---------------------------------------- Intel Woes November 28th, 2020 ---------------------------------------- My wife and I own computers that have an Intel wireless card in them. My wife's computer is a laptop and mine is a desktop. Within the past year, we've had impactful bugs hit both computers that have caused wireless QoL to be severely degraded. In my wife's example, the wireless driver simply fails to load with an ambiguous status code. Enough sleuthing around online shows that many other people have hit this bug which remained open up to the last time I checked a month ago. This particular fix was a major regression and for some reason a fix wasn't easily backportable to earlier GNU/Linux kernel versions. I've tried compiling custom forks of mainline Ubuntu kernels with iwlwifi patches to no avail. Something is so seriously wrong that reverting back to even a 3.x kernel version doesn't fix the issue. She's been forced to use a USB wifi card which works reliably through kernel updates but sticks out from a laptop and is asking to be snapped off. The bug I experience on my desktop is much more sneaky. I will have intermittent network drops as my card disconnects and reconnects after a couple of seconds. I notice this much more when playing real time online games as it usually results in a server disconnect or in a loss of about 20 seconds or so of game time for more forgiving game network code. I am on Wi-Fi, but the router is in the room right below where the desktop is kept; it's probably no more than 10 feet away with a floor in between. For my desktop computer, I can easily run some stats on my bug because it has an easy to grep message of the number of times in a particular day that this issue has affected me, sorted by the highest total number of occurrences in a day. $ sudo journalctl | grep "No beacon heard" | cut -c 1-6 | uniq -c | sort -h -r 372 Jul 19 277 Jul 16 224 Jul 18 198 Jun 27 163 Aug 08 140 Jul 15 131 Jun 29 106 Feb 02 100 Jun 28 99 Jul 01 91 Feb 01 83 Nov 28 78 Mar 22 69 Mar 21 62 Nov 27 62 Jul 02 61 Nov 14 61 Jul 11 60 May 24 53 Jul 13 53 Jan 22 51 Sep 13 50 Aug 11 48 Aug 09 47 Jul 12 45 Jan 24 43 Nov 15 41 Jun 05 41 Jul 10 40 Jan 17 39 Jul 20 39 Jul 06 36 May 28 36 Feb 03 35 Jul 05 35 Feb 23 35 Aug 07 34 Jul 04 33 Sep 27 33 Jul 17 32 Aug 04 31 Feb 04 30 Mar 27 29 Nov 16 29 Nov 11 28 Nov 23 28 Nov 20 28 May 25 28 Jul 08 28 Aug 05 27 May 23 24 Mar 30 23 Feb 24 22 Nov 26 22 May 27 21 May 22 21 Jul 07 20 Sep 12 20 Nov 12 19 Mar 29 18 Aug 06 17 Aug 12 16 Nov 22 16 Mar 20 15 Jun 07 15 Aug 13 14 Oct 26 13 Aug 10 12 Nov 19 12 Jul 03 12 Jan 15 12 Feb 09 11 Sep 28 11 Sep 26 11 Nov 17 10 Sep 24 10 Nov 26 10 May 31 9 Oct 29 8 Jun 01 7 May 26 7 Mar 04 7 Feb 22 6 Oct 08 6 Mar 28 6 Jul 09 6 Jan 09 6 Aug 14 5 Mar 15 5 Mar 08 5 Jun 16 5 Jan 29 5 Jan 23 5 Jan 12 5 Feb 21 5 Feb 16 4 Oct 28 4 Nov 13 4 May 02 4 Mar 01 4 Jun 11 4 Jan 26 4 Jan 11 4 Dec 01 3 Sep 25 3 Oct 22 3 Oct 20 3 Nov 07 3 Mar 14 3 Mar 07 3 Jun 26 3 Jul 26 3 Jan 30 3 Jan 14 3 Jan 04 3 Jan 02 3 Feb 18 3 Feb 17 3 Feb 08 3 Apr 29 2 Oct 25 2 Oct 09 2 Oct 03 2 Nov 20 2 May 16 2 Mar 16 2 Jan 06 2 Feb 28 2 Dec 06 2 Apr 26 2 Apr 11 2 Apr 10 1 Oct 24 1 Oct 23 1 Oct 19 1 Oct 03 1 Nov 19 1 Nov 17 1 May 03 1 Jul 14 1 Jan 13 1 Jan 05 1 Feb 29 1 Feb 14 1 Dec 07 1 Aug 02 1 Apr 09 Seeing the bug tracker is even worse -- bugs that have been open for more than a year with little follow-up from the developer's side. Most of these bug trackers are filled with people discussing hacks as workarounds when nothing is being done. Both of the open bugs affecting my wife and I are regressions to existing functionality; both of us were plagued with these bugs after kernel updates. It makes me wonder what sort of testing (if any) that Intel is doing with their GNU/Linux drivers. Is there something that can be done to help? I'm usually not perterbed when it comes to hardware and Linux capability but running into these issues will make me very cautious about buying laptops or components that include Intel technology in the future. I've had my fair share of issues with Realtek-based wireless chipsets as well, but those bugs seem to be patched more easily and readily than anything Intel puts out.