a already available matching control structure is used.
- * If a structure is generated then the polynomial arrays for
- * fast encoding / decoding are built. This can take some time so
- * make sure not to call this function from a time critical path.
- * Usually a module / driver should initialize the necessary
- * rs_control structure on module / driver init and release it
- * on exit.
- * The encoding puts the calculated syndrome into a given syndrome
- * buffer.
- * The decoding is a two step process. The first step calculates
- * the syndrome over the received (data + syndrome) and calls the
- * second stage, which does the decoding / error correction itself.
- * Many hw encoders provide a syndrome calculation over the received
- * data + syndrome and can call the second stage directly.
+ *
+ * Each user must call init_rs to get a pointer to a rs_control structure
+ * for the given rs parameters. The control struct is unique per instance.
+ * It points to a codec which can be shared by multiple control structures.
+ * If a codec is newly allocated then the polynomial arrays for fast
+ * encoding / decoding are built. This can take some time so make sure not
+ * to call this function from a time critical path.  Usually a module /
+ * driver should initialize the necessary rs_control structure on module /
+ * driver init and release it on exit.
+ *
+ * The encoding puts the calculated syndrome into a given syndrome buffer.
+ *
+ * The decoding is a two step process. The first step calculates the
+ * syndrome over the received (data + syndrome) and calls the second stage,
+ * which does the decoding / error correction itself.  Many hw encoders
+ * provide a syndrome calculation over the received data + syndrome and can
+ * call the second stage directly.
  */
 #include <linux/errno.h>
 #include <linux/kernel.h>
@@ -36,13 +37,13 @@
 #include <linux/slab.h>
 #include <linux/mutex.h>
 
-/* This list holds all currently allocated rs control structures */
-static LIST_HEAD (rslist);
+/* This list holds all currently allocated rs codec structures */
+static LIST_HEAD(codec_list);
 /* Protection for the list */
 static DEFINE_MUTEX(rslistlock);
 
 /**
- * rs_init - Initialize a Reed-Solomon codec
+ * codec_init - Initialize a Reed-Solomon codec
  * @symsize:	symbol size, bits (1-8)
  * @gfpoly:	Field generator polynomial coefficients
  * @gffunc:	Field generator function
@@ -51,14 +52,14 @@ static DEFINE_MUTEX(rslistlock);
  * @nroots:	RS code generator polynomial degree (number of roots)
  * @gfp:	GFP_ flags for allocations
  *
- * Allocate a control structure and the polynom arrays for faster
+ * Allocate a codec structure and the polynom arrays for faster
  * en/decoding. Fill the arrays according to the given parameters.
  */
-static struct rs_control *rs_init(int symsize, int gfpoly, int (*gffunc)(int),
-				  int fcr, int prim, int nroots, gfp_t gfp)
+static struct rs_codec *codec_init(int symsize, int gfpoly, int (*gffunc)(int),
+				   int fcr, int prim, int nroots, gfp_t gfp)
 {
-	struct rs_control *rs;
 	int i, j, sr, root, iprim;
+	struct rs_codec *rs;
 
 	rs = kzalloc(sizeof(*rs), gfp);
 	if (!rs)
@@ -138,6 +139,9 @@ static struct rs_control *rs_init(int symsize, int gfpoly, int (*gffunc)(int),
 	/* convert rs->genpoly[] to index form for quicker encoding */
 	for (i = 0; i <= nroots; i++)
 		rs->genpoly[i] = rs->index_of[rs->genpoly[i]];
+
+	rs->users = 1;
+	list_add(&rs->list, &codec_list);
 	return rs;
 
 err:
@@ -150,27 +154,37 @@ err:
 
 
 /**
- *  free_rs - Free the rs control structure, if it is no longer used
- *  @rs:	the control structure which is not longer used by the
+ *  free_rs - Free the rs control structure
+ *  @rs:	The control structure which is not longer used by the
  *		caller
+ *
+ * Free the control structure. If @rs is the last user of the associated
+ * codec, free the codec as well.
  */
 void free_rs(struct rs_control *rs)
 {
+	struct rs_codec *cd;
+
+	if (!rs)
+		return;
+
+	cd = rs->codec;
 	mutex_lock(&rslistlock);
-	rs->users--;
-	if(!rs->users) {
-		list_del(&rs->list);
-		kfree(rs->alpha_to);
-		kfree(rs->index_of);
-		kfree(rs->genpoly);
-		kfree(rs);
+	cd->users--;
+	if(!cd->users) {
+		list_del(&cd->list);
+		kfree(cd->alpha_to);
+		kfree(cd->index_of);
+		kfree(cd->genpoly);
+		kfree(cd);
 	}
 	mutex_unlock(&rslistlock);
+	kfree(rs);
 }
 EXPORT_SYMBOL_GPL(free_rs);
 
 /**
- * init_rs_internal - Find a matching or allocate a new rs control structure
+ * init_rs_internal - Allocate rs control, find a matching codec or allocate a new one
  *  @symsize:	the symbol size (number of bits)
  *  @gfpoly:	the extended Galois field generator polynomial coefficients,
  *		with the 0th coefficient in the low order bit. The polynomial
@@ -201,33 +215,39 @@ static struct rs_control *init_rs_internal(int symsize, int gfpoly,
 	if (nroots < 0 || nroots >= (1<<symsize))
 		return NULL;
 
+	rs = kzalloc(sizeof(*rs), GFP_KERNEL);
+	if (!rs)
+		return NULL;
+
 	mutex_lock(&rslistlock);
 
 	/* Walk through the list and look for a matching entry */
-	list_for_each(tmp, &rslist) {
-		rs = list_entry(tmp, struct rs_control, list);
-		if (symsize != rs->mm)
+	list_for_each(tmp, &codec_list) {
+		struct rs_codec *cd = list_entry(tmp, struct rs_codec, list);
+
+		if (symsize != cd->mm)
 			continue;
-		if (gfpoly != rs->gfpoly)
+		if (gfpoly != cd->gfpoly)
 			continue;
-		if (gffunc != rs->gffunc)
+		if (gffunc != cd->gffunc)
 			continue;
-		if (fcr != rs->fcr)
+		if (fcr != cd->fcr)
 			continue;
-		if (prim != rs->prim)
+		if (prim != cd->prim)
 			continue;
-		if (nroots != rs->nroots)
+		if (nroots != cd->nroots)
 			continue;
 		/* We have a matching one already */
-		rs->users++;
+		cd->users++;
+		rs->codec = cd;
 		goto out;
 	}
 
 	/* Create a new one */
-	rs = rs_init(symsize, gfpoly, gffunc, fcr, prim, nroots, gfp);
-	if (rs) {
-		rs->users = 1;
-		list_add(&rs->list, &rslist);
+	rs->codec = codec_init(symsize, gfpoly, gffunc, fcr, prim, nroots, gfp);
+	if (!rs->codec) {
+		kfree(rs);
+		rs = NULL;
 	}
 out:
 	mutex_unlock(&rslistlock);
@@ -235,7 +255,7 @@ out:
 }
 
 /**
- * init_rs_gfp - Find a matching or allocate a new rs control structure
+ * init_rs_gfp - Create a RS control struct and initialize it
  *  @symsize:	the symbol size (number of bits)
  *  @gfpoly:	the extended Galois field generator polynomial coefficients,
  *		with the 0th coefficient in the low order bit. The polynomial
@@ -254,9 +274,8 @@ struct rs_control *init_rs_gfp(int symsize, int gfpoly, int fcr, int prim,
 EXPORT_SYMBOL_GPL(init_rs_gfp);
 
