Re: CXL Namespaces of ACPI disappearing in Qemu demo

[Jonathan Cameron]

On Thu, 10 Aug 2023 17:30:43 +0800
Yuquan Wang <wangyuquan1236@phytium.com.cn> wrote:

> Hi, Jonathan
> 
> When I tested the CXL topology in qemu, I found the linux kernel could find 
> and match the ACPI0016&ACPI0017 in relevant drivers, however, I could not 
> find these information in dumped ACPI tables (like DSDT). 
> 
> Hence, I guess there is some problem in my method, maybe you have some 
> suggestions if you have some time to look at this email.
> 
> My environment:
> qemu: Jonathan_qemu/cxl-2023-08-07
> Linux: v6.3.0
> ACPI Disassembler: iasl/version 20230628
> 
> My qemu script is the same as the first example in 
> qemu/docs/system/devices/cxl.rst, 
> as below:
>   qemu-system-x86_64 -M q35,cxl=on -m 4G,maxmem=8G,slots=8 -smp 4 \
>   ...
>   -object 
> memory-backend-file,id=cxl-mem1,share=on,mem-path=/tmp/cxltest.raw,size=256M \
>   -object 
> memory-backend-file,id=cxl-lsa1,share=on,mem-path=/tmp/lsa.raw,size=256M \
>   -device pxb-cxl,bus_nr=12,bus=pcie.0,id=cxl.1 \
>   -device cxl-rp,port=0,bus=cxl.1,id=root_port13,chassis=0,slot=2 \
>   -device 
> cxl-type3,bus=root_port13,persistent-memdev=cxl-mem1,lsa=cxl-lsa1,id=cxl-pmem0
>  \
>   -M cxl-fmw.0.targets.0=cxl.1,cxl-fmw.0.size=4G

Hi Yuquan,

Ah. I may have messed up dropping a patch set that was factoring out all the 
PXB stuff to SSDT
in my latest tree.

I dropped it because it was nasty to rebase and didn’t seem likely to go 
upstream particularly
quickly as we pushed the decision on it back behind a load of other PCI changes.

I tested the latest tree heavily on arm, but was lazy on the x86 front :(
as I forgot that that change might have broken things (and all the new stuff 
was architecture
agnostic).

Right now for x86 tests with my tree I'm getting a freeze on booting so some 
debug needed.
Even better, on upstream qemu with near vanilla kernel I get a segfault instead 
in pretty
much the same place.
This is going to be a 'fun' day I guess. 

Jonathan



> 
> 
> Many thanks
> Yuquan