Re: iwlwifi: AC9560 firmware crash with iwlwifi-QuZ-a0-jf-b0 versions above 55

From: Johannes Berg <johannes@sipsolutions.net>
To: "Íñigo Huguet" <ihuguet@redhat.com>
Cc: gregory.greenman@intel.com, golan.ben.ami@intel.com,
	jonathan.sade@intel.com,
	Linux List Kernel Mailing <linux-wireless@vger.kernel.org>,
	linux-bluetooth@vger.kernel.org,
	Jose Ignacio Tornos Martinez <jtornosm@redhat.com>,
	Laura Trivelloni <ltrivell@redhat.com>
Subject: Re: iwlwifi: AC9560 firmware crash with iwlwifi-QuZ-a0-jf-b0 versions above 55
Date: Wed, 05 Oct 2022 12:06:34 +0200	[thread overview]
Message-ID: <5c47143aea07c34dbe95fdb2cab677f2a7a753fb.camel@sipsolutions.net> (raw)
In-Reply-To: <CACT4oufKgUG5n6Ocgi6r-Z=+Nt9vNVK1e5dJDccogmd8h7xM8Q@mail.gmail.com>

On Wed, 2022-10-05 at 11:50 +0200, Íñigo Huguet wrote:
> On Wed, Oct 5, 2022 at 11:35 AM Johannes Berg <johannes@sipsolutions.net> wrote:
> > 
> > On Fri, 2022-09-09 at 12:41 +0200, Íñigo Huguet wrote:
> > > With an AC9560 installed on a system with Quasar (11th gen processor,
> > > in my case) we're getting a firmware crash as soon as you try to do
> > > something with the card. The problem doesn't happen with the card
> > > installed in older platforms (8th / 9th gen), and it neither happens
> > > with firmware iwlwifi-QuZ-a0-jf-b0-55 version, but yes with any other
> > > newer version.
> > > 
> > > I've tried different versions of kernel and firmware, including the
> > > latest kernel and linux-firmware available today, and all of them fail
> > > unless I remove all fw files above version 55.
> > > 
> > 
> > I believe this is a duplicate of
> > 
> > https://bugzilla.kernel.org/show_bug.cgi?id=214343
> > 
> > which was supposed to be fixed in firmware, but evidently that fix was
> > never backported to older images.
> > 
> > Internally, FW commit d8887278 fixed this, so FW version 69 should be
> > OK?
> 
> I don't think so, we've reproduced it with iwlwifi-QuZ-a0-jf-b0-71.ucode
> 

Indeed, my git-fu was bad. Looks like 74 only fixes it (or will, since
it hasn't been released, apparently).

We could do some awful workaround like this I guess:

diff --git a/drivers/net/wireless/intel/iwlwifi/fw/acpi.c b/drivers/net/wireless/intel/iwlwifi/fw/acpi.c
index 009a4074d062..0c0a28d748f2 100644
--- a/drivers/net/wireless/intel/iwlwifi/fw/acpi.c
+++ b/drivers/net/wireless/intel/iwlwifi/fw/acpi.c
@@ -1119,6 +1119,11 @@ int iwl_read_ppag_table(struct iwl_fw_runtime *fwrt, union iwl_ppag_table_cmd *c
         int i, j, num_sub_bands;
         s8 *gain;
 
+	/* many firmware images for JF lie about this */
+	if (CSR_HW_RFID_TYPE(fwrt->trans->hw_rf_id) ==
+	    CSR_HW_RFID_TYPE(CSR_HW_RF_ID_TYPE_JF))
+		return -EOPNOTSUPP;
+
         if (!fw_has_capa(&fwrt->fw->ucode_capa, IWL_UCODE_TLV_CAPA_SET_PPAG)) {
                 IWL_DEBUG_RADIO(fwrt,
                                 "PPAG capability not supported by FW, command not sent.\n");


johannes
