Skip to contentSkip to conversation

clk: k210: Fix checking if ulongs are less than 0

PATCH Superseded Sean Anderson 3 participants 7 messages Jul 27, 2021 – Jul 31, 2021 Common clk

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/4 clk: k210: Fix checking if ulongs are less than 0

Some clock functions return ulong but still have "negative" errors. To deal with this, cast the relevant arguments to long.

Fixes 609bd60b94 ("clk: k210: Rewrite to remove CCF")

---

 drivers/clk/clk_kendryte.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/clk/clk_kendryte.c +3 -3
index 3148756968..37bd624eca 100644
@@ -439,7 +439,7 @@ static const struct k210_clk_params k210_clks[] = {
 #ifdef CONFIG_CLK_K210_SET_RATE
 static int k210_pll_enable(struct k210_clk_priv *priv, int id);
 static int k210_pll_disable(struct k210_clk_priv *priv, int id);
-static ulong k210_pll_get_rate(struct k210_clk_priv *priv, int id, ulong rate_in);
+static ulong k210_pll_get_rate(struct k210_clk_priv *priv, int id, long rate_in);
 
 /*
  * The PLL included with the Kendryte K210 appears to be a True Circuits, Inc.
@@ -841,7 +841,7 @@ TEST_STATIC int k210_pll_calc_config(u32 rate, u32 rate_in,
 }
 
 static ulong k210_pll_set_rate(struct k210_clk_priv *priv, int id, ulong rate,
-			       ulong rate_in)
+			       long rate_in)
 {
 	int err;
 	const struct k210_pll_params *pll = &k210_plls[id];
@@ -890,7 +890,7 @@ static ulong k210_pll_set_rate(struct k210_clk_priv *priv, int id, ulong rate,
 #endif /* CONFIG_CLK_K210_SET_RATE */
 
 static ulong k210_pll_get_rate(struct k210_clk_priv *priv, int id,
-			       ulong rate_in)
+			       long rate_in)
 {
 	u64 r, f, od;
 	u32 reg = readl(priv->base + k210_plls[id].off);
Signature · 1 line
2.32.0
#1 lore raw

PATCH 2/4 k210: clk: Refactor out_of_spec tests

Everything here sits in a while (true) loop. However, this introduces a couple of layers of indentation. We can simplify the code by introducing a single goto instead of using continue/break. This will also make adding loops in the next patch easier.

---

 drivers/clk/clk_kendryte.c | 105 ++++++++++++++++++-------------------
 1 file changed, 52 insertions(+), 53 deletions(-)
drivers/clk/clk_kendryte.c +52 -53
index 37bd624eca..d2cee2cf97 100644
@@ -709,6 +709,10 @@ TEST_STATIC int k210_pll_calc_config(u32 rate, u32 rate_in,
 		 * Whether we swapped r and od while enforcing frequency limits
 		 */
 		bool swapped = false;
+		/*
+		 * Whether the intermediate frequencies are out-of-spec
+		 */
+		bool out_of_spec;
 		u64 last_od = od;
 		u64 last_r = r;
 
@@ -767,76 +771,71 @@ TEST_STATIC int k210_pll_calc_config(u32 rate, u32 rate_in,
 		 * aren't in spec, try swapping r and od. If everything is
 		 * in-spec, calculate the relative error.
 		 */
-		while (true) {
+again:
+		out_of_spec = false;
+		if (r > max_r) {
+			out_of_spec = true;
+		} else {
 			/*
-			 * Whether the intermediate frequencies are out-of-spec
+			 * There is no way to only divide once; we need
+			 * to examine the frequency with and without the
+			 * effect of od.
 			 */
-			bool out_of_spec = false;
+			u64 vco = DIV_ROUND_CLOSEST_ULL(rate_in * f, r);
 
-			if (r > max_r) {
+			if (vco > 1750000000 || vco < 340000000)
 				out_of_spec = true;
-			} else {
-				/*
-				 * There is no way to only divide once; we need
-				 * to examine the frequency with and without the
-				 * effect of od.
-				 */
-				u64 vco = DIV_ROUND_CLOSEST_ULL(rate_in * f, r);
+		}
 
-				if (vco > 1750000000 || vco < 340000000)
-					out_of_spec = true;
+		if (out_of_spec) {
+			u64 new_r, new_od;
+
+			if (!swapped) {
+				u64 tmp = r;
+
+				r = od;
+				od = tmp;
+				swapped = true;
+				goto again;
 			}
 
