[PATCH v2] arm64: kdump: simplify the reservation behaviour of crashkernel=,high

[Catalin Marinas]

On Fri, Feb 03, 2023 at 03:57:23PM +0800, Baoquan He wrote:
> On arm64, reservation for 'crashkernel=xM,high' is taken by searching for
> suitable memory region top down. If the 'xM' of crashkernel high memory
> is reserved from high memory successfully, it will try to reserve
> crashkernel low memory later accoringly. Otherwise, it will try to search
> low memory area for the 'xM' suitable region. Please see the details in
> Documentation/admin-guide/kernel-parameters.txt.
> 
> While we observed an unexpected case where a reserved region crosses the
> high and low meomry boundary. E.g on a system with 4G as low memory end,
> user added the kernel parameters like: 'crashkernel=512M,high', it could
> finally have [4G-126M, 4G+386M], [1G, 1G+128M] regions in running kernel.
> The crossing 4G boudary of crashkernel high region will bring issues:
> 
> 1) For crashkernel=x,high, if getting crashkernel high region across
> 4G boudary, then user will see two memory regions under 4G, and one
> memory region above 4G. The two crashkernel low memory regions are
> confusing.

Looking at your patch, I just realised that the 4G boundary between
'low' and 'high' reservations is not always true. On RPi4, that would be
1GB, the limit of ZONE_DMA. Are there user-space tools that rely on this
32-bit boundary? If they do, they'd get confused on RPi4, not sure they
have the notion of the actual ZONE_DMA that the kernel has enabled. If
we do want ,high to mean always 4G or higher, we'd need to change the
logic a bit so that the search_base starts from 4G rather than
CRASH_ADDR_LOW_MAX. We could leave the latter when ,high was not
specified.

> 2) If people explicityly specify "crashkernel=x,high crashkernel=y,low"
> and y <= 128M, when crashkernel high region crosses 4G boudary and the
> part below 4G of crashkernel high reservation is bigger than y, the
> expected crahskernel low reservation will be skipped. But the expected
> crashkernel high reservation is shrank and could not satisfy user space
> requirement.

I guess if the user passes both high and low, we should honour that and
ignore any y <= 128M checks.

-- 
Catalin