x86/xen: attempt to inflate the memory balloon on PVH

When running as PVH or HVM Linux will use holes in the memory map as scratch
space to map grants, foreign domain pages and possibly miscellaneous other
stuff.  However the usage of such memory map holes for Xen purposes can be
problematic.  The request of holesby Xen happen quite early in the kernel boot
process (grant table setup already uses scratch map space), and it's possible
that by then not all devices have reclaimed their MMIO space.  It's not
unlikely for chunks of Xen scratch map space to end up using PCI bridge MMIO
window memory, which (as expected) causes quite a lot of issues in the system.

At least for PVH dom0 we have the possibility of using regions marked as
UNUSABLE in the e820 memory map.  Either if the region is UNUSABLE in the
native memory map, or it has been converted into UNUSABLE in order to hide RAM
regions from dom0, the second stage translation page-tables can populate those
areas without issues.

PV already has this kind of logic, where the balloon driver is inflated at
boot.  Re-use the current logic in order to also inflate it when running as
PVH.  onvert UNUSABLE regions up to the ratio specified in EXTRA_MEM_RATIO to
RAM, while reserving them using xen_add_extra_mem() (which is also moved so
it's no longer tied to CONFIG_PV).

[jgross: fixed build for CONFIG_PVH without CONFIG_XEN_PVH]

Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/20240220174341.56131-1-roger.pau@citrix.com
Signed-off-by: Juergen Gross <jgross@suse.com>

1 files changed, 68 insertions, 0 deletions

diff --git a/arch/x86/xen/enlighten_pvh.c b/arch/x86/xen/enlighten_pvh.c
index ada3868c02c2..c28f073c1df5 100644
--- a/arch/x86/xen/enlighten_pvh.c
+++ b/arch/x86/xen/enlighten_pvh.c
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 #include <linux/acpi.h>
 #include <linux/export.h>
+#include <linux/mm.h>
 
 #include <xen/hvc-console.h>
 
@@ -72,3 +73,70 @@ void __init mem_map_via_hcall(struct boot_params *boot_params_p)
 	}
 	boot_params_p->e820_entries = memmap.nr_entries;
 }
+
+/*
+ * Reserve e820 UNUSABLE regions to inflate the memory balloon.
+ *
+ * On PVH dom0 the host memory map is used, RAM regions available to dom0 are
+ * located as the same place as in the native memory map, but since dom0 gets
+ * less memory than the total amount of host RAM the ranges that can't be
+ * populated are converted from RAM -> UNUSABLE.  Use such regions (up to the
+ * ratio signaled in EXTRA_MEM_RATIO) in order to inflate the balloon driver at
+ * boot.  Doing so prevents the guest (even if just temporary) from using holes
+ * in the memory map in order to map grants or foreign addresses, and
+ * hopefully limits the risk of a clash with a device MMIO region.  Ideally the
+ * hypervisor should notify us which memory ranges are suitable for creating
+ * foreign mappings, but that's not yet implemented.
+ */
+void __init xen_reserve_extra_memory(struct boot_params *bootp)
+{
+	unsigned int i, ram_pages = 0, extra_pages;
+
+	for (i = 0; i < bootp->e820_entries; i++) {
+		struct boot_e820_entry *e = &bootp->e820_table[i];
+
+		if (e->type != E820_TYPE_RAM)
+			continue;
+		ram_pages += PFN_DOWN(e->addr + e->size) - PFN_UP(e->addr);
+	}
+
+	/* Max amount of extra memory. */
+	extra_pages = EXTRA_MEM_RATIO * ram_pages;
+
+	/*
+	 * Convert UNUSABLE ranges to RAM and reserve them for foreign mapping
+	 * purposes.
+	 */
+	for (i = 0; i < bootp->e820_entries && extra_pages; i++) {
+		struct boot_e820_entry *e = &bootp->e820_table[i];
+		unsigned long pages;
+
+		if (e->type != E820_TYPE_UNUSABLE)
+			continue;
+
+		pages = min(extra_pages,
+			PFN_DOWN(e->addr + e->size) - PFN_UP(e->addr));
+
+		if (pages != (PFN_DOWN(e->addr + e->size) - PFN_UP(e->addr))) {
+			struct boot_e820_entry *next;
+
+			if (bootp->e820_entries ==
+			    ARRAY_SIZE(bootp->e820_table))
+				/* No space left to split - skip region. */
+				continue;
+
+			/* Split entry. */
+			next = e + 1;
+			memmove(next, e,
+				(bootp->e820_entries - i) * sizeof(*e));
+			bootp->e820_entries++;
+			next->addr = PAGE_ALIGN(e->addr) + PFN_PHYS(pages);
+			e->size = next->addr - e->addr;
+			next->size -= e->size;
+		}
+		e->type = E820_TYPE_RAM;
+		extra_pages -= pages;
+
+		xen_add_extra_mem(PFN_UP(e->addr), pages);
+	}
+}