FedoraForum.org - Fedora Support Forums and Community
Page 1 of 2 1 2 LastLast
Results 1 to 15 of 16
  1. #1
    Join Date
    Feb 2009
    Location
    Florida
    Posts
    629

    Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    Recently I have been experiencing problems booting kernel updates on my GigaByte GA-990FXA-UD7 Rev1.1 motherboard running with a AMD Phenom II x6 1100T.
    I currently have 4.14.14, 4.14.16, and 4.14.18 kernels listed on my grub boot menu. Only 4.14.14 will boot without freezing at a black screen. For the last couple of months it has been hit or miss whether an upgrade will successfully boot. Since I now have the last 2 upgrades failing, I decided I needed to isolate this problem before the next update wipes me out.

    The 4.14.18 boot process is not getting far enough to begin the journal logging process. I deduce this from the fact that requesting a "journalctl -b -1" only returns my last successful boot, no info at all from the intervening unsuccessful boot. Editing the kernel command line to remove "rhgb" and "quiet" and then add " rd.shell rd.debug log_buf_len=1M" results in good news and bad news.

    The good news is that the 4.14.18 kernel now boots successfully. The bad news is no issues are identified in the debug logs (probably because the boot was successful).

    Anyone else experiencing similar issues? Any insights to share on how to isolate this issue? Better yet anyone solved this issue?
    Laptop: Toshiba / Intel B960 2.20Ghz x2/ 4GB/ 320GB SataII/ Intel HD/ fc28.x86_64
    Tower: GigaByte (970A) / AMD FX 8320 3.5Ghz x8/ 16GB/ 9TB Sata III/ AMD 6770HD/ fc28.x86_64
    Bookshelf: Shuttle DS61 (H61)/ i3-3225 3.3Ghz x2/ 16GB/ 320GB Sata II/ Intel HD 4000/ fc27.x86_64
    Embedded: BeagleBone Blk / ARM AM3358 1 GHz x1/ 512MB/ 2GB eMMC/ PowerVR SGX530/ fc27.armv7hl

  2. #2
    Join Date
    May 2011
    Posts
    222

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    Quote Originally Posted by Kobuck
    Since I now have the last 2 upgrades failing, I decided I needed to isolate this problem before the next update wipes me out.
    You can always manually remove the middle kernel version using dnf, which always leaves you with two kernel versions. The next kernel update will not remove any since you now have 3 versions.

    You can repeat the manual removal of middle version until the latest kernel fixes your issue or you are always left with two kernel versions: latest version and your working version.

  3. #3
    Join Date
    Feb 2009
    Location
    Florida
    Posts
    629

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    Still have the issue with the 4.15.3 update.

    Activating debug logging still allows successful boot. Tried just removing "quiet" from command line, but no joy until debug was enabled.

    This is my workaround for now. Logging is relatively painless and the settings are automatically carried forward in kernel updates.
    Laptop: Toshiba / Intel B960 2.20Ghz x2/ 4GB/ 320GB SataII/ Intel HD/ fc28.x86_64
    Tower: GigaByte (970A) / AMD FX 8320 3.5Ghz x8/ 16GB/ 9TB Sata III/ AMD 6770HD/ fc28.x86_64
    Bookshelf: Shuttle DS61 (H61)/ i3-3225 3.3Ghz x2/ 16GB/ 320GB Sata II/ Intel HD 4000/ fc27.x86_64
    Embedded: BeagleBone Blk / ARM AM3358 1 GHz x1/ 512MB/ 2GB eMMC/ PowerVR SGX530/ fc27.armv7hl

  4. #4
    Join Date
    Feb 2009
    Location
    Florida
    Posts
    629

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    Still have the issue with 4.15.4.

    Tried changing all kinds of BIOS settings to see if I could have an affect ( CPU Lock/Unlock, boot disk priority list, IOMMU On/Off, memory speed... ) anything I thought would change to system operating profile. NO JOY!

    The only thing that consistantly works is activating the debug log. Going to review BugZilla and add my 2 cents. They're just gonna complain that I am providing insufficient info. duhhh!!
    Laptop: Toshiba / Intel B960 2.20Ghz x2/ 4GB/ 320GB SataII/ Intel HD/ fc28.x86_64
    Tower: GigaByte (970A) / AMD FX 8320 3.5Ghz x8/ 16GB/ 9TB Sata III/ AMD 6770HD/ fc28.x86_64
    Bookshelf: Shuttle DS61 (H61)/ i3-3225 3.3Ghz x2/ 16GB/ 320GB Sata II/ Intel HD 4000/ fc27.x86_64
    Embedded: BeagleBone Blk / ARM AM3358 1 GHz x1/ 512MB/ 2GB eMMC/ PowerVR SGX530/ fc27.armv7hl

