In order to see the SPD (detailed memory information) data, the user currently has to manually load the needed kernel driver. Which driver to load depends on the memory type. Depending on the driver user, the devices may even have to be instantiated manually and this is a non-trivial multi-step task. Plus you need to be root to do it.
I would like to attempt to automatize all this at least in the most common and simple cases like Intel x86 desktop. The idea would be to figure out the memory type and the I2C address of the SPD EEPROMs based on DMI data. If the DMI data is of good quality then it should be possible to automatically figure out which driver to use and to instantiate the devices at boot time.
If this works then running "decode-dimms" (or any other equivalent tool) should just work after boot without any preparatory work, for all users.
I plan to start implementing this for DDR3 memory and the i2c-i801 SMBus controller driver because that's what I have on my workstation. If it works, doing the same for DDR4 shouldn't be too difficult. Once this works for the i2c-i801, it should be pretty trivial to do the same with other SMBus controller drivers, for example i2c-piix4.
Excluded from the scope are large server systems with multiple SMBus controllers or multiplexed SMBus. Also excluded are OF/DT systems as I would expect SPD EEPROMs to be declared in the device tree so they would already be instantiated without further effort.
This project is part of:
Hack Week 18
Activity
Comments
-
over 4 years ago by jdelvare | Reply
For the record, this hack week project was successful. Relevant upstream kernel commits:
commit 9e0afe3910ff7e5493c5d8ebe3b499994b5e0272 Author: Jean Delvare Date: Tue Dec 3 11:20:37 2019 +0100 firmware: dmi: Remember the memory type commit 7c2378800cf7ac87e2663afa7f39d102871f0c28 Author: Jean Delvare Date: Tue Dec 3 11:20:37 2019 +0100 firmware: dmi: Add dmi_memdev_handle commit 5ace60859e84113b7a185c117fbf2c429d381b59 Author: Jean Delvare Date: Mon Mar 16 11:22:24 2020 +0100 i2c: smbus: Add a way to instantiate SPD EEPROMs automatically commit 01590f361e94a01e9b9868fa81d4079d255c681f Author: Jean Delvare Date: Mon Mar 16 11:24:48 2020 +0100 i2c: i801: Instantiate SPD EEPROMs automatically
So the feature is supported since kernel v5.8. Next step is to extend support to systems with 5-8 memory slots (still on a single SMBus segment), and to add support to other SMBus controller drivers.
Similar Projects
early stage kdump support by mbrugger
Project Description
When we experience a early boot crash, we are not able to analyze the kernel dump, as user-space wasn't able to load the crash system. The idea is to make the crash system compiled into the host kernel (think of initramfs) so that we can create a kernel dump really early in the boot process.
Goal for the Hackweeks
- Investigate if this is possible and the implications it would have (done in HW21)
- Hack up a PoC (done in HW22 and HW23)
- Prepare RFC series (giving it's only one week, we are entering wishful thinking territory here).
update HW23
- I was able to include the crash kernel into the kernel Image.
- I'll need to find a way to load that from
init/main.c:start_kernel()
probably afterkcsan_init()
- I workaround for a smoke test was to hack
kexec_file_load()
systemcall which has two problems:- My initramfs in the porduction kernel does not have a new enough kexec version, that's not a blocker but where the week ended
- As the crash kernel is part of init.data it will be already stale once I can call
kexec_file_load()
from user-space.
The solution is probably to rewrite the POC so that the invocation can be done from init.text (that's my theory) but I'm not sure if I can reuse the kexec infrastructure in the kernel from there, which I rely on heavily.
update HW24
- Day1
- rebased on v6.12 with no problems others then me breaking the config
- setting up a new compilation and qemu/virtme env
- getting desperate as nothing works that used to work
- Day 2
- getting to call the invocation of loading the early kernel from
__init
afterkcsan_init()
- getting to call the invocation of loading the early kernel from
Day 3
- fix problem of memdup not being able to alloc so much memory... use 64K page sizes for now
- code refactoring
- I'm now able to load the crash kernel
- When using virtme I can boot into the crash kernel, also it doesn't boot completely (major milestone!), crash in
elfcorehdr_read_notes()
Day 4
- crash systems crashes (no pun intended) in
copy_old_mempage()
link; will need to understand elfcorehdr... - call path
vmcore_init() -> parse_crash_elf_headers() -> elfcorehdr_read() -> read_from_oldmem() -> copy_oldmem_page() -> copy_to_iter()
- crash systems crashes (no pun intended) in
Day 5
- hacking
arch/arm64/kernel/crash_dump.c:copy_old_mempage()
to see if crash system really starts. It does. - fun fact: retested with more reserved memory and with UEFI FW, host kernel crashes in init but directly starts the crash kernel, so it works (somehow) \o/
- hacking
TODOs
- fix elfcorehdr so that we actually can make use of all this...
- test where in the boot
__init()
chain we can/should callkexec_early_dump()