On Sat, 2020-03-14 at 10:57 +0000, Jonathan McDowell wrote: > Commit 39c1a9728f93 refactored the SAR table support and inadvertently > added a call to GEO_TX_POWER_LIMIT without checking that the firmware > supported it. This caused a regression for device + firmware > combinations lacking the support. Add a check in iwl_mvm_sar_init to > prevent this. > > Fixes: 39c1a9728f93 (iwlwifi: refactor the SAR tables from mvm to acpi) > Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=206395 > Cc: stable@xxxxxxxxxxxxxxx # 5.5 > Signed-Off-By: Jonathan McDowell <noodles@xxxxxxxx> > > diff --git a/drivers/net/wireless/intel/iwlwifi/mvm/fw.c b/drivers/net/wireless/intel/iwlwifi/mvm/fw.c > index 54c094e88474..4a8a4cb2c553 100644 > --- a/drivers/net/wireless/intel/iwlwifi/mvm/fw.c > +++ b/drivers/net/wireless/intel/iwlwifi/mvm/fw.c > @@ -1007,6 +1007,13 @@ static int iwl_mvm_sar_init(struct iwl_mvm *mvm) > if (ret == -ENOENT) > return 1; > > + /* > + * If our firmware doesn't support GEO_TX_POWER_LIMIT then skip trying > + * to use it. > + */ > + if (!iwl_sar_geo_support(&mvm->fwrt)) > + return 1; > + > return ret; > } Thanks, Jonathan! We already have an internal patch that fixes this (also fixes another symptom of the same bug). I'll send it out now. -- Cheers, Luca.