-			if (out_of_spec) {
-				if (!swapped) {
-					u64 tmp = r;
-
-					r = od;
-					od = tmp;
-					swapped = true;
-					continue;
-				} else {
-					/*
-					 * Try looking ahead to see if there are
-					 * additional factors for the same
-					 * product.
-					 */
-					if (i + 1 < ARRAY_SIZE(factors)) {
-						u64 new_r, new_od;
-
-						i++;
-						new_r = UNPACK_R(factors[i]);
-						new_od = UNPACK_OD(factors[i]);
-						if (r * od == new_r * new_od) {
-							r = new_r;
-							od = new_od;
-							swapped = false;
-							continue;
-						}
-						i--;
-					}
-					break;
+			/*
+			 * Try looking ahead to see if there are additional
+			 * factors for the same product.
+			 */
+			if (i + 1 < ARRAY_SIZE(factors)) {
+				i++;
+				new_r = UNPACK_R(factors[i]);
+				new_od = UNPACK_OD(factors[i]);
+				if (r * od == new_r * new_od) {
+					r = new_r;
+					od = new_od;
+					swapped = false;
+					goto again;
 				}
+				i--;
 			}
 
-			error = DIV_ROUND_CLOSEST_ULL(f * inv_ratio, r * od);
-			/* The lower 16 bits are spurious */
-			error = abs((error - BIT(32))) >> 16;
+			/* We ran out of things to try */
+			continue;
+		}
 
-			if (error < best_error) {
-				best->r = r;
-				best->f = f;
-				best->od = od;
-				best_error = error;
-			}
-			break;
+		error = DIV_ROUND_CLOSEST_ULL(f * inv_ratio, r * od);
+		/* The lower 16 bits are spurious */
+		error = abs((error - BIT(32))) >> 16;
+
+		if (error < best_error) {
+			best->r = r;
+			best->f = f;
+			best->od = od;
+			best_error = error;
 		}
 	} while (f < 64 && i + 1 < ARRAY_SIZE(factors) && error != 0);
 
+	log_debug("best error %lld\n", best_error);
 	if (best_error == S64_MAX)
 		return -EINVAL;
 
-	log_debug("best error %lld\n", best_error);
 	return 0;
 }
 
Signature · 1 line
2.32.0
#2 lore raw

PATCH 3/4 test: dm: k210: Reduce duplication in test cases

In reply to Sean Anderson

Having to copy-paste the same 3 lines makes adding new test cases error-prone. Use a macro.

---

 test/dm/k210_pll.c | 30 ++++++++++++------------------
 1 file changed, 12 insertions(+), 18 deletions(-)
test/dm/k210_pll.c +12 -18
index 54764f269c..5574ac96fa 100644
@@ -69,27 +69,21 @@ static int dm_test_k210_pll(struct unit_test_state *uts)
 						  &theirs));
 	ut_asserteq(-EINVAL, k210_pll_calc_config(1500000000, 20000000,
 						  &theirs));
+	ut_asserteq(-EINVAL, k210_pll_calc_config(1750000000, 13300000,
+						  &theirs));
 
 	/* Verify we get the same output with brute-force */
-	ut_assertok(dm_test_k210_pll_calc_config(390000000, 26000000, &ours));
-	ut_assertok(k210_pll_calc_config(390000000, 26000000, &theirs));
-	ut_assertok(dm_test_k210_pll_compare(&ours, &theirs));
+#define compare(rate, rate_in) do { \
+	ut_assertok(dm_test_k210_pll_calc_config(rate, rate_in, &ours)); \
+	ut_assertok(k210_pll_calc_config(rate, rate_in, &theirs)); \
+	ut_assertok(dm_test_k210_pll_compare(&ours, &theirs)); \
+} while (0)
 
-	ut_assertok(dm_test_k210_pll_calc_config(26000000, 390000000, &ours));
-	ut_assertok(k210_pll_calc_config(26000000, 390000000, &theirs));
-	ut_assertok(dm_test_k210_pll_compare(&ours, &theirs));
-
-	ut_assertok(dm_test_k210_pll_calc_config(400000000, 26000000, &ours));
-	ut_assertok(k210_pll_calc_config(400000000, 26000000, &theirs));
-	ut_assertok(dm_test_k210_pll_compare(&ours, &theirs));
-
-	ut_assertok(dm_test_k210_pll_calc_config(27000000, 26000000, &ours));
-	ut_assertok(k210_pll_calc_config(27000000, 26000000, &theirs));
-	ut_assertok(dm_test_k210_pll_compare(&ours, &theirs));
-
-	ut_assertok(dm_test_k210_pll_calc_config(26000000, 27000000, &ours));
-	ut_assertok(k210_pll_calc_config(26000000, 27000000, &theirs));
-	ut_assertok(dm_test_k210_pll_compare(&ours, &theirs));
+	compare(390000000, 26000000);
+	compare(26000000, 390000000);
+	compare(400000000, 26000000);
+	compare(27000000, 26000000);
+	compare(26000000, 27000000);
 
 	return 0;
 }
Signature · 1 line
2.32.0
Reviewed-by on v1 3/4
#3 lore raw

Re: [PATCH 3/4] test: dm: k210: Reduce duplication in test cases

Quoted 9 lines · Sean Anderson

Having to copy-paste the same 3 lines makes adding new test cases error-prone. Use a macro.

