On 8/21/19 9:53 PM, Kairui Song wrote:
> Since commit c7753208a94c ("x86, swiotlb: Add memory encryption support"),
> SWIOTLB will be enabled even if there is less than 4G of memory when SME
> is active, to support DMA of devices that not support address with the
> encrypt bit.
>
> And commit aba2d9a6385a ("iommu/amd: Do not disable SWIOTLB if SME is
> active") make the kernel keep SWIOTLB enabled even if there is an IOMMU.
>
> Then commit d7b417fa08d1 ("x86/mm: Add DMA support for SEV memory
> encryption") will always force SWIOTLB to be enabled when SEV is active
> in all cases.
>
> Now, when either SME or SEV is active, SWIOTLB will be force enabled,
> and this is also true for kdump kernel. As a result kdump kernel will
> run out of already scarce pre-reserved memory easily.
>
> So when SME/SEV is active, reserve extra memory for SWIOTLB to ensure
> kdump kernel have enough memory, except when "crashkernel=size[KMG],high"
> is specified or any offset is used. As for the high reservation case, an
> extra low memory region will always be reserved and that is enough for
> SWIOTLB. Else if the offset format is used, user should be fully aware
> of any possible kdump kernel memory requirement and have to organize the
> memory usage carefully.
>
> Signed-off-by: Kairui Song <[email protected]>
> ---
> arch/x86/kernel/setup.c | 26 +++++++++++++++++++++++---
> 1 file changed, 23 insertions(+), 3 deletions(-)
>
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index bbe35bf879f5..ed91fa9d9f6e 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -528,7 +528,7 @@ static int __init reserve_crashkernel_low(void)
>
> static void __init reserve_crashkernel(void)
> {
> - unsigned long long crash_size, crash_base, total_mem;
> + unsigned long long crash_size, crash_base, total_mem, mem_enc_req;
> bool high = false;
> int ret;
>
> @@ -550,6 +550,17 @@ static void __init reserve_crashkernel(void)
> return;
> }
>
> + /*
> + * When SME/SEV is active, it will always required an extra SWIOTLB
> + * region.
> + */
> + if (sme_active() || sev_active()) {
You can use mem_encrypt_active() here in place of the two checks.
> + mem_enc_req = ALIGN(swiotlb_size_or_default(), SZ_1M);
> + pr_info("Memory encryption is active, crashkernel needs %ldMB
> extra memory\n",
> + (unsigned long)(mem_enc_req >> 20));
There is a point below where you zero out this value, so should this
be issued later only if mem_enc_req is non-zero?
Also, looks like one too many tabs.
> + } else
Since you used braces on the if path, you need braces on the else path.
Thanks,
Tom
> + mem_enc_req = 0;
> +
> /* 0 means: find the address automatically */
> if (!crash_base) {
> /*
> @@ -563,11 +574,19 @@ static void __init reserve_crashkernel(void)
> if (!high)
> crash_base = memblock_find_in_range(CRASH_ALIGN,
> CRASH_ADDR_LOW_MAX,
> - crash_size, CRASH_ALIGN);
> - if (!crash_base)
> + crash_size + mem_enc_req,
> + CRASH_ALIGN);
> + /*
> + * For high reservation, an extra low memory for SWIOTLB will
> + * always be reserved later, so no need to reserve extra
> + * memory for memory encryption case here.
> + */
> + if (!crash_base) {
> + mem_enc_req = 0;
> crash_base = memblock_find_in_range(CRASH_ALIGN,
> CRASH_ADDR_HIGH_MAX,
> crash_size, CRASH_ALIGN);
> + }
> if (!crash_base) {
> pr_info("crashkernel reservation failed - No suitable
> area found.\n");
> return;
> @@ -583,6 +602,7 @@ static void __init reserve_crashkernel(void)
> return;
> }
> }
> + crash_size += mem_enc_req;
> ret = memblock_reserve(crash_base, crash_size);
> if (ret) {
> pr_err("%s: Error reserving crashkernel memblock.\n", __func__);
>
