Hey Folks,

I have a bit of a problem that is getting out of hand. It is that my syslog file grows many gigs in day. Last Saturday it grew by 31 gigs in one day and filled up my hard drive. So, I finally realized I needed to get the root of this problem fixed. So I displayed a few lines . . . and was totally stumped. I have no idea what is causing these messages. Here is a small sample of what I am getting.

Sep 29 08:12:14 prometheus kernel: [ 255.944943] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.948648] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.948871] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.960571] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.960823] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.964953] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.968647] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.980051] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.980342] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.984487] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.988858] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 255.989264] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.000919] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.001160] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.005289] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.005620] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.005626] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.006115] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.018942] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.019146] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.023296] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.303040] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]
Sep 29 08:12:14 prometheus kernel: [ 256.321177] AMD-Vi: Event logged [IO_PAGE_FAULT device=01:00.0 domain=0x0016 address=0x0000000000001000 flags=0x0000]


Prometheus is the name of my computer. I have an AMD processor. Other than that, what do you folks think gives?

Thanks,
Bob

If it started recently, you might check to see if the machine is full of dust bunnies.

Otherwise, there are a few boards that are helped with this option in the kernel boot flags (added in, for example the append= section of lilo.conf and then reinstalling lilo):

iommu=soft

3 members found this post helpful.

A quick search indicates this may be the result of a known bug with some mmu implementations. You might have luck with iommu=pt or iommu=soft in your kernel arguments if that is the issue.

However it seems like you are implying this was a working system, chugging along normally for some time before this problem suddenly started, and there is no new hardware, no change to kernel config? If that's the case I am afraid it's sounding like it could be a sign of failing hardware too.

EDIT: Ah and the master posted while I was typing. He's right too. Check for dust bunnies. I've solved several apparent hardware failures over the years by giving the case and fans a good clean. An overheated system can do all kinds of crazy stuff, and dust bunnies can really impede your cooling.

Thanks for the quick response guys! You know, I did have to tweak my iommu settings in my bios, as I recall, so you both are right in telling me to look there.

Arkerless, yes this system is old, but this problem has been there since the beginning -ish. I just ignored it before because I was using this system for trading currencies so software design and testing was a higher priority. I quickly wrote a script that I put in the cron.daily directory that would delete the old syslog.X files. And I'm afraid the problem sort of fell into the "out of sight out of mind" category -- until this last weekend. Pretty sloppy sys admin, I know.

Anyway, I can't shut this system down until Friday night this week, as it is trading even as I write. So I'll try Patrick's and Arkerless suggestions then and post back my results.

Thanks for the responses again. And Patrick, thanks for the awesome distribution. When I get asked why I use Slackware, my easy answer is: "it works." A lot of heartache has been put into those two words. Actually, I have some pretty unique requirements and other distributions just don't seem to be able to stay up and running for days/weeks on end.

Thanks again,
Bob

1 members found this post helpful.

I had a similar problem on my slackware64 14.1 running on a AMD-a6.. After updating my system to "current" the problem has gone away. The newer kernel fixed my problem.

Thank you so much Patrick. I had the same issue with an AMD A10-7850K APU with an MSI A88XM E45 Motherboard on a new build.

My symptoms were that eth0 would drop out at random, frequent times and sometimes restarting rc.inet1 would help, sometimes not.

I got to the point where I was going to flash the bios even though it did not look like that would have helped.

I finally came across the AMD-Vi report in dmesg and found this thread. Your fix of appending iommu=soft to lilo.conf solved the issue.

This box is running very sweet now on Slackware64-current.

I have the same hardware (A10-7850K / A88XM-E45 with BIOS 25.3) running -current, and the same problem:

[ 0.344119] AMD-Vi: Found IOMMU at 0000:00:00.2 cap 0x40
[ 0.344183] AMD-Vi: Extended features: PPR GT IA PC
[ 0.344387] AMD-Vi: Interrupt remapping enabled
[ 0.344467] pci 0000:00:00.2: irq 72 for MSI/MSI-X
[ 0.345374] AMD-Vi: Using passthrough domain for device 0000:00:01.0
[ 0.350388] AMD-Vi: Event logged [IO_PAGE_FAULT device=00:12.0 domain=0x000d address=0x000000007ddfe040 flags=0x0010]
... repeat many times ...
[ 0.361452] AMD-Vi: Event logged [IO_PAGE_FAULT device=00:13.0 domain=0x000f address=0x00000000ffffffc0 flags=0x0010]
[ 0.409717] GHES: HEST is not enabled!

And then it quits. The two devices are built-in OHCI controllers. I also had this problem with a Broadcom wireless controller until I changed it out with an Atheros. I put this system together in January. I played with BIOS settings without success. Seems like there should have been a fix by now (vs. turning it off).

Andy

All,

As I mentioned in my last post I had to wait until this weekend to try Patrick's and Arkerless' suggestions. So, I implemented their recommendations today and it fixed my problems too! WAHOO!!!! Thanks all.

Now, I want to mark this bad boy as solved. I swear I've seen some button or something around here to do that . . .

Heck, I'm just gonna write it in the title.

Thanks again!

Bob

Hah! Finally found the Solved link after I posed my last reply.

I think the purpose of my reply was misinterpreted. In my world FOUND_WORKAROUND != FIXED.

Andy

it's a kernel bug
try finding it here https://bugzilla.kernel.org/