Lenovo G505s

dispuser5132 · October 23, 2021, 10:44pm

---
layout:
  'hcl'
type:
  'laptop'
hvm:
  'yes'
iommu:
  'yes'
slat:
  'yes'
tpm:
  'no'
remap:
  'yes'
brand: |
  LENOVO
model: |
  G505s
bios: |
  CBET4000 Coreboot 4.12 + SeaBios 1.13
cpu: |
  AMD A10-5750M APU with Radeon(tm) HD Graphics
cpu-short: |
  AMD A10-5750M
chipset: |
  Advanced Micro Devices, Inc. [AMD] Family 15h (Models 10h-1fh) Processor Root Complex [1022:1410]
chipset-short: |
  FIXME
gpu: |
  Advanced Micro Devices, Inc. [AMD/ATI] Richland [Radeon HD 8650G] [1002:990b] (prog-if 00 [VGA controller])
  Advanced Micro Devices, Inc. [AMD/ATI] Sun PRO [Radeon HD 8570A/8570M] [1002:6663]
gpu-short: |
  Radeon HD 8650G
network: |
  Qualcomm Atheros QCA8172 Fast Ethernet (rev 10)
  Qualcomm Atheros QCA9565 / AR9565 Wireless Network Adapter (rev 01)
memory: |
  15853
scsi: |
  Samsung SSD 840  Rev: BB6Q
  DVD-RAM UJ8DB    Rev: 8.51
usb: |
  5
versions:

- works:
    'partial'
  qubes: |
    R4.0
  xen: |
    4.8.5-35.fc25
  kernel: |
    5.13.6-1
  remark: |
    Too much
  credit: |
    disp5132
  link: |
    FIXLINK

---

Remarks

Suspend broken. Random freezes maybe issue with usb. sys-usb needs kernel option “irqpoll”. USB3 ports on left side broken due to missing firmware blob should be working as USB2 but really unstable could not assign any device to a VM. USB2 port on right side works.

All in all it is pretty unstable. Thinking about switching back to original bios.

If anyone is interested in my “coreboot build story” for this device just let me know

Attachments

Qubes-HCL-LENOVO-INVALID-20211023-192635.yml (1.0 KB) (should be Qubes-HCL-LENOVO-G505s-20211023-192635.yml)

Raphael_Balthazar · October 24, 2021, 8:01am

Remarks

Suspend working. No freezes at all. All USB ports working (USB2), no problems assigning any device to a VM (USB is slow, noticeable when copying files); dedicated GPU working (sys-gui) and so on.

All in all pretty stable. I’ve been using it for more than 2 years. Never thought about switching back to the original BIOS.

Maybe your laptop wasn’t working well anymore from the start (?).

It isn’t the fastest laptop out there (I’ve posted the time period for starting a VM somewhere in another thread) but it’s one of the cheapest (if you’re lucky) to be compatible with Qubes 4.0 and 4.1, it has no ME or PSP and with 16GB RAM it works well enough for me. I’ve put 2 SSDs inside and I use a Yubikey for U2F and to secure my keepass.

The only major problem I am having with this laptop is that the Qualcomm Ethernet controller is not working anymore above kernel 5.4.98. I am still hoping that a later kernel will solve the problem.

Sven · October 26, 2021, 12:25am

Thank you @dispuser5132 for your HCL report, which is now online.

Maybe you can still get it going with some help from @Raphael_Balthazar. I am sure you saw the multiple positive R3.2 reports of identical machines with Coreboot. So it ought to be possible to make it work.

In any case, negative reports are equally appreciated!

dispuser5132 · October 29, 2021, 12:08pm

Some problems seemed to be fixed. The 2 ports on left side seem to work now without freezing the system anymore. I didn’t change anything not even update. But still only USB2 flashing again with xhci amd firmware makes the ports unusable. Assigning to VMs works now. Random screensaver problem didn’t happen again ( came up while i was typing, so that was strange).

Suspend still doesn’t work or to be more clear. laptop goes to sleep when closing the lid but does not wake up again. Blackscreen.

@Raphael_Balthazar helped me that some coreboot builds are broken. Seemed that we both followed the recommended build guides from Mike Banon on dangerousprototypes. When i was building with his patches my build failed. Thanks to Raphael

For now i used my own dumped vgabios firmware from stock bios and an older coreboot version.

git reset --hard e6fb1344ed9188e19be4b54bdf1a76680b8c4523

based on mike banons github GitHub - mikebdp2/coreboot-g505s-builds: Some coreboot g505s builds for demo purposes. BUILD YOUR OWN !

for now it works ok for me. maybe i will try again a newer coreboot version with mikes patches.

mike_banon · September 27, 2022, 9:05am

Hi there @dispuser5132 and @Raphael_Balthazar , just found this thread! Thank you for sharing your experiences, I am sorry to hear about your problems 1 year ago:

Seemed that we both followed the recommended build guides from Mike Banon on DangerousPrototypes. When i was building with his patches my build failed.

Have you reported this problem to my [ mikebdp2 [at] gmail [d0t] c0m ] e-mail mentioned at my DP guide? IMHO my guide+script are quite refined, and when they do not work - despite you did everything right and your build environment is OK - a true cause of breakage is some new commit at the coreboot repository.

For an example, take a look at this small commit. This tiny change may look innocent, however it made a coreboot’s ./g505s/irq_tables.c incompatible with my AMD good IRQs patch, causing

...
patching file src/mainboard/lenovo/g505s/irq_tables.c
Hunk #1 FAILED at 1.
1 out of 1 hunk FAILED -- saving rejects to file src/mainboard/lenovo/g505s/irq_tables.c.rej
...

And, since I’m not rebuilding my BIOSes all the time and have the other things to do, this problem stayed for two weeks until a user reported it to me! As a temporary solution I told him to do a git revert for this particular commit and updated my custom patches to restore the compatibility.

Another problem, which happened around 1 year ago (so may have been a cause of your troubles) : a coreboot’s GCC got upgraded to 11.x version, but for unknown-yet-reasons it builds the unstable ROMs for our AMD AGESA boards including G505S. The only known solution is to additionally do

git revert ce134ababd6a444082962ccdfcd34415a647f41e
[press Ctrl+X then Y if your default git editor is nano, to apply this]
git revert b0d87f753c9c517ba906115362d32aa4422fd188
[press Ctrl+X then Y if your default git editor is nano, to apply this]

to downgrade a coreboot toolchain’s GCC to old trusty 8.3.0 version, before building this toolchain with make crossgcc-i386 . Of course this info is a part of a guide already.

I can assure you that, if you’ll try my guide right now, it should be working flawlessly! At least it worked 3 days ago when I rebuilt my BIOS, and it’s unlikely the compatibility has been lost so fast: i.e. last time it lasted 5 months (until the things got broken for 2 weeks because of a minor commit which I mentioned above)