  5. #5
    Join Date
    Jun 2005
    Location
    Montreal, Que, Canada
    Posts
    5,053

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    Hi Kobuck

    When issues you describe occur for me, it is because I used the computer bios to overclock the cpu and the ram. Are you using the vanilla bios settings?
    On the install ISO there is / was a memtest program. Try running it for a few minutes to see if anything abnormal occurs.
    Leslie in Montreal

    Interesting web sites list
    http://forums.fedoraforum.org/showth...40#post1697840

  6. #6
    Join Date
    Feb 2009
    Location
    Florida
    Posts
    629

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    I think I have a slight overclock in place to enable the use of the 1600MHz memory. This has been in place since I built the machine in 2011. Why would it suddenly be an issue??
    I did change memory settings as part of testing this AM, but I think I will go back and test with a VERY conservative setting.
    Will report back results here.
    Laptop: Toshiba / Intel B960 2.20Ghz x2/ 4GB/ 320GB SataII/ Intel HD/ fc28.x86_64
    Tower: GigaByte (970A) / AMD FX 8320 3.5Ghz x8/ 16GB/ 9TB Sata III/ AMD 6770HD/ fc28.x86_64
    Bookshelf: Shuttle DS61 (H61)/ i3-3225 3.3Ghz x2/ 16GB/ 320GB Sata II/ Intel HD 4000/ fc27.x86_64
    Embedded: BeagleBone Blk / ARM AM3358 1 GHz x1/ 512MB/ 2GB eMMC/ PowerVR SGX530/ fc27.armv7hl

  7. #7
    Join Date
    Feb 2009
    Location
    Florida
    Posts
    629

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    Well... Thanks Leslie that was a path I hadn't solidly evaluated.
    However, still no Joy.
    I walked back to the basic Motherboard settings of 1333MHz for memory but there was no effect on boot ( 3 tries )

    The CPU is an unlocked Black, the memory is rated at 1600MHz, and the MB supports overclocking. The settings I use are just a slight bump to enable 1600MHz memory which no "real" overclocker would accept as an adequate challenge. Plus I have been operating with these settings for 6 years.

    Nice thought, but I don't think this is the answer.
    I have submitted a Bugzilla since I didn't see anything there yet. I do see several posts on this forum re: AMD equipment which seem similar so there may be hope for a fix.
    Laptop: Toshiba / Intel B960 2.20Ghz x2/ 4GB/ 320GB SataII/ Intel HD/ fc28.x86_64
    Tower: GigaByte (970A) / AMD FX 8320 3.5Ghz x8/ 16GB/ 9TB Sata III/ AMD 6770HD/ fc28.x86_64
    Bookshelf: Shuttle DS61 (H61)/ i3-3225 3.3Ghz x2/ 16GB/ 320GB Sata II/ Intel HD 4000/ fc27.x86_64
    Embedded: BeagleBone Blk / ARM AM3358 1 GHz x1/ 512MB/ 2GB eMMC/ PowerVR SGX530/ fc27.armv7hl

  8. #8
    Join Date
    Feb 2009
    Location
    Florida
    Posts
    629

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    Well I updated to 4.15.6 this AM and still no progress. Reading bugzilla on this gives no confidence that even the work coming out in 4.16.x releases will address this area.

    I think I'm going to invest ~$200 to bring my MB and processor forward a few years to 2014-2015 timeframe. I don't want to trigger a complete rebuild just move forward enough to get past some the design issues that seem to be in play here. ( ACPI tables and PCIe v2.0 )

    Wish me luck!!
    Laptop: Toshiba / Intel B960 2.20Ghz x2/ 4GB/ 320GB SataII/ Intel HD/ fc28.x86_64
    Tower: GigaByte (970A) / AMD FX 8320 3.5Ghz x8/ 16GB/ 9TB Sata III/ AMD 6770HD/ fc28.x86_64
    Bookshelf: Shuttle DS61 (H61)/ i3-3225 3.3Ghz x2/ 16GB/ 320GB Sata II/ Intel HD 4000/ fc27.x86_64
    Embedded: BeagleBone Blk / ARM AM3358 1 GHz x1/ 512MB/ 2GB eMMC/ PowerVR SGX530/ fc27.armv7hl

  9. #9
    Join Date
    Feb 2009
    Location
    Florida
    Posts
    629

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    Update:
    I've replaced my AMD 1100Tx6 with a more recent model AMD FX8320x8 and, replaced the GigaByte 990FXA MB with a GigaByte 970A-DS3P FX. The motherboard replacement provided a more recent AMI BIOS dated 7/13/2017 (5 years newer).