 /**
- * init_rs_non_canonical - Find a matching or allocate a new rs control
- *                         structure, for fields with non-canonical
- *                         representation
+ * init_rs_non_canonical - Allocate rs control struct for fields with
+ *                         non-canonical representation
  *  @symsize:	the symbol size (number of bits)
  *  @gffunc:	pointer to function to generate the next field element,
  *		or the multiplicative identity element if given 0.  Used
@@ -277,7 +296,7 @@ EXPORT_SYMBOL_GPL(init_rs_non_canonical);
 #ifdef CONFIG_REED_SOLOMON_ENC8
 /**
  *  encode_rs8 - Calculate the parity for data values (8bit data width)
- *  @rs:	the rs control structure
+ *  @rsc:	the rs control structure
  *  @data:	data field of a given type
  *  @len:	data length
  *  @par:	parity data, must be initialized by caller (usually all 0)
@@ -287,7 +306,7 @@ EXPORT_SYMBOL_GPL(init_rs_non_canonical);
  *  symbol size > 8. The calling code must take care of encoding of the
  *  syndrome result for storage itself.
  */
-int encode_rs8(struct rs_control *rs, uint8_t *data, int len, uint16_t *par,
+int encode_rs8(struct rs_control *rsc, uint8_t *data, int len, uint16_t *par,
 	       uint16_t invmsk)
 {
 #include "encode_rs.c"
@@ -298,7 +317,7 @@ EXPORT_SYMBOL_GPL(encode_rs8);
 #ifdef CONFIG_REED_SOLOMON_DEC8
 /**
  *  decode_rs8 - Decode codeword (8bit data width)
- *  @rs:	the rs control structure
+ *  @rsc:	the rs control structure
  *  @data:	data field of a given type
  *  @par:	received parity data field
  *  @len:	data length
@@ -313,7 +332,7 @@ EXPORT_SYMBOL_GPL(encode_rs8);
  *  syndrome result and the received parity before calling this code.
  *  Returns the number of corrected bits or -EBADMSG for uncorrectable errors.
  */
-int decode_rs8(struct rs_control *rs, uint8_t *data, uint16_t *par, int len,
+int decode_rs8(struct rs_control *rsc, uint8_t *data, uint16_t *par, int len,
 	       uint16_t *s, int no_eras, int *eras_pos, uint16_t invmsk,
 	       uint16_t *corr)
 {
@@ -325,7 +344,7 @@ EXPORT_SYMBOL_GPL(decode_rs8);
 #ifdef CONFIG_REED_SOLOMON_ENC16
 /**
  *  encode_rs16 - Calculate the parity for data values (16bit data width)
- *  @rs:	the rs control structure
+ *  @rsc:	the rs control structure
  *  @data:	data field of a given type
  *  @len:	data length
  *  @par:	parity data, must be initialized by caller (usually all 0)
@@ -333,7 +352,7 @@ EXPORT_SYMBOL_GPL(decode_rs8);
  *
  *  Each field in the data array contains up to symbol size bits of valid data.
  */
-int encode_rs16(struct rs_control *rs, uint16_t *data, int len, uint16_t *par,
+int encode_rs16(struct rs_control *rsc, uint16_t *data, int len, uint16_t *par,
 	uint16_t invmsk)
 {
 #include "encode_rs.c"
@@ -344,7 +363,7 @@ EXPORT_SYMBOL_GPL(encode_rs16);
 #ifdef CONFIG_REED_SOLOMON_DEC16
 /**
  *  decode_rs16 - Decode codeword (16bit data width)
- *  @rs:	the rs control structure
+ *  @rsc:	the rs control structure
  *  @data:	data field of a given type
  *  @par:	received parity data field
  *  @len:	data length
@@ -357,7 +376,7 @@ EXPORT_SYMBOL_GPL(encode_rs16);
  *  Each field in the data array contains up to symbol size bits of valid data.
  *  Returns the number of corrected bits or -EBADMSG for uncorrectable errors.
  */
-int decode_rs16(struct rs_control *rs, uint16_t *data, uint16_t *par, int len,
+int decode_rs16(struct rs_control *rsc, uint16_t *data, uint16_t *par, int len,
 		uint16_t *s, int no_eras, int *eras_pos, uint16_t invmsk,
 		uint16_t *corr)
 {