Skip to contentSkip to conversation

staging: rtl8723bs: remove dead 5GHz fallback in GetDeltaSwingTable_8723B

PATCH In review Hank Yang 2 participants 3 messages Sep 6 Staging

lore
Share & tools
lore

Hank Yang

author
#0 lore raw

PATCH 1/1 staging: rtl8723bs: remove dead 5GHz fallback in GetDeltaSwingTable_8723B

GetDeltaSwingTable_8723B() falls back to calibration tables from a different chip (8188E) when channel is outside the range of 1 to 14:

  } else {
          *TemperatureUP_A   = (u8 *)DeltaSwingTableIdx_2GA_P_8188E;
          ...
  }

This branch is unreachable. channel comes from CurrentChannel, which is only updated in rtl8723b_phycfg.c after passing HAL_IsLegalChannel(), which enforces a range of 1 to 14. Since RTL8723BS is a 2.4GHz-only chip [1][2], with no 5GHz support, CurrentChannel can never exceed the range of 1 to 14, so this fallback branch can never execute.

See discussion at [3] for the full reasoning.

Also remove the two DeltaSwingTableIdx_*_8188E arrays that were only used by this branch, along with a comment referencing a TxPowerTrack.txt config file — no such file-reading logic exists anywhere in this driver, so the comment is stale as well.

[1] https://www.realtek.com/Product/Index?id=610&cate_id=194
[2] https://fccid.io/TX2-RTL8723BS/User-Manual/User-Manual-System-pdf-2365682
[3] https://lore.kernel.org/linux-staging/2026090527-outdoors-press-9602@gregkh/T/#t

Assisted-by Claude:claude-sonnet-5
---
 .../staging/rtl8723bs/hal/HalPhyRf_8723B.c    | 19 -------------------
 1 file changed, 19 deletions(-)
drivers/staging/rtl8723bs/hal/HalPhyRf_8723B.c +0 -19
index 6c5f56d5a1f4..394ee4e732a6 100644
@@ -21,20 +21,6 @@
 #define		PATH_S1							0 /*  RF_PATH_A */
 #define		IDX_0xC4C						2
 
-/*---------------------------Define Local Constant---------------------------*/
-
-/* In the case that we fail to read TxPowerTrack.txt, we use the table for
- * 88E as the default table.
- */
-static u8 DeltaSwingTableIdx_2GA_N_8188E[] = {
-	0, 0, 0, 2, 2, 3, 3, 4,  4,  4,  4,  5,  5,  6,  6,
-	7, 7, 7, 7, 8, 8, 9, 9, 10, 10, 10, 11, 11, 11, 11
-};
-static u8 DeltaSwingTableIdx_2GA_P_8188E[] = {
-	0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 4, 4, 4,
-	4, 4, 4, 5, 5, 7, 7, 8, 8, 8, 9, 9, 9, 9, 9
-};
-
 /* 3 ============================================================ */
 /* 3 Tx Power Tracking */
 /* 3 ============================================================ */
@@ -312,11 +298,6 @@ static void GetDeltaSwingTable_8723B(
 			*TemperatureUP_B   = pRFCalibrateInfo->DeltaSwingTableIdx_2GB_P;
 			*TemperatureDOWN_B = pRFCalibrateInfo->DeltaSwingTableIdx_2GB_N;
 		}
-	} else {
-		*TemperatureUP_A   = (u8 *)DeltaSwingTableIdx_2GA_P_8188E;
-		*TemperatureDOWN_A = (u8 *)DeltaSwingTableIdx_2GA_N_8188E;
-		*TemperatureUP_B   = (u8 *)DeltaSwingTableIdx_2GA_P_8188E;
-		*TemperatureDOWN_B = (u8 *)DeltaSwingTableIdx_2GA_N_8188E;
 	}
 }
 
Signature · 1 line
2.43.0
#1 lore raw
Quoted 3 lines

This branch is unreachable. channel comes from CurrentChannel, which is only updated in rtl8723b_phycfg.c after passing HAL_IsLegalChannel(), which enforces a range of 1 to 14

	if ((!Adapter->driver_stopped) && (!Adapter->bSurpriseRemoved)) {
		phy_SwChnlAndSetBwMode8723B(Adapter);
	} else {
		if (pHalData->bSwChnl) {
			pHalData->CurrentChannel = tmpChannel;
			pHalData->CurrentCenterFrequencyIndex1 = tmpChannel;
		}

		if (pHalData->bSetChnlBW) {
			pHalData->CurrentChannelBW = tmpBW;
			pHalData->nCur40MhzPrimeSC = tmpnCur40MhzPrimeSC;
			pHalData->nCur80MhzPrimeSC = tmpnCur80MhzPrimeSC;
			pHalData->CurrentCenterFrequencyIndex1 = tmpCenterFrequencyIndex1;
		}
	}

In else branch pHalData->bSwChnl is still true.
It is possible to write an invalid value to pHalData->CurrentChannel with PHY_HandleSwChnlAndSetBW8723B().

Thanks,
Bera

author
#2 lore raw

Hi Bera,

Thanks for catching this — you're right, and I missed something important in my analysis. bSwChnl isn't reset to false when HAL_IsLegalChannel() rejects a channel, so an unvalidated ChannelNum can still reach CurrentChannel through a stale bSwChnl state. This branch is a defensive fallback, not dead code.

I'm withdrawing this patch. Thanks again for the careful review.

Thanks,
Hank