A lot of this sounds like my issues, so I thought I'd give my side. I run an AMD thinkpad A485 with ryzen 2500u pro. Not what I'd call a low-end laptop, but the issues are mostly the same.
On 7/22/19 11:09 PM, Claudia wrote: > Claudia: >> Chris Laprise: >>> On 7/22/19 11:38 AM, Claudia wrote: >>>> So I finally got around to doing this. >>>> >>>> Qubes works and all the basic features are supported, VT-x VT-d, and >>>> so on, as far as I can tell. >>>> >>>> One major issue, hardware/firmware-wise: >>>> >>>> 1) It doesn't come back from suspend. The fan stops, but there are >>>> no blinking lights (actually, no lights besides AC and caps lock), >>>> and nothing I do wakes it. I have to long-press the power button, >>>> then press again to turn on the machine. It's probably an ACPI >>>> issue, probably not a graphics driver issue as there's no dGPU. I >>>> played around with acpi_osi= and some basic troubleshooting but the >>>> odds of me being able to fix it are slim to none. >>> >>> This is good to hear! >> I've got the same issue on my thinkpad. It seems to suspend fine, but when I try to wake it, only the fan and keyboard backlight turns on. Nothing else. It seems, however, that when I hard reset it, it seems to think it's resuming from suspend. Not sure if that's a clue. I did try fedora on this machine though, and it had the same issues there, so it doesn't look qubes-specific, at least for me. >>> >>>> >>>> Minor issues: >>>> >>>> 2) Buggy BIOS / ACPI impl. Dom0 kernel dmesg complains about >>>> "[Firmware bug]" and ACPI issues. Though no noticeable problems >>>> except suspend. The firmware's OS-less update-from-USB-drive feature >>>> seems broken, I tried several times. Still might be able to update >>>> from fwupd, though. No UI for managing secure boot keys, etc., it >>>> seems to only have the bare minimum options. >>>> >>>> So, it appears you were totally right about consumer-grade laptops >>>> and buggy firmware. But suspend/resume is problematic even among >>>> high-end/business-class laptops, too, isn't it? It's just something >>>> Linux has never been good at. I think I have the same issue, I get ACPI errors, one for each CPU kernel. So I'd say this is not a "cheap laptop" issue, mine is mid-range, branded towards IT security staff. >>> >>> TBH, I haven't had long-term problems with suspend on Thinkpads >>> _except_ when anti-evil-maid is enabled; that combo hasn't worked for >>> about 2 years. >>> >>> My experience is that consumer BIOS will result in poor suspend >>> compatibility. >> >> Not what I wanted to hear of course, but it does seem that way. I >> can't say you didn't warn me. But that's the *only* firmware/hardware >> issue I've run into so far. I feel like I'm so close! >> >>> >>>> >>>> 3) USB qube isn't working. I installed with USB qube, and the >>>> microphone shows up fine. But flash drives and the card reader don't >>>> show up. When I plug in a USB drive, its LED blinks on for a >>>> fraction of a second, then turns off (on other machines it stays >>>> on). No sign of it in lsusb or lsblk in either sys-usb or dom0. >>>> >>>> However, when I remove the qubes.rd.hide_all_usb kernel flag, it >>>> works normally, so I think this is just a software issue. >>> >>> This could have to do with how the USB controllers respond to the >>> steps involved in placing it under IOMMU passthrough. I think without >>> hide_all_usb set, they get reset twice (once in dom0 and once in >>> sys-usb)? >> >> As long as it's not strictly firmware/hardware-related, I'll worry >> about it later and start another thread for it. If I have to get rid >> of the USB qube or something it's not that big of a deal. Same issue for me here as well. Mine looks to me to be IOMMU-group-related. I believe that something in the usbs' IOMMU group is connected to amdgpu. If I boot with rd.qubes.hide_all_usb, I have to have nomodeset. It looks like a collision with something in the graphics driver, think I saw somehting floating past the screen during one crash. Sadly have no record. >> >>> >>>> >>>> 4) Screen power management (turn off display) doesn't work, although >>>> I had the same problem with a machine where suspend does work, and I >>>> think I narrowed it down to a fedora/X11 issue. The display does >>>> turn off when the lid is closed and lid-switch is set to "do >>>> nothing," though. >>> >>> I usually have to switch to KDE with sddm to get this working. >> >> I think I had it working temporarily on my current machine by directly >> using X11 commands, it was just XFCE not using them correctly or >> something. It must happen to a lot of people, if you ran into it as >> well. That's