From: Juha Leppanen <juha_motorsportcom@luukku.com>
Date: Mon, 4 Jan 2010 20:52:50 +0000 (-0500)
Subject: wl1271: fix timeout in wl1271_top_reg_read
X-Git-Url: http://git.osdn.net/view?a=commitdiff_plain;h=51633632ef614ef045f25d76fc7f0133d7cc60c6;p=sagit-ice-cold%2Fkernel_xiaomi_msm8998.git

wl1271: fix timeout in wl1271_top_reg_read

I noticed a timeout bug in

/drivers/net/wireless/wl12xx/wl1271_spi.c

In the current code you cannot tell why you exited
the "poll for data ready" do-while loop if exiting
was done after the last possible loop.

Then timeout==0 regardless of (val & OCP_READY_MASK) or
!(val & OCP_READY_MASK), leading to possible false timeout...

Simple correction could be decreasing timeout after checking
for !(val & OCP_READY_MASK), not before

(Manually converted from email to an actual patch by me. -- JWL)

Reported-by: "Juha Leppanen" <juha_motorsportcom@luukku.com>
Signed-off-by: "Juha Leppanen" <juha_motorsportcom@luukku.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
---

diff --git a/drivers/net/wireless/wl12xx/wl1271_spi.c b/drivers/net/wireless/wl12xx/wl1271_spi.c
index 02978a16e732..ee9564aa6ecc 100644
--- a/drivers/net/wireless/wl12xx/wl1271_spi.c
+++ b/drivers/net/wireless/wl12xx/wl1271_spi.c
@@ -397,8 +397,7 @@ u16 wl1271_top_reg_read(struct wl1271 *wl, int addr)
 	/* poll for data ready */
 	do {
 		val = wl1271_spi_read32(wl, OCP_DATA_READ);
-		timeout--;
-	} while (!(val & OCP_READY_MASK) && timeout);
+	} while (!(val & OCP_READY_MASK) && --timeout);
 
 	if (!timeout) {
 		wl1271_warning("Top register access timed out.");