aspm-troubleshooting
This Repository provides an Overview with some Tips/Tricks into getting ASPM to Work in some/many/most Circumstances.
WARNING: I'm not responsible for any breakages of BIOS, Motherboards, CPUs, or whatever else. Proceed at your own Risk and do your own Research & Investigation.
There are several Steps involved in getting ASPM to work.
At the very least, make sure that:
- CPU PCIe Slots have ASPM Enabled (at least ASPM L1)
- PCH PCIe Slots have ASPM Enabled (at least ASPM L1)
- DMI Link has ASPM Enabled (at least ASPM L1 - some People Reported getting ASPM working by forcing L1 only, instead of L0sL1)
- CPU C-States have been enabled at least up to C6/C7/C7s (Haswell), higher on more recent CPUs
- CPU Package C-States (PC-States) have been enabled at least up to PC6/PC7 (Haswell), higher on more recent CPUs
Reccomendations / no definitive Answer:
- C1 auto-Demotion is Disabled
- C3 auto-Demotion is Disabled
- Pre-Wake is Disabled
- LakeTiny is Enabled
- C6 Latency is Long
- C7 Latency is Long
At the Moment I just put my Configuration in /etc/rc.local and configured a Basic Systemd Service for that.
IMPORTANT: if any other Commands in that file Fail, ASPM won't be Enabled/Activated/Configured.
# PowerSave Mode
powertop --auto-tune
A quick Test showed that the powersave Governor is really only suitable for lighly loaded Systems.
For Systems that are relatively loaded, that can be counter-productive, as with a low CPU Frequency there is not enough "Speed" or "Time" to complete the Tasks, thus the CPU never enters PC6/PC7.
# cpupower frequency-set --governor powersave
cpupower frequency-set --governor ondemand
# cpupower frequency-set --governor conservative
# Set the CPUIdle Governor
# "menu" (Default) or "ladder"
echo "menu" > /sys/devices/system/cpu/cpuidle/current_governor
If ASPM is not forcefully disabled by BIOS, FADT, uncooperative PCIe Devices, etc, then it is Enabled. However, one must still Activate ASPM to make use of it:
# Activate ASPM
# "powersave" or "powersupersave"
echo -n powersupersave > /sys/module/pcie_aspm/parameters/policy
One cause why ASPM is not working/doing anything is because the CPU is pinned to its maximum Frequency all the Time.
This can be because the CPU Frequency Governor is still set to Performance.
Refer to the Previous Section on how to Configure the CPU Frequency Governor.
If you have a high load on the System, your System will NOT enter ASPM to begin with.
100% Load on all Cores due to e.g. Kernel Compilation or other CPU-Intensive Processes will prevent the CPU from Sleeping.
If you run a Test with iperf3 or Similar Tools, the CPU will not go to sleep and your Package will be stuck to PC2/PC3 and never go into PC6/PC7.
In case you get a Message like this in dmesg:
ACPI FADT declares the system doesn't support PCIe ASPM, so disable
Then you need to manually Patch the ACPI Tables.
This can be done in InitramFS or by using a custom UEFI Tool (pre-Bootloader or that chainloads the GRUB Bootloader).
For a InitramFS Patch, refer to My other Repository.
For other Tools, refer to the Z8.re Website (see References) on how the same could be achieved via Other Methods.
In case you get a Message like this in dmesg:
pci 0000:03:00.0: disabling ASPM on pre-1.1 PCIe device. You can enable it with 'pcie_aspm=force'
Then, as the Message Outlines, you need to put pcie_aspm=force in /etc/default/grub.d/aspm.cnf:
GRUB_CMDLINE_LINUX="${GRUB_CMDLINE_LINUX} pcie_aspm=force"
GRUB_CMDLINE_LINUX_DEFAULT="${GRUB_CMDLINE_LINUX_DEFAULT} pcie_aspm=force"
In my Case I was getting this Message with an Intel i350-T4 4x1gbps NIC.
For instance you might want to "Kick Out" of the PCIe BUS a Device that Explicitely does NOT support ASPM.
IMPORTANT:
- Make sure that is the only device that is on the PCIe BUS and that no other Device shares the same Root Port !!
- Removing a PCIe Device and/or PCIe Root Port can actually make the situation WORSE (e.g. Package might go from PC3 to PC2 or even PC0 !!!)