---

 test/dm/k210_pll.c | 30 ++++++++++++------------------
 1 file changed, 12 insertions(+), 18 deletions(-)
#4 lore raw

PATCH 4/4 clk: k210: Try harder to get the best config

In reply to Sean Anderson

In some cases, the best config cannot be used because the VCO would be out-of-spec. In these cases, we may need to try a worse combination of r/od in order to find the best representable config. This also adds a few test cases to catch this and other (possible) unlikely errors.

---

 drivers/clk/clk_kendryte.c | 24 ++++++++++++++++++++++++
 test/dm/k210_pll.c         |  4 ++++
 2 files changed, 28 insertions(+)
drivers/clk/clk_kendryte.c +24 -0
index d2cee2cf97..e90550ef33 100644
@@ -816,6 +816,30 @@ again:
 				i--;
 			}
 
+			/*
+			 * Try looking back to see if there is a worse ratio
+			 * that we could try anyway
+			 */
+			while (i > 0) {
+				i--;
+				new_r = UNPACK_R(factors[i]);
+				new_od = UNPACK_OD(factors[i]);
+				/*
+				 * Don't loop over factors for the same product
+				 * to avoid getting stuck because of the above
+				 * clause
+				 */
+				if (r * od != new_r * new_od) {
+					if (new_r * new_od > last_r * last_od) {
+						r = new_r;
+						od = new_od;
+						swapped = false;
+						goto again;
+					}
+					break;
+				}
+			}
+
 			/* We ran out of things to try */
 			continue;
 		}
test/dm/k210_pll.c +4 -0
index 5574ac96fa..f55379f336 100644
@@ -84,6 +84,10 @@ static int dm_test_k210_pll(struct unit_test_state *uts)
 	compare(400000000, 26000000);
 	compare(27000000, 26000000);
 	compare(26000000, 27000000);
+	compare(13300000 * 64, 13300000);
+	compare(21250000, 21250000 * 70);
+	compare(21250000, 1750000000);
+	compare(1750000000, 1750000000);
 
 	return 0;
 }
Signature · 1 line
2.32.0
#5 lore raw
In reply to Sean Anderson
Quoted 25 lines · Sean Anderson

Some clock functions return ulong but still have "negative" errors. To deal with this, cast the relevant arguments to long.

Fixes 609bd60b94 ("clk: k210: Rewrite to remove CCF")

---

 drivers/clk/clk_kendryte.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/clk/clk_kendryte.c +1 -1
index 3148756968..37bd624eca 100644
@@ -439,7 +439,7 @@ static const struct k210_clk_params k210_clks[] = {
 #ifdef CONFIG_CLK_K210_SET_RATE
 static int k210_pll_enable(struct k210_clk_priv *priv, int id);
 static int k210_pll_disable(struct k210_clk_priv *priv, int id);
-static ulong k210_pll_get_rate(struct k210_clk_priv *priv, int id, ulong rate_in);
+static ulong k210_pll_get_rate(struct k210_clk_priv *priv, int id, long rate_in);
 
 /*
  * The PLL included with the Kendryte K210 appears to be a True Circuits, Inc.

drivers/clk/clk_kendryte.c TEST_STATIC int k210_pll_calc_config(u32 rate, u32 rate_in, · line 843

Quoted 3 lines
drivers/clk/clk_kendryte.c +0 -0
@@ -841,7 +841,7 @@ TEST_STATIC int k210_pll_calc_config(u32 rate, u32 rate_in,
 }
 
 static ulong k210_pll_set_rate(struct k210_clk_priv *priv, int id, ulong rate,

Shouldn't this one return a long, in case of error ? It seems that the commit messages hints at such a change, but you are changing the argument type instead. A little confusing. What am I missing ?

drivers/clk/clk_kendryte.c static ulong k210_pll_set_rate(struct k210_clk_priv *priv, int id, ulong rate, · line 892