something I can worry about later too. I'll keep sddm in >> mind and see if that fixes it. >> This one works for me, but I also run KDE. Don't remember it from XFCE though, I think that worked for me there. >>> >>>> >>>> 5) ...plus a few other minor issues probably not hardware related. >>>> >>>> >>>> >>>> Right now I'm trying to decide if I can live without suspend. But, >>>> this is such a common problem that I'm afraid the next one I trade >>>> it in for would have the same problem, and the next one after that. >>>> Then I spent twice the money and got nowhere. This issue is >>>> all-too-common on laptops running Linux. It could be fixed (or >>>> broken) on any machine at any time in a random kernel update, too, >>>> but who knows. >>>> >>>> This is especially a problem because Xen doesn't support hibernation >>>> at all (not to mention whether it would actually work), and Qubes >>>> doesn't support Xen's "save VM state" feature, either of which I >>>> could live with instead. So my only choices are "on" and "off." >>> >>> This is an excellent point, and I think there is a Qubes issue about >>> VM hibernation... >> >> #2414, which hasn't had any activity in two years. >> >>> >>>> >>>> Besides suspend being broken, I actually really like it, and you >>>> can't go wrong for the price. >>>> >>>> I think I'm going to try installing Ubuntu and testing suspend from >>>> there, and also trying to update the firmware from fwupd, but I'm >>>> not holding my breath. >>> >>> That's what I would also try first. Qubes 2.0 used to make my >>> ethernet NIC go dead, but booting temporarily with an Ubuntu live cd >>> would get it working again and I could use it in Qubes after that >>> until I did a Qubes-to-Qubes reboot. Problem stopped around Qubes >>> 4.0. :) >>> >>>> >>>> So, any advice on troubleshooting suspend... or advice on what to do >>>> next, I guess... would be appreciated. Ugh, this is totally >>>> frustrating. >>> >>> You should try these: >>> >>> * Find which wifi modules are being used in sys-net (i.e. do "sudo >>> lsmod") then add them to /rw/config/suspend-module-blacklist. I find >>> this is usually required to get suspend working right. For an Intel >>> wifi card, you would add both 'iwldvm' and 'iwlwifi' in that order. >> >> I didn't think of wifi modules preventing suspend. I'll definitely >> give it a try, but I'm not sure it could cause the kind of problem I'm >> having. It doesn't seem like it's even trying to wake up when I press >> a key or the power button. It just stays sleeping. The only observable >> difference between sleep and off is that in sleep pressing the power >> button doesn't turn the machine on until I power cycle. Also, I tried >> enabling "USB Wake Support" in BIOS, but it didn't seem to make a >> difference. This looks interesting, I'll give this a try. I just shelved my suspension issues, since I thought it wasn't qubes-specific, was going to nag lenovo about it... >> >>> >>> * Upgrade the dom0 and vm kernels to 4.19 or later. The 4.19 versions >>> from qubes*testing have been very stable for me. OTOH, there are also >>> 5.x versions available. >>> >> >> I did `qubes-dom0-upgrade --enablerepo=qubes-dom0-unstable kernel` and >> I still couldn't get a newer kernel. 4.14.199-2 popped up and it said >> "nothing to do." What am I doing wrong? > > Got it. current-testing, not -unstable. Sometimes I don't know how to > read, haha. I guess I just assumed unstable was newer than testing. > Confused me as well... But yeah, 4.19 seems really important for ryzen, a lot apparently happened there. >> >> It doesn't seem like VM kernels would make a difference with suspend, >> but I can try upgrading them anyways. >> >> Thanks again > > > ------------------------------------------------- > This free account was provided by VFEmail.net - report spam to > ab...@vfemail.net > > ONLY AT VFEmail! - Use our Metadata Mitigator to keep your email out of > the NSA's hands! > $24.95 ONETIME Lifetime accounts with Privacy Features! 15GB disk! No > bandwidth quotas! > Commercial and Bulk Mail Options! So yeah, a lot of this looks like it's ryzen-specific, and not due to consumer-grade hardware. Mine isn't the most expensive thinkpad around, I guess, but I can't really call it cheapish. <3 /panina -- You received this message because you are subscribed to the Google Groups "qubes-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to qubes-users+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/qubes-users/97da8240-54be-4086-a22f-45b595065868%40nonbinary.me.
0x6648B5C5E394CC24.asc
Description: application/pgp-keys
signature.asc
Description: OpenPGP digital signature