For instance, on my ASUS P9D WS:
# Remove the ASPEED Bridge Device first of All
echo "1" > /sys/devices/pci0000:00/0000:00:1c.1/0000:04:00.0/remove
# Remove the Parent PCIe Root Port
echo "1" > /sys/devices/pci0000:00/0000:00:1c.1/remove
The Parent PCIe Root Port can be found using lspci -t.
Or alternatively, just look at what can be Kicked out of the PCIe Bus using find /sys -iwholename *04:00.0*remove and also Remove the Parent Device in that Tree (Parent "Folder").
For the ASUS P9D with a PCIe-to-PCI Bridge ASM1083, there is a Quirk in the Linux Kernel such that if any Vendor ID 0x1080 matches ASM1083/ASM1085, then ASPM is forcefully disabled.
This is how it could look like in dmesg:
pci 0000:04:00.0: Disabling ASPM L0s/L1
This trick will ONLY work if you do NOT have any PCI Device in those Legacy PCI Slots.
I care only about ASPM and NOT to use these Slots, so it's OK for me to do it :).
Note that while this "Fixes" the Issue about ASPM being disabled by "force", it does NOT guarantee that there are not other Issues preventing ASPM on the System !
Refer to My Separate Repository.
lspci -vvv | grep -A60 -B60 --color -Ei "ASPM\sDisabled|L1 Unlimited|L0 Unlimited"
For Instance for a Hailo-8L Accelerator it looks like this:
08:00.0 Co-processor: Hailo Technologies Ltd. Hailo-8 AI Processor (rev 01)
Subsystem: Hailo Technologies Ltd. Hailo-8 AI Processor
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
Latency: 0, Cache Line Size: 64 bytes
Interrupt: pin A routed to IRQ 42
IOMMU group: 16
Region 0: Memory at f0004000 (64-bit, prefetchable) [size=16K]
Region 2: Memory at f0008000 (64-bit, prefetchable) [size=4K]
Region 4: Memory at f0000000 (64-bit, prefetchable) [size=16K]
Capabilities: [80] Express (v2) Endpoint, MSI 00
DevCap: MaxPayload 256 bytes, PhantFunc 0, Latency L0s <64ns, L1 unlimited
ExtTag+ AttnBtn- AttnInd- PwrInd- RBE+ FLReset+ SlotPowerLimit 25W
DevCtl: CorrErr+ NonFatalErr+ FatalErr+ UnsupReq+
RlxdOrd- ExtTag+ PhantFunc- AuxPwr- NoSnoop+ FLReset-
MaxPayload 128 bytes, MaxReadReq 512 bytes
DevSta: CorrErr- NonFatalErr- FatalErr- UnsupReq- AuxPwr- TransPend-
LnkCap: Port #0, Speed 8GT/s, Width x4, ASPM L0s L1, Exit Latency L0s <1us, L1 <2us
ClockPM- Surprise- LLActRep- BwNot- ASPMOptComp+
LnkCtl: ASPM L1 Enabled; RCB 64 bytes, Disabled- CommClk+
ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt-
LnkSta: Speed 5GT/s (downgraded), Width x2 (downgraded)
TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
DevCap2: Completion Timeout: Not Supported, TimeoutDis+ NROPrPrP- LTR+
10BitTagComp- 10BitTagReq- OBFF Not Supported, ExtFmt+ EETLPPrefix-
EmergencyPowerReduction Not Supported, EmergencyPowerReductionInit-
FRS- TPHComp- ExtTPHComp-
AtomicOpsCap: 32bit- 64bit- 128bitCAS-
DevCtl2: Completion Timeout: 50us to 50ms, TimeoutDis- LTR+ 10BitTagReq- OBFF Disabled,
AtomicOpsCtl: ReqEn-