Quoted 9 lines
drivers/clk/clk_kendryte.c +1 -1
-			       ulong rate_in)
+			       long rate_in)
 {
 	int err;
 	const struct k210_pll_params *pll = &k210_plls[id];
@@ -890,7 +890,7 @@ static ulong k210_pll_set_rate(struct k210_clk_priv *priv, int id, ulong rate,
 #endif /* CONFIG_CLK_K210_SET_RATE */
 
 static ulong k210_pll_get_rate(struct k210_clk_priv *priv, int id,

Same here ?

drivers/clk/clk_kendryte.c static ulong k210_pll_set_rate(struct k210_clk_priv *priv, int id, ulong rate, · line 893

Quoted 2 lines
drivers/clk/clk_kendryte.c +1 -1
-			       ulong rate_in)
+			       long rate_in)

I would assume that these functions are called if the rate_in argument is correct, so I do not really understand why the argument type needs to be changed...

Quoted 4 lines
drivers/clk/clk_kendryte.c +0 -0
 {
 	u64 r, f, od;
 	u32 reg = readl(priv->base + k210_plls[id].off);
Signature · 2 lines
Damien Le Moal
Western Digital Research
#6 lore raw
Quoted 28 lines · Damien Le Moal
Quoted 27 lines · Sean Anderson

Some clock functions return ulong but still have "negative" errors. To deal with this, cast the relevant arguments to long.

Fixes 609bd60b94 ("clk: k210: Rewrite to remove CCF")

---

  drivers/clk/clk_kendryte.c | 6 +++---
  1 file changed, 3 insertions(+), 3 deletions(-)
drivers/clk/clk_kendryte.c +1 -1
index 3148756968..37bd624eca 100644
@@ -439,7 +439,7 @@ static const struct k210_clk_params k210_clks[] = {
  #ifdef CONFIG_CLK_K210_SET_RATE
  static int k210_pll_enable(struct k210_clk_priv *priv, int id);
  static int k210_pll_disable(struct k210_clk_priv *priv, int id);
-static ulong k210_pll_get_rate(struct k210_clk_priv *priv, int id, ulong rate_in);
+static ulong k210_pll_get_rate(struct k210_clk_priv *priv, int id, long rate_in);
 
  /*
   * The PLL included with the Kendryte K210 appears to be a True Circuits, Inc.
@@ -841,7 +841,7 @@ TEST_STATIC int k210_pll_calc_config(u32 rate, u32 rate_in,
  }
 
Quoted 5 lines

drivers/clk/clk_kendryte.c TEST_STATIC int k210_pll_calc_config(u32 rate, u32 rate_in, · line 843

Quoted 1 line
drivers/clk/clk_kendryte.c +0 -0
  static ulong k210_pll_set_rate(struct k210_clk_priv *priv, int id, ulong rate,

Shouldn't this one return a long, in case of error ? It seems that the commit messages hints at such a change, but you are changing the argument type instead. A little confusing. What am I missing ?

Perhaps they should return long, but these are basically matching the prototypes in include/clk-uclass.h. And there, get_rate and set_rate take ulong arguments and return ulongs.

Quoted 13 lines

drivers/clk/clk_kendryte.c static ulong k210_pll_set_rate(struct k210_clk_priv *priv, int id, ulong rate, · line 892

Quoted 10 lines
drivers/clk/clk_kendryte.c +1 -1
-			       ulong rate_in)
+			       long rate_in)
  {
  	int err;
  	const struct k210_pll_params *pll = &k210_plls[id];
@@ -890,7 +890,7 @@ static ulong k210_pll_set_rate(struct k210_clk_priv *priv, int id, ulong rate,
  #endif /* CONFIG_CLK_K210_SET_RATE */
 
  static ulong k210_pll_get_rate(struct k210_clk_priv *priv, int id,

Same here ?

Quoted 5 lines

drivers/clk/clk_kendryte.c static ulong k210_pll_set_rate(struct k210_clk_priv *priv, int id, ulong rate, · line 893

Quoted 2 lines
drivers/clk/clk_kendryte.c +1 -1
-			       ulong rate_in)
+			       long rate_in)

I would assume that these functions are called if the rate_in argument is correct, so I do not really understand why the argument type needs to be changed...

Hm, I suppose the better patch would be to check the return of get_rate and set_rate when we call them. I think my intent here was to allow subsequent functions to be no-ops in case of error, but it looks like I act on these values directly. Will fix in v2.

--Sean

Quoted 7 lines
Quoted 5 lines
drivers/clk/clk_kendryte.c +0 -0
  {
  	u64 r, f, od;
  	u32 reg = readl(priv->base + k210_plls[id].off);