Skip to contentSkip to conversation

staging: vt6655: rewrite CARDvUpdateBasicTopRate()

PATCH In review Dan Carpenter 2 participants 4 messages Feb 13, 2010

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 staging: vt6655: rewrite CARDvUpdateBasicTopRate()

Originally, I wrote a patch to add some missing curly braces, but Walter Harms suggested that I should just rewrite the whole function to use the kernel version of ffs().

It is odd to subtract one from the result of ffs() but that was present in the original code as well (once you add the missing curly braces).

---
This patch replaces my previous patch. I don't have the hardware to actually test this change. :/

drivers/staging/vt6655/card.c +24 -20
index db78614..3a1f2be 100644
@@ -2780,28 +2780,32 @@ void vUpdateIFS (PVOID pDeviceHandler)
     VNSvOutPortB(pDevice->PortOffset + MAC_REG_CWMAXMIN0, (BYTE)byMaxMin);
 }
 
+static int top_OFDM_basic_rate(u16 basic_rate)
+{
+	int rate;
+
+	rate = ffs(basic_rate) - 1;
+	if (rate > RATE_54M || rate < RATE_6M)
+		return RATE_24M;
+	return rate;
+}
+
+static int top_CCK(u16 basic_rate)
+{
+	int rate;
+
+	rate = ffs(basic_rate) - 1;
+	if (rate > RATE_11M || rate < 0)
+		return RATE_1M;
+	return rate;
+}
+
 void CARDvUpdateBasicTopRate (PVOID pDeviceHandler)
 {
-    PSDevice pDevice = (PSDevice) pDeviceHandler;
-    BYTE byTopOFDM = RATE_24M, byTopCCK = RATE_1M;
-    BYTE ii;
-
-     //Determines the highest basic rate.
-     for (ii = RATE_54M; ii >= RATE_6M; ii --) {
-         if ( (pDevice->wBasicRate) & ((WORD)(1<<ii)) )
-             byTopOFDM = ii;
-             break;
-     }
-     pDevice->byTopOFDMBasicRate = byTopOFDM;
-
-     for (ii = RATE_11M;; ii --) {
-         if ( (pDevice->wBasicRate) & ((WORD)(1<<ii)) )
-             byTopCCK = ii;
-             break;
-         if (ii = RATE_1M)
-            break;
-     }
-     pDevice->byTopCCKBasicRate = byTopCCK;
+	PSDevice pDevice = (PSDevice) pDeviceHandler;
+
+	pDevice->byTopOFDMBasicRate = top_OFDM_basic_rate(pDevice->wBasicRate);
+	pDevice->byTopCCKBasicRate = top_CCK(pDevice->wBasicRate);
 }
 
 
#1 lore raw
Quoted 51 lines · Dan Carpenter

Originally, I wrote a patch to add some missing curly braces, but Walter Harms suggested that I should just rewrite the whole function to use the kernel version of ffs().

It is odd to subtract one from the result of ffs() but that was present in the original code as well (once you add the missing curly braces).

---
This patch replaces my previous patch. I don't have the hardware to actually test this change. :/

drivers/staging/vt6655/card.c +20 -9
index db78614..3a1f2be 100644
@@ -2780,28 +2780,32 @@ void vUpdateIFS (PVOID pDeviceHandler)
     VNSvOutPortB(pDevice->PortOffset + MAC_REG_CWMAXMIN0, (BYTE)byMaxMin);
 }
 
+static int top_OFDM_basic_rate(u16 basic_rate)
+{
+	int rate;
+
+	rate = ffs(basic_rate) - 1;
+	if (rate > RATE_54M || rate < RATE_6M)
+		return RATE_24M;
+	return rate;
+}
+
+static int top_CCK(u16 basic_rate)
+{
+	int rate;
+
+	rate = ffs(basic_rate) - 1;
+	if (rate > RATE_11M || rate < 0)
+		return RATE_1M;
+	return rate;
+}
+
 void CARDvUpdateBasicTopRate (PVOID pDeviceHandler)
 {
-    PSDevice pDevice = (PSDevice) pDeviceHandler;
-    BYTE byTopOFDM = RATE_24M, byTopCCK = RATE_1M;
-    BYTE ii;
-
-     //Determines the highest basic rate.
-     for (ii = RATE_54M; ii >= RATE_6M; ii --) {
-         if ( (pDevice->wBasicRate) & ((WORD)(1<<ii)) )
-             byTopOFDM = ii;
-             break;

drivers/staging/vt6655/card.c void vUpdateIFS (PVOID pDeviceHandler) · line 2803

Quoted 10 lines
drivers/staging/vt6655/card.c +0 -10
-     }
-     pDevice->byTopOFDMBasicRate = byTopOFDM;
-
-     for (ii = RATE_11M;; ii --) {
-         if ( (pDevice->wBasicRate) & ((WORD)(1<<ii)) )
-             byTopCCK = ii;
-             break;
-         if (ii = RATE_1M)
-            break;
-     }

Hmm, but this is different from ffs in that it starts from higher bits and go down and hence finds the highest possible rate (there may be more basic rates). Your patch selects the lowest possible one.

You cannot use ffs and fls here.

Signature · 1 line
js
#2 lore raw
In reply to Dan Carpenter
Quoted 1 line · Jiri Slaby

You cannot use ffs and fls here.

Actually you can (fls), if you add proper masks there like fls(pDevice->wBasicRate & ((1 << RATE_24M) - 1) & ~((1 << RATE_6M) - 1)), but with some macro-ification inside. Dunno if it gets better though.

Signature · 1 line
js
#3 lore raw
In reply to Dan Carpenter
Quoted 7 lines · Jiri Slaby
Quoted 1 line · Jiri Slaby

You cannot use ffs and fls here.

Actually you can (fls), if you add proper masks there like fls(pDevice->wBasicRate & ((1 << RATE_24M) - 1) & ~((1 << RATE_6M) - 1)), but with some macro-ification inside. Dunno if it gets better though.

Ok. I didn't understand how ffs() and fls() worked before. Sorry about that.

Let's just apply my original patch because that is an obvious one.

If I could test my changes, I would feel more comfortable rewriting stuff but I don't have the hardware.

regards,
dan carpenter

Quoted 2 lines
Signature · 1 line
js