On 5/25/2022 13:45, Pavel Machek wrote:
Hi!
From: Sasha Neftin sasha.neftin@intel.com
commit 79cc8322b6d82747cb63ea464146c0bf5b5a6bc1 upstream.
The device ID for I226_K was incorrectly assigned, update the device ID to the correct one.
Fixes: bfa5e98c9de4 ("igc: Add new device ID")
I don't see updating the ID, I see adding an unused define. I don't think this is suitable for stable. Same thing goes for previous two patches, they don't really fix anything.
Pavel, 1. Commit bfa5e98c9de4 added new device ID for i225/226 parts. Commit 79cc8322b6d8 just fixed number for IGC_DEV_ID_I226_K. This number comes from NVM and PCIe configuration space initialized with this number. (In case you will use wrong number SW won't work) 2. Regards PHY_ID: here is two things: i225/i226 parts have only one PHY (there is no option for another PHY) some unit upon specific power up condition could wrong represent PHY ID and we do not want block SW. 3. phy->type - just clean. you indeed can skip it.
5106 O Greg Kroah ├─>[PATCH 5.10 05/97] igc: Remove _I_PHY_ID checking 5107 O Greg Kroah ├─>[PATCH 5.10 06/97] igc: Remove phy->type checking 5108 Greg Kroah ├─>[PATCH 5.10 07/97] igc: Update I226_K device ID
Best regards, Pavel
+++ b/drivers/net/ethernet/intel/igc/igc_hw.h @@ -22,6 +22,7 @@ #define IGC_DEV_ID_I220_V 0x15F7 #define IGC_DEV_ID_I225_K 0x3100 #define IGC_DEV_ID_I225_K2 0x3101 +#define IGC_DEV_ID_I226_K 0x3102 #define IGC_DEV_ID_I225_LMVP 0x5502 #define IGC_DEV_ID_I225_IT 0x0D9F #define IGC_DEV_ID_I226_LM 0x125B
Thanks, Sasha