Re: [PATCH v6 1/4] mm: add NR_SECONDARY_PAGETABLE to count secondary page table uses.

From: Yosry Ahmed
Date: Mon Aug 22 2022 - 20:05:41 EST


On Wed, Aug 17, 2022 at 3:27 PM Yosry Ahmed <yosryahmed@xxxxxxxxxx> wrote:
>
> On Wed, Aug 17, 2022 at 10:24 AM Andrew Morton
> <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
> >
> > On Tue, 28 Jun 2022 22:09:35 +0000 Yosry Ahmed <yosryahmed@xxxxxxxxxx> wrote:
> >
> > > We keep track of several kernel memory stats (total kernel memory, page
> > > tables, stack, vmalloc, etc) on multiple levels (global, per-node,