cancel
Showing results for 
Search instead for 
Did you mean: 

Drivers & Software

nop
Adept II

Re: gcc segmentation faults on Ryzen / Linux

pjssilva

Do you boot your Linux from a drive connected over USB 3?

I'm asking because malakudi​ also had those freezes during idle. He found out that it had to do with booting the system from a device connected to a USB 3 port.

The problem was more evident when it was connected to a USB 3 Port from the chipset than one from the CPU.

The freezes are gone since he is using a drive connected directly to SATA for booting Linux.

I could not reproduce those freezes by just booting from a USB 3.0 memory stick so maybe it was a problem with the Linux driver for the USB 3.0 <-> SATA converter chip of his external SSD enclosure?

pjssilva wrote:

Hello,

I think I have a final follow up about the Segfaults in my processor. It works reliably in auto with the memory at 2100MHz. I can also achieve stability with the memory using XMP 2 profile (2400MHz) if I increase its voltage to 1.3V and SOC to 1.05V.

However my processor still suffers from the freeze when idle. For the first time since the RMAed processor arrived I left it idle during the weekend (since I stopped testing for segfault). I arrived today at work and found the computer frozen. It did not recognize the mouse and keyboard anymore and I had to force a reboot. I believe this is another problem that may be solved disabling C6 state. I have just disable it in the BIOS and I will see if the freezes also go away.

0 Kudos
flyinryzen1700
Adept II

Re: gcc segmentation faults on Ryzen / Linux

That would make sense because the system probably can't access any 'system' files to wake up from hibernate which it likely needs to reestablish the USB drivers.  The hibernate file is usually written onto the system disk, so it needs bios level access to that file.

0 Kudos
bsp2020
Elite

Re: gcc segmentation faults on Ryzen / Linux

I got my 1800X replace and it seems not to have seg fault problem. However, I'm still experiencing random hang and reboot problem. I had seg fault problem as well as hang/reboot. I thought that when I get my CPU replaced, both problems would go away. Well...

I have seem some people claiming that hang/reboot problem is related to motherboard, rather the CPU. I have GigaByte GA-AB350M-D3H with the latest BIOS F5a. Is anyone here using the same motherboard? If so, are you having random hang/reboot problem?

0 Kudos
raydude
Elite

Re: gcc segmentation faults on Ryzen / Linux

bsp2020 wrote:

I got my 1800X replace and it seems not to have seg fault problem. However, I'm still experiencing random hang and reboot problem. I had seg fault problem as well as hang/reboot. I thought that when I get my CPU replaced, both problems would go away. Well...

I have seem some people claiming that hang/reboot problem is related to motherboard, rather the CPU. I have GigaByte GA-AB350M-D3H with the latest BIOS F5a. Is anyone here using the same motherboard? If so, are you having random hang/reboot problem?

I have the same motherboard and am running the same BIOS. I have an RMAed R5-1600 and I don't have the hang / reboot problem. I always figured it was a BIOS issue. You just proved me wrong.

What power supply are you using?

Have you tried disabling power savings in the BIOS (I can't remember what it's called). By the way I did that once and it keeps all the cores at full frequency which really makes my system run hot...

0 Kudos
bsp2020
Elite

Re: gcc segmentation faults on Ryzen / Linux

At the moment, I'm using Optimized Default setting. That means I'm running my 3000MHz ram at 2133MHz. But I sill got a reboot this morning. The new processor does seem a bit more stable in that the old one used to hang/reboot almost every other day. The new one went 5 days before hanging...

0 Kudos
pjssilva
Adept III

Re: gcc segmentation faults on Ryzen / Linux

nop​, no, I am not using an USB drive. I am using an NVMe SSD.

Thanks for asking, though.

0 Kudos
flyinryzen1700
Adept II

Re: gcc segmentation faults on Ryzen / Linux

I would suggest disabling hibernate and sleep, see if the problem goes away.  Then, re-enable sleep and see if the problem appears.  I'm guessing it's a problem with hibernation.  I used to have a laptop that would lock-up like that if my Linux system tried to hibernate, turning it off fixed it, of course, it used a little more power in sleep mode.

P.S. Just sent my 1800X back for RMA to Miami.  Is AMD still cross-shipping these?   I'm not sure where they are running this support from, but I only get 1 email per day on my ticket.  Anywhere between 1am CST and 6am CST from Shibu.  He just said I'd get a tracking# when available.

I'm curious, the testing they are doing on the replacement processors.  Am I going to get basically the same quality silicon as the 1700 processor replacements as my 1800X, both capable of 3.6-3.8 base clock?   That would suck.

0 Kudos
synthor
Adept II

Re: gcc segmentation faults on Ryzen / Linux

I read of a few people with the Gigabyte AB350M boards who have those errors.

Never noticed any stability problem with my Gigabyte AB350 Gaming 3. It already has BIOS version F9d. Maybe you all with the M boards should contact Gigabyte to ask for an update.

Edit: F9d is from 2017/09/12 and F5 from 2017/09/15 for the D3H. Very strange..

0 Kudos
ahartmetz
Adept II

Re: gcc segmentation faults on Ryzen / Linux

After about two weeks of delay I received a replacement (UA37SUS, no traces of use / testing) that STILL segfaults while compiling.

I asked for an advance replacement part so I could just swap the CPUs in one go and have minimal downtime. That was denied after a week. I sent in the CPU and I got a message that replacements were out of stock, about two weeks ago. Last Thursday I got a message that my returned CPU had been inspected and the replacement authorized (what happened to we'll send the replacement as soon as the original has a tracking number?). Maybe the connection between the reason for the RMA and the RMA procedure had been lost.


What is a really big deal here is that there is actually no known manufacturing date after which all CPUs are defect free. Yeah, and that I'll have to RMA again.

0 Kudos
excieve
Adept I

Re: gcc segmentation faults on Ryzen / Linux

Replacement still segfaulting is worrying. I'm still waiting for my replacement for almost two weeks now and they haven't even sent me the tracking number yet. All I know it's been delivered to their NL warehouse on 29th Sep. Anticipating customs procedures, fees and taxes I'll likely need to pay for it (I'm in Europe but not EU), I really hope it won't be failing... It's already very frustrating, doing this second time will be just awful.