From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from m42-4.mailgun.net ([69.72.42.4]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1kZxEl-0005vi-Ix for ath11k@lists.infradead.org; Tue, 03 Nov 2020 14:24:43 +0000 From: Kalle Valo Subject: Re: PROBLEM: unable to get QCA6390 to work References: <87lffjodu7.fsf@codeaurora.org> <87ft5rszcs.fsf@codeaurora.org> Date: Tue, 03 Nov 2020 16:23:58 +0200 Message-ID: <87ft5qsem9.fsf@codeaurora.org> MIME-Version: 1.0 List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: "ath11k" Errors-To: ath11k-bounces+kvalo=adurom.com@lists.infradead.org To: Pavel Procopiuc Cc: ath11k@lists.infradead.org Kalle Valo writes: > Pavel Procopiuc writes: > >> After doing some testing I was able to get it running under some >> specific conditions. I took the v5.9 tag of the main kernel tree and >> merged in the ath11k-qca6390-bringup-202009250825 tag from the ath >> tree. I had to fix some trivial conflicts, but the resulting kernel >> together with the firmware (amss.bin, m3.bin and board-2.bin) gave me >> a working driver that I was able to use to connect to my AP in both >> 2,4 and 5GHz. >> >> I then tried to copy over the ath11k driver source resulting from that >> merge on top of the 5.10-rc2 tag. After fixing some API breakage >> (IEEE80211_TX_CTRL_HW_80211_ENCAP and ieee80211_set_hw_80211_encap) to >> make it compile I booted into the resulting kernel, but ath11k driver >> was giving me the same firmware errors I had initially on 5.10-rc2. So >> it looks like there might be some change in 5.10 that broke the logic. > > Interesting, and worrisome. I have not yet tested v5.10-rc releases, > I'll do that ASAP and try to reproduce your issue. If I see the same > problem then it should be possible to bisect the issue pretty easily. For me v5.10-rc2 (no extra patches) was working fine. A bisection would be a good next step. What I suggest that is that you first test these three commits from Linus' tree in this order: fd5ad4d1e980 wcn36xx: Advertise beacon filtering support in bmps 3650b228f83a Linux 5.10-rc1 3cea11cd5e3b Linux 5.10-rc2 Commit fd5ad4d1e980 is based on v5.9-rc6 so try that first. In these three commits ath11k is practically identical, the only change is a small bugfix in commit 8431350eee2e ("ath11k: Fix memory leak on error path"). If commit fd5ad4d1e980 works for you then use that as "good" commit for git-bisect. And use 3650b228f83a or 3cea11cd5e3b as the "bad" commit, depending on which one was broken first. Hopefully the bisect would give us more clues what's happening. -- https://patchwork.kernel.org/project/linux-wireless/list/ https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k