x86/PVH: improve Dom0 memory size calculation

Assuming that the accounting for IOMMU page tables will also take care
of the P2M needs was wrong: dom0_paging_pages() can determine a far
higher value, high enough for the system to run out of memory while
setting up Dom0. Hence in the case of shared page tables the larger of
the two values needs to be used (without shared page tables the sum of
both continues to be applicable).

To not further complicate the logic, eliminate the up-to-2-iteration
loop in favor of doing a few calculations twice (before and after
calling dom0_paging_pages()). While this will lead to slightly too high
a value in "cpu_pages", it is deemed better to account a few too many
than a few too little.

As a result the calculation is now deemed good enough to no longer
warrant the warning message, which therefore gets dropped.

Also uniformly use paging_mode_enabled(), not is_hvm_domain().

While there also account for two further aspects in the PV case: With
"iommu=dom0-passthrough" no IOMMU page tables would get allocated, so
none need accounting for. And if shadow mode is to be enabled (including
only potentially, because of "pv-l1tf=dom0"), setting aside a suitable
amount for the P2M pool to get populated is also necessary (i.e. similar
to the non-shared-page-tables case of PVH).

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>

diff --git a/xen/arch/x86/dom0_build.c b/xen/arch/x86/dom0_build.c
index 5a7441ed5b79b0c8fe82f874ccf7f6cabdf774e8..a7fec05956c155ec12ad86c5e737b135b4dce4d8 100644 (file)
--- a/xen/arch/x86/dom0_build.c
+++ b/xen/arch/x86/dom0_build.c
@@ -19,6 +19,7 @@
 #include <asm/io_apic.h>
 #include <asm/p2m.h>
 #include <asm/setup.h>
+#include <asm/spec_ctrl.h>
 
 struct memsize {
     long nr_pages;
@@ -321,12 +322,22 @@ unsigned long __init dom0_paging_pages(const struct domain *d,
     return ((memkb + 1023) / 1024) << (20 - PAGE_SHIFT);
 }
 
+
+/*
+ * If allocation isn't specified, reserve 1/16th of available memory for
+ * things like DMA buffers. This reservation is clamped to a maximum of 128MB.
+ */
+static unsigned long __init default_nr_pages(unsigned long avail)
+{
+    return avail - (pv_shim ? pv_shim_mem(avail)
+                            : min(avail / 16, 128UL << (20 - PAGE_SHIFT)));
+}
+
 unsigned long __init dom0_compute_nr_pages(
     struct domain *d, struct elf_dom_parms *parms, unsigned long initrd_len)
 {
     nodeid_t node;
-    unsigned long avail = 0, nr_pages, min_pages, max_pages;
-    bool need_paging;
+    unsigned long avail = 0, nr_pages, min_pages, max_pages, iommu_pages = 0;
 
     /* The ordering of operands is to work around a clang5 issue. */
     if ( CONFIG_DOM0_MEM[0] && !dom0_mem_set )
@@ -344,53 +355,47 @@ unsigned long __init dom0_compute_nr_pages(
         avail -= d->max_vcpus - 1;
 
     /* Reserve memory for iommu_dom0_init() (rough estimate). */
-    if ( is_iommu_enabled(d) )
+    if ( is_iommu_enabled(d) && !iommu_hwdom_passthrough )
     {
         unsigned int s;
 
         for ( s = 9; s < BITS_PER_LONG; s += 9 )
-            avail -= max_pdx >> s;
+            iommu_pages += max_pdx >> s;
+
+        avail -= iommu_pages;
     }
 
-    need_paging = is_hvm_domain(d) &&
-        (!iommu_use_hap_pt(d) || !paging_mode_hap(d));
-    for ( ; ; need_paging = false )
+    if ( paging_mode_enabled(d) || opt_dom0_shadow || opt_pv_l1tf_hwdom )
     {
-        nr_pages = get_memsize(&dom0_size, avail);
-        min_pages = get_memsize(&dom0_min_size, avail);
-        max_pages = get_memsize(&dom0_max_size, avail);
+        unsigned long cpu_pages;
+
+        nr_pages = get_memsize(&dom0_size, avail) ?: default_nr_pages(avail);
 
         /*
-         * If allocation isn't specified, reserve 1/16th of available memory
-         * for things like DMA buffers. This reservation is clamped to a
-         * maximum of 128MB.
+         * Clamp according to min/max limits and available memory
+         * (preliminary).
          */
-        if ( !nr_pages )
-        {
-            nr_pages = avail - (pv_shim ? pv_shim_mem(avail)
-                                 : min(avail / 16, 128UL << (20 - PAGE_SHIFT)));
-            if ( is_hvm_domain(d) && !need_paging )
-                /*
-                 * Temporary workaround message until internal (paging) memory
-                 * accounting required to build a pvh dom0 is improved.
-                 */
-                printk("WARNING: PVH dom0 without dom0_mem set is still unstable. "
-                       "If you get crashes during boot, try adding a dom0_mem parameter\n");
-        }
-
-
-        /* Clamp according to min/max limits and available memory. */
-        nr_pages = max(nr_pages, min_pages);
-        nr_pages = min(nr_pages, max_pages);
+        nr_pages = max(nr_pages, get_memsize(&dom0_min_size, avail));
+        nr_pages = min(nr_pages, get_memsize(&dom0_max_size, avail));
         nr_pages = min(nr_pages, avail);
 
-        if ( !need_paging )
-            break;
+        cpu_pages = dom0_paging_pages(d, nr_pages);
 
-        /* Reserve memory for shadow or HAP. */
-        avail -= dom0_paging_pages(d, nr_pages);
+        if ( !iommu_use_hap_pt(d) )
+            avail -= cpu_pages;
+        else if ( cpu_pages > iommu_pages )
+            avail -= cpu_pages - iommu_pages;
     }
 
+    nr_pages = get_memsize(&dom0_size, avail) ?: default_nr_pages(avail);
+    min_pages = get_memsize(&dom0_min_size, avail);
+    max_pages = get_memsize(&dom0_max_size, avail);
+
+    /* Clamp according to min/max limits and available memory (final). */
+    nr_pages = max(nr_pages, min_pages);
+    nr_pages = min(nr_pages, max_pages);
+    nr_pages = min(nr_pages, avail);
+
     if ( is_pv_domain(d) &&
          (parms->p2m_base == UNSET_ADDR) && !memsize_gt_zero(&dom0_size) &&
          (!memsize_gt_zero(&dom0_min_size) || (nr_pages > min_pages)) )