LnkCap2: Supported Link Speeds: 2.5-8GT/s, Crosslink- Retimer- 2Retimers- DRS-
LnkCtl2: Target Link Speed: 8GT/s, EnterCompliance- SpeedDis-
Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-
Compliance Preset/De-emphasis: -6dB de-emphasis, 0dB preshoot
LnkSta2: Current De-emphasis Level: -6dB, EqualizationComplete- EqualizationPhase1-
EqualizationPhase2- EqualizationPhase3- LinkEqualizationRequest-
Retimer- 2Retimers- CrosslinkRes: unsupported
Capabilities: [e0] MSI: Enable+ Count=1/1 Maskable- 64bit+
Address: 00000000fee003d8 Data: 0000
Capabilities: [f8] Power Management version 3
Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot+,D3cold-)
Status: D0 NoSoftRst+ PME-Enable- DSel=0 DScale=1 PME-
Capabilities: [100 v1] Vendor Specific Information: ID=1556 Rev=1 Len=008 <?>
Capabilities: [108 v1] Latency Tolerance Reporting
Max snoop latency: 71680ns
Max no snoop latency: 71680ns
Capabilities: [110 v1] L1 PM Substates
L1SubCap: PCI-PM_L1.2+ PCI-PM_L1.1+ ASPM_L1.2+ ASPM_L1.1+ L1_PM_Substates+
PortCommonModeRestoreTime=10us PortTPowerOnTime=10us
L1SubCtl1: PCI-PM_L1.2- PCI-PM_L1.1- ASPM_L1.2- ASPM_L1.1-
T_CommonMode=0us LTR1.2_Threshold=0ns
L1SubCtl2: T_PwrOn=10us
Capabilities: [128 v1] Alternative Routing-ID Interpretation (ARI)
ARICap: MFVC- ACS-, Next Function: 0
ARICtl: MFVC- ACS-, Function Group: 0
Capabilities: [200 v2] Advanced Error Reporting
UESta: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol-
UEMsk: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol-
UESvrt: DLP+ SDES- TLP- FCP+ CmpltTO- CmpltAbrt- UnxCmplt- RxOF+ MalfTLP+ ECRC- UnsupReq- ACSViol-
CESta: RxErr- BadTLP- BadDLLP- Rollover- Timeout- AdvNonFatalErr-
CEMsk: RxErr- BadTLP- BadDLLP- Rollover- Timeout- AdvNonFatalErr+
AERCap: First Error Pointer: 00, ECRCGenCap+ ECRCGenEn- ECRCChkCap+ ECRCChkEn-
MultHdrRecCap- MultHdrRecEn- TLPPfxPres- HdrLogCap-
HeaderLog: 00000000 00000000 00000000 00000000
Capabilities: [300 v1] Secondary PCI Express
LnkCtl3: LnkEquIntrruptEn- PerformEqu-
LaneErrStat: 0
Kernel driver in use: hailo
Kernel modules: hailo_pci
Note how the Link Capabilities report that LnkCtl: ASPM L1 Enabled yet the Device Capabilities report that DevCap: MaxPayload 256 bytes, PhantFunc 0, Latency L0s <64ns, L1 unlimited.
And to top it Off dmesg states that the Hailo-8L forces the System to Disable L0s, YET it should support L1, except that it has an Infinite Exit Latency from L1. Meaning it will never be able to Recover.
[ 21.925841] hailo 0000:01:00.0: Disabling ASPM L0s
[ 21.925919] hailo 0000:01:00.0: Successfully disabled ASPM L0s
[ 22.060060] hailo 0000:02:00.0: Disabling ASPM L0s
[ 22.060068] hailo 0000:02:00.0: Successfully disabled ASPM L0s
[ 22.197020] hailo 0000:08:00.0: Disabling ASPM L0s
[ 22.197088] hailo 0000:08:00.0: Successfully disabled ASPM L0s
I can try to Patch it using the Customized Python Script, so that I can directly feed in Command Line Arguments.
To find the Root Port, either use lspci -t or, maybe easier, jusdt look at find /sys -iwholename *08:00.0* which will return something like this:
/sys/kernel/iommu_groups/16/devices/0000:08:00.0
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/uevent
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/resource0_wc
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power_state
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/broken_parity_status
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/subsystem_device
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/dma_mask_bits
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/vendor
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/iommu_group
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/local_cpus
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/firmware_node
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/runtime_active_time
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/wakeup
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/wakeup_count
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/wakeup_abort_count
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/wakeup_expire_count
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/runtime_active_kids
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/wakeup_total_time_ms
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/wakeup_active_count
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/runtime_usage
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/runtime_status
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/autosuspend_delay_ms
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/async
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/runtime_suspended_time
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/wakeup_max_time_ms
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/wakeup_active
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/runtime_enabled
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/control
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/power/wakeup_last_time_ms
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/class
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/reset
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/numa_node
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/resource
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/rescan
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/max_link_width
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/resource2_wc
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/msi_bus
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/device
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/aer_dev_nonfatal
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/current_link_width
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/driver
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/max_link_speed
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/resource4
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/local_cpulist
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/driver_override
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/subsystem
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/d3cold_allowed
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/irq
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/revision
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/resource2
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/current_link_speed
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/reset_method
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/resource4_wc
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/aer_dev_correctable
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/consistent_dma_mask_bits
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/resource0
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/config
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/ari_enabled
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/msi_irqs
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/msi_irqs/42
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/remove
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/iommu
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/aer_dev_fatal
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/enable
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/link
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/link/l1_aspm
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/modalias
/sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/subsystem_vendor
/sys/devices/virtual/iommu/dmar1/devices/0000:08:00.0
/sys/bus/pci/devices/0000:08:00.0
/sys/bus/pci/drivers/hailo/0000:08:00.0
Thus the Root Port can be Found looking at e.g. the Line /sys/devices/pci0000:00/0000:00:1c.4/0000:08:00.0/iommu so Device 08:00.0 corresponds to Root Port 00:1c.4.
Similarly for the Other Devices:
find /sys -iwholename *02:00.0*->/sys/devices/pci0000:00/0000:00:01.1/0000:02:00.0/iommu-> Device02:00.0and Root Port00:01.1find /sys -iwholename *01:00.0*->/sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/iommu-> Device01:00.0and Root Port00:01.0
So it can be set with:
python3 aspm.py --device "01:00.0" --root "00:01.0"
python3 aspm.py --device "02:00.0" --root "00:01.1"
python3 aspm.py --device "08:00.0" --root "00:1c.4"
Unfortunately, that still does NOT make ASPM work :(.
To patch only the Root Port one can do:
python3 aspm.py --root "00:1c.0" --setting "ASPM_L1_AND_L0s"
python3 aspm.py --root "00:1c.1" --setting "ASPM_L1_AND_L0s"
The Device works in PCH-connected PCIe Slot (as confirmed by other People on the Internet). However, unless you have a DMI 3.0 x4 Link and/or only want e.g. 2.5-5.0 gbps (10 gbps could also work, if SATA bandwidth is Low) NIC Bandwidth,, the Bandwidth will be severly restricted and shared with SATA Controller, other Ethernet Controllers, etc.
On CPU-connected PCIe Slots however, it seems that the System just doesn't want to go into ASPM, no matter what one does. Even though lspci still statees (exactly like on PCH-connected PCIe Slots) that the Mellanox ConnectX-4 NIC does Support ASPM L1.
For very Recent Intel 12th/13th Generation i5/i7 CPUs, the Issue might be related to PCIe "Multi-VC" (Multiple Virtual Channels), which need to be disabled in the BIOS.
See for Instance:
- Intel Community
- Matt Gadient Blog (According to etnicor on February 15, 2024)
Very often that Feature is NOT exposed in the BIOS, so one needs to build a Custom Modded BIOS and flash that onto the Board.
Alternatively, one can Decompile an Existing BIOS using UEFITool + IFRExtractor, then use a Tool like setup_var.efi or possibly uvt / UefiVarTool to set such Setting from a UEFI Shell.
Unfortunately, neither Option worked for me, presumably because I don't have Multi-VC to beging with on older Platforms such as Intel Xeon E3 v3 and possibly also v5/v6 (I didn't test those yet).
Some Users including me attempted to play with mlxconfig (see reconfigure_mellanox_connectx-4.sh) but no Success was reported thus far.
See for Instance:
Both the Cause and the Solution to this Issue still remains unknown.
The Intel X710-DA2 NIC seems to support ASPM out of the Box, both in CPU-connected PCIe Slots, as well as in PCH-connected PCIe Slots.
I use a modified version of the Excellent ASPM Script with a simple Option to feed-in the Arguments via Command Line (argparse).
Note: based on some Reddit User Comments, the Reason why "Unkown Register" Error shows up when using a Shell Script called aspm-enabler is because the bc Package was not installed.
This Example Supposes the Following Folder Structure and shows the Process of Analysing a Supermicro X10SLM+-F BIOS Version 3.3, although of course other Folder Structures can be used of course:
.
├── ifrextractor
│ ├── ifrextractor
│ ├── Section_Freeform_subtype_GUID_MiniSetupResourceSection_Setup_Setup_body.bin
│ ├── Section_Freeform_subtype_GUID_MiniSetupResourceSection_Setup_Setup_body.bin.0.0.en-US.ifr.txt
│ ├── Section_Freeform_subtype_GUID_MiniSetupResourceSection_Setup_Setup_body.bin.1.0.en-US.ifr.txt
│ ├── Section_Freeform_subtype_GUID_MiniSetupResourceSection_Setup_Setup_body.bin.2.0.en-US.ifr.txt
│ ├── Section_Freeform_subtype_GUID_MiniSetupResourceSection_Setup_Setup_body.bin.3.0.en-US.ifr.txt
│ ├── Section_Freeform_subtype_GUID_MiniSetupResourceSection_Setup_Setup_body.bin.4.0.en-US.ifr.txt
│ └── Section_Freeform_subtype_GUID_MiniSetupResourceSection_Setup_Setup_body.bin.5.0.en-US.ifr.txt
├── UEFITool
│ ├── Section_Freeform_subtype_GUID_MiniSetupResourceSection_Setup_Setup_body.bin
│ ├── uefitool
│ └── x10slh0.327
└── x10slh0.327
The easiest (by far) Way to Patch BIOS Settings is as the following Section Describes.
Download your BIOS Firmware from the Manufacturer, making sure to select the correct Motherboard Model and BIOS Version
- Download the Tool from the Official Repository
- Open the Tool:
./UEFITool/uefitool - "File" -> "Open Image File"
- File Type: "All Files"
- Select your BIOS Image
- Click "Open"
- "Action" -> "Search"
- Go to "Text" Tab
- Enter "Setup"
- Click "OK"
- Click on one of the Results that States
Unicode text "Setup" found in Setup/.../MiniSetupResourceSection at header-offset 2E3EBh" - Select the Item
MiniSetupResourceSectionwhich should be located underSetup/Compressed Section/MiniSetupResourceSection - Right Click -> Extract Body -> Save this File in the same Folder as
uefitool
- Download the Tool from the Official Repository
- Suggest to copy the File saved from
uefitoolinto the ifrextractor Folder - Switch Folder to the Tool Executable:
cd ./ifrextractor - Run
ifrextractorto Extract the BIOS Data as Text:./ifrextractor Section_Freeform_subtype_GUID_MiniSetupResourceSection_Setup_Setup_body.bin - A lot of
.txtFiles will be Generated, each of which (0.0,1.0, ...) should correspond to a different Tab in the BIOS Screen
Each BIOS Setting is represented by a list of Parameters, e.g.:
OneOf Prompt: "PCH SLOT4 PCI-E 2.0 X2/4 (IN X8) - ASPM", Help: "Set the ASPM Level: Force L0s - Force all links to L0s State : AUTO - BIOS auto configure : DISABLE - Disables ASPM", QuestionFlags: 0x10, QuestionId: 0x118, VarStoreId: 0x1, VarOffset: 0xBA, Flags: 0x10, Size: 8, Min: 0x0, Max: 0x0, Step: 0x0
OneOfOption Option: "Disabled" Value: 0
OneOfOption Option: "L0s" Value: 1
OneOfOption Option: "L1" Value: 2
OneOfOption Option: "L0sL1" Value: 3
OneOfOption Option: "Auto" Value: 4, Default, MfgDefault
End
The Setting itself is defined by:
VarOffset(in Example:0xBA): in which "Position" or "Register" it's locatedVarStoreId(in Example:0x1): in which "Register Group" it's locatedValue: the possible Values that this BIOS Setting can take, in Example:0x0forDisabled0x1forL0s0x2forL10x3forL0sL10x4forAuto
USUALLY (for MOST Settings):
<VarStoreId>is0x1<Name>isSetup
However, to make Sure, since <Name> is usually Defined in the First File, so you can Find <Name> by doing:
grep -ri "VarStoreId: <VarStoreId>," Section_Freeform_*.0.0* | sed -E "s|.*, Name: \"([a-zA-Z0-9]*?)\"|\1|"
To search for settings you might be Interested in, you can e.g. do to pre-generate a list of Fields to be manually adressed:
grep -rih "<Your_Search_Criteria>" *.txt | grep "VarOffset" | sed -E "s|.*?VarStoreId: 0x([0-9A-F]+?), VarOffset: 0x([0-9A-F]+?),.*|setup_var.efi Setup(0x\1)\:0x\2=<To_Be_Manually_Filled_In>|" | uniq
Then of course you'll have to manually scan the Files by the correct VarOffset to determine which Value you want to use.
The UEFI Boot Drive might NOT be at all required, if your system has already a built-in UEFI Shell.
Furthermore, many of the different File Copies / Locations described below might NOT be required, but this was the only way I could get it to boot on an ASUS P9D WS Motherboard with the "Standard" EFI Shell Executable Layout.
- Format the Drive as FAT32
- Get a
shellx64.efiUEFI Shell Executable- (Old) TianoCore
- (Verify Build Process / Artifacts for Safety) pbatard Repository
- Create a
./efi/bootFolder Structure in your newly formatted FAT32 Drive - Copy
shellx64.efito:./efi/shellx64.efi./efi/bootx64.efi./efi/boot/shellx64.efi./efi/boot/bootx64.efi./shellx64.efi./bootx64.efi
Get setup_var.efi from the Official Repository.
Alternatively you might consider getting uvt / UefiVarTool from their Official Repository (untested).
Copy it to the UEFI Boot Drive or any other USB Flash Drive in case your System has a built-in UEFI Shell.
For easier Use, Track-Record of your Modifications, ability to share with the Community, etc, you might consider writing an EFI Shell Script e.g. patch.nsh to automate all Operations from UEFI Shell:
You can refer to some Actualy Files (some totally/partially NOT Tested !) in the patching Subfolder.
# Tested with BIOS Version <BIOS_Version> on <Manufactorer> <Motherboard_Model>
# Disable printing of Commands
@echo -off
#############################################
# <Section_Name> #
#############################################
# <Menu_Title> - <Help_Text>
# <Value_0>: <Meaning_0> / <Value_1>: <Meaning_1> / ...
setup_var.efi <Name>(<VarStoreId>):<VarOffset>=<Value>
# <Menu_Title> - <Help_Text>
# <Value_0>: <Meaning_0> / <Value_1>: <Meaning_1> / ...
setup_var.efi <Name>(<VarStoreId>):<VarOffset>=<Value>
# ...
# ...
# ...
Once the Preparations are Done:
- Boot the System
- Either Choose "Select Boot Device" (Typically F10 or F11 on Supermicro Motherboards) or "Enter BIOS Setup" (Typically F2/DELETE)
- Choose your USB FlashDrive
- Enter the FlashDrive Root Folder. This is typically done by entering
fs0: - Change Folder where you put your
patch.nshEFI Script and thesetup_var.efiExecutable using e.g.cd TOPDIR\SUBDIR\MYDIR - Execute the Script by issueing
patch.nsh - Reboot the System by issueing
reset
WORK IN PROGRESS - CURRENTLY THIS DOES NOT WORK
It's also possible to use ufk to get the Required Information.
Extract Information using ufk and Export into JSON Format:
utk ./x10slh0.327 json > x10slh0.327.json
Then extract the GUID like e.g.
GUID=$(cat x10slh0.327.json | jq '.. | objects | select(.Name=="Setup"?) | .GUID.GUID')
Locate the File to be Analysed:
FILENAME=$(find x10slh0.327.utk.extract/ -iwholename *EC87D643-EBA4-4BB5-A1E5-3F3E36B20DA9/Setup.bin)
Extract the Information with ifrextractor:
ifrextractor
cpupowerturbostat
vmstatprocinfo
- https://docs.redhat.com/en/documentation/red_hat_enterprise_linux/7/html/performance_tuning_guide/sect-red_hat_enterprise_linux-performance_tuning_guide-cpu-monitoring_and_diagnosing_performance_problems#sect-Red_Hat_Enterprise_Linux-Performance_Tuning_Guide-Monitoring_and_diagnosing_performance_problems-numastat
- https://docs.redhat.com/en/documentation/red_hat_enterprise_linux/8/html-single/monitoring_and_managing_system_status_and_performance/index#overview-of-performance-monitoring-options_monitoring-and-managing-system-status-and-performance
- https://www.site24x7.com/learn/linux/context-switching.html
- https://www.suse.com/support/kb/doc/?id=000016914