    Results: All of the ACPI errors and all of the "do_IRQ: 1.55 No irq handler for vector" errors are gone. the IRQ errors are what caused me to upgrade my system.

    However, the basic problem still exists. Plus 2 new ones which gave me insight as to what's going on.
    A new conflict with the sp5100_tco watchdog process lead me to the "i2c_piix4" module.

    This reminded me of the problems I had installing a 4 port SATA III expansion card several months back. It used two lanes in a PCIe ver 2 slot and did not play nice with the AMD IOMMU controller.
    The newer kernel releases seem to be reopening an issue I thought I had solved.
    The second issue is the new MB uses a VIA VL805 chip to implement USB3.1 support and it was also affected by this expansion card problem.

    Adding file "i2c_piix4.conf" containing the line "blacklist i2c_piix4" to the "/etc/modprobe.d/" directory solved both the sp5100_tco and VIA chip issues. This seems to have no effect on the operation of the expansion card???

    I really like my system configuration enabled by the extra SATA drives so I'm going to continue trying to knock this problem down. I'm currently trying to discover to what extent expansion cards can be managed/configured by Linux with the hope that moving some address assignments can relieve the conflicts which are arising. Maybe learning more about the debug kernel will help.

    Sigh!!! I thought PCIe was supposed to make this all painless.
    Last edited by Kobuck; 6th March 2018 at 03:06 AM.
    Laptop: Toshiba / Intel B960 2.20Ghz x2/ 4GB/ 320GB SataII/ Intel HD/ fc28.x86_64
    Tower: GigaByte (970A) / AMD FX 8320 3.5Ghz x8/ 16GB/ 9TB Sata III/ AMD 6770HD/ fc28.x86_64
    Bookshelf: Shuttle DS61 (H61)/ i3-3225 3.3Ghz x2/ 16GB/ 320GB Sata II/ Intel HD 4000/ fc27.x86_64
    Embedded: BeagleBone Blk / ARM AM3358 1 GHz x1/ 512MB/ 2GB eMMC/ PowerVR SGX530/ fc27.armv7hl

  10. #10
    Join Date
    Dec 2013
    Location
    United Kingdom
    Posts
    5,987

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    interesting, i'll try blacklisting that i2c_piix4 module on my AMD Athlon 5350 system as well. I've had that sp5100_tco conflict for some time now and just blacklisted the sp5100_tco module. it didn't seem to have any negative impact but I'll try your method and see what happens

  11. #11
    Join Date
    Dec 2013
    Location
    United Kingdom
    Posts
    5,987

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    yep sp5100_tco loads okay now so I've switched the blacklisted module in my boot command. the joys of linux kernel modules being configured by default for loading intel platform modules...

  12. #12
    Join Date
    Feb 2009
    Location
    Florida
    Posts
    629

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    Thanks for the confirmation.

    It seems this is the right area to be working, but I'm confused as to why this works.
    If you read:
    https://www.kernel.org/doc/Documenta...sses/i2c-piix4

    You will read that i2c_piix4 implements the PCI bus. So how does blacklisting it enable our AMD PCI devices to work? Is there a basic level of PCI in the kernel that better supports AMD?

    The other thing I haven't mentioned previously is that my system is still a legacy BIOS Boot machine. Maybe a UEFI boot will do better at properly initializing these PCIe devices/chips.
    Given my new BIOS is UEFI capable, I'm going to try that next. ( maybe UEFI uses different 32 bit protected code than legacy BIOS 16 bit real mode for initialization )

    Update: I've learned that UEFI uses a completely different set of 64 bit code to preform initialization than the 16 bit real mode legacy BIOS code
    Last edited by Kobuck; 11th March 2018 at 03:35 PM. Reason: Update info in post
    Laptop: Toshiba / Intel B960 2.20Ghz x2/ 4GB/ 320GB SataII/ Intel HD/ fc28.x86_64
    Tower: GigaByte (970A) / AMD FX 8320 3.5Ghz x8/ 16GB/ 9TB Sata III/ AMD 6770HD/ fc28.x86_64
    Bookshelf: Shuttle DS61 (H61)/ i3-3225 3.3Ghz x2/ 16GB/ 320GB Sata II/ Intel HD 4000/ fc27.x86_64
    Embedded: BeagleBone Blk / ARM AM3358 1 GHz x1/ 512MB/ 2GB eMMC/ PowerVR SGX530/ fc27.armv7hl

  13. #13
    Join Date
    Dec 2013
    Location
    United Kingdom
    Posts
    5,987

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    Quote Originally Posted by Kobuck
    Thanks for the confirmation.

    It seems this is the right area to be working, but I'm confused as to why this works.
    If you read:
    https://www.kernel.org/doc/Documenta...sses/i2c-piix4

    You will read that i2c_piix4 implements the PCI bus. So how does blacklisting it enable our AMD PCI devices to work? Is there a basic level of PCI in the kernel that better supports AMD?

    The other thing I haven't mentioned previously is that my system is still a legacy BIOS Boot machine. Maybe a UEFI boot will do better at properly initializing these PCIe devices/chips.
    Given my new BIOS is UEFI capable, I'm going to try that next. ( maybe UEFI uses different 32 bit protected code than legacy BIOS 16 bit real mode for initialization )
    my machine is UEFI compliant and has always run in UEFI mode, I'm no expert on this either but since you mentioned i2c_piix4 I tried finding out more about it. It seems to be specific to controllers in Intel chipsets anyway and for some reason gets loaded before the AMD sp5100_tco. sp5100_tco is listed as an SMBus driver if you execute lspci -v so there clearly isn't the need to have both running.

    The Athlon 5350 is a SoC design like the newer 2200G and 2400G Ryzen models

    Code:
    00:02.3 PCI bridge: Advanced Micro Devices, Inc. [AMD] Family 16h Processor Functions 5:1 (prog-if 00 [Normal decode])
            Flags: bus master, fast devsel, latency 0, IRQ 25
            Bus: primary=00, secondary=01, subordinate=01, sec-latency=0
            I/O behind bridge: 0000e000-0000efff [size=4K]
            Memory behind bridge: fea00000-feafffff [size=1M]
            Prefetchable memory behind bridge: 00000000d0800000-00000000d08fffff [size=1M]
            Capabilities: <access denied>
            Kernel driver in use: pcieport
            Kernel modules: shpchp
    ...
    
    0:14.0 SMBus: Advanced Micro Devices, Inc. [AMD] FCH SMBus Controller (rev 3a)
            Subsystem: Gigabyte Technology Co., Ltd FCH SMBus Controller
            Flags: 66MHz, medium devsel
            Kernel modules: i2c_piix4, sp5100_tco
    Code:
    06/03/2018 15:45    shpchp    Standard Hot Plug PCI Controller Driver version: 0.4
    Code:
    [antikythera@Aures ~]$ inxi -Fxms
    Resuming in non X mode: xrandr not found. For package install advice run: inxi --recommends
    System:    Host: Aures Kernel: 4.15.7-1-MANJARO x86_64 bits: 64 gcc: 7.3.0
               Desktop: KDE Plasma 5.12.2 (Qt 5.10.1) Distro: Manjaro Linux                                                     
    Machine:   Device: desktop Mobo: Gigabyte model: AM1M-S2H v: x.x serial: N/A UEFI: American Megatrends v: F2 date: 06/20/2014
    CPU:       Quad core AMD Athlon 5350 APU with Radeon R3 (-MCP-) arch: Jaguar rev.1 cache: 2048 KB
               flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm) bmips: 16368
               clock speeds: max: 2050 MHz 1: 1949 MHz 2: 1985 MHz 3: 1667 MHz 4: 1198 MHz
    Memory:    Using dmidecode: root required for dmidecode
    Graphics:  Card: Advanced Micro Devices [AMD/ATI] Kabini [Radeon HD 8400 / R3 Series] bus-ID: 00:01.0
               Display Server: N/A drivers: ati,radeon (unloaded: modesetting) tty size: 124x25
    Audio:     Card-1 Advanced Micro Devices [AMD] FCH Azalia Controller driver: snd_hda_intel bus-ID: 00:14.2
               Card-2 Advanced Micro Devices [AMD/ATI] Kabini HDMI/DP Audio driver: snd_hda_intel bus-ID: 00:01.1
               Sound: Advanced Linux Sound Architecture v: k4.15.7-1-MANJARO
    Network:   Card: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller
               driver: r8169 v: 2.3LK-NAPI port: e000 bus-ID: 01:00.0
               IF: enp1s0 state: up speed: 1000 Mbps duplex: full mac: 74:d4:35:a6:5a:1f
    Drives:    HDD Total Size: 1000.2GB (3.6% used)
               ID-1: /dev/sda model: ST1000DM003 size: 1000.2GB
    Partition: ID-1: / size: 92G used: 12G (13%) fs: btrfs dev: /dev/sda3
               ID-2: /boot size: 969M used: 90M (10%) fs: ext4 dev: /dev/sda2
               ID-3: /home size: 827G used: 11G (2%) fs: xfs dev: /dev/sda4
               ID-4: swap-1 size: 12.82GB used: 0.00GB (0%) fs: swap dev: /dev/sda5
    Sensors:   System Temperatures: cpu: 25.0C mobo: N/A gpu: 26.0
               Fan Speeds (in rpm): cpu: N/A
    Info:      Processes: 189 Uptime: 12 min Memory: 954.3/7415.3MB Init: systemd Gcc sys: 7.3.0
               Client: Shell (bash 4.4.191) inxi: 2.3.56
    Code:
    06/03/2018 14:48    Command line    BOOT_IMAGE=/vmlinuz-4.15-x86_64 root=UUID=d6eb2da9-53e8-41d7-82d1-50e439548fd9 rw rootflags=subvol=@ sysrq_always_enabled=1 modprobe.blacklist=i2c_piix4 radeon.dpm=1 resume=UUID=a636c1a3-2cdf-4cac-80aa-b1ea812dc44f

  14. #14
    Join Date
    Feb 2009
    Location
    Florida
    Posts
    629

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    The UEFI boot completes without hangups. No kernel line debug command required to successfully complete the boot. So there is an issue in the BIOS boot path that does not exist in the UEFI boot path.

    This solves the original purpose of this post.

    However:
    With my new board I still can't turn on IOMMU without killing the VIA chip which delivers USB 3.1. I consider this a different issue which came with my switch to the GA 970A MB. Its not really a new issue either. I can remember Bob having problems with this way back in 2015. This whole exercise does not leave me feeling "warm and snuggly" about the attention AMD issues are getting in the kernel. The AMD IOMMU has had longstanding problems with little sign of progress to resolution.

    Any way, since I can now boot the production releases without making mods, I think my system is now back on the rails
    Last edited by Kobuck; 8th March 2018 at 03:48 PM.
    Laptop: Toshiba / Intel B960 2.20Ghz x2/ 4GB/ 320GB SataII/ Intel HD/ fc28.x86_64
    Tower: GigaByte (970A) / AMD FX 8320 3.5Ghz x8/ 16GB/ 9TB Sata III/ AMD 6770HD/ fc28.x86_64
    Bookshelf: Shuttle DS61 (H61)/ i3-3225 3.3Ghz x2/ 16GB/ 320GB Sata II/ Intel HD 4000/ fc27.x86_64
    Embedded: BeagleBone Blk / ARM AM3358 1 GHz x1/ 512MB/ 2GB eMMC/ PowerVR SGX530/ fc27.armv7hl

  15. #15
    Join Date
    Feb 2009
    Location
    Florida
    Posts
    629

    Re: Problems booting 4.14.18.300.fc27.x86_64 on AMD processor

    After running on this for a while, I'm withdrawing my complaint about the IOMMU problem.

    The system defaults to using SWIOTLB (just like Intel systems) when no IOMMU is found. Whether the "fresh" install, UEFI initialization, a slightly faster CPU, dropping the IOMMU, or all of the above Gnome desktop just seems a little snappier.

    I'm a happy camper
    Laptop: Toshiba / Intel B960 2.20Ghz x2/ 4GB/ 320GB SataII/ Intel HD/ fc28.x86_64
    Tower: GigaByte (970A) / AMD FX 8320 3.5Ghz x8/ 16GB/ 9TB Sata III/ AMD 6770HD/ fc28.x86_64
    Bookshelf: Shuttle DS61 (H61)/ i3-3225 3.3Ghz x2/ 16GB/ 320GB Sata II/ Intel HD 4000/ fc27.x86_64
    Embedded: BeagleBone Blk / ARM AM3358 1 GHz x1/ 512MB/ 2GB eMMC/ PowerVR SGX530/ fc27.armv7hl

Page 1 of 2 1 2 LastLast

Similar Threads

  1. [SOLVED]
    Fedora 27: Kernel 4.14.14-300.fc27.x86_64 stops during booting.
    By zethan191 in forum Using Fedora
    Replies: 8
    Last Post: 18th February 2018, 05:15 PM
  2. Kernel crash - 4.14.6-300.fc27.x86_64
    By lsatenstein in forum Installation, Upgrades and Live Media
    Replies: 1
    Last Post: 20th December 2017, 12:28 PM
  3. kernel 4.13.0-0.rc3.git4.1.fc27.x86_64 is available
    By martyfelker in forum Rawhide
    Replies: 1
    Last Post: 7th August 2017, 09:16 PM
  4. 4.13.0-0.rc0.git3.1.fc27.x86_64 is available
    By martyfelker in forum Rawhide
    Replies: 2
    Last Post: 16th July 2017, 03:21 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •