Re: [PATCH 0/5] cxl: Initialization and shutdown fixes

[Alejandro Lucero Palau <alucerop@amd.com>]

Hi Dan,


I think this is the same issue one of the patches in type2 support tries 
to deal with:


https://lore.kernel.org/linux-cxl/20240907081836.5801-1-alejandro.lucero-palau@amd.com/T/#m9357a559c1a3cc7869ecce44a1801d51518d106e


If this fixes that situation, I guess I can drop that one from v4 which 
is ready to be sent.


The other problem I try to fix in that patch, the endpoint not being 
there when that code tries to use it, it is likely not needed either, 
although I have a trivial fix for it now instead of that ugly loop with 
delays. The solution is to add PROBE_FORCE_SYNCHRONOUS as probe_type for 
the cxl_mem_driver which implies the device_add will only return when 
the device is really created. Maybe that is worth it for other potential 
situations suffering the delayed creation.


On 10/11/24 06:33, Dan Williams wrote:
> Gregory's modest proposal to fix CXL cxl_mem_probe() failures due to
> delayed arrival of the CXL "root" infrastructure [1] prompted questions
> of how the existing mechanism for retrying cxl_mem_probe() could be
> failing.
>
> The critical missing piece in the debug was that Gregory's setup had
> almost all CXL modules built-in to the kernel.
>
> On the way to that discovery several other bugs and init-order corner
> cases were discovered.
>
> The main fix is to make sure the drivers/cxl/Makefile object order
> supports root CXL ports being fully initialized upon cxl_acpi_probe()
> exit. The modular case has some similar potential holes that are fixed
> with MODULE_SOFTDEP() and other fix ups. Finally, an attempt to update
> cxl_test to reproduce the original report resulted in the discovery of a
> separate long standing use after free bug in cxl_region_detach().
>
> [1]: http://lore.kernel.org/20241004212504.1246-1-gourry@gourry.net
>
> ---
>
> Dan Williams (5):
>        cxl/port: Fix CXL port initialization order when the subsystem is built-in
>        cxl/port: Fix cxl_bus_rescan() vs bus_rescan_devices()
>        cxl/acpi: Ensure ports ready at cxl_acpi_probe() return
>        cxl/port: Fix use-after-free, permit out-of-order decoder shutdown
>        cxl/test: Improve init-order fidelity relative to real-world systems
>
>
>   drivers/base/core.c          |   35 +++++++
>   drivers/cxl/Kconfig          |    1
>   drivers/cxl/Makefile         |   12 +--
>   drivers/cxl/acpi.c           |    7 +
>   drivers/cxl/core/hdm.c       |   50 +++++++++--
>   drivers/cxl/core/port.c      |   13 ++-
>   drivers/cxl/core/region.c    |   48 +++-------
>   drivers/cxl/cxl.h            |    3 -
>   include/linux/device.h       |    3 +
>   tools/testing/cxl/test/cxl.c |  200 +++++++++++++++++++++++-------------------
>   tools/testing/cxl/test/mem.c |    1
>   11 files changed, 228 insertions(+), 145 deletions(-)
>
> base-commit: 8cf0b93919e13d1e8d4466eb4080a4c4d9d66d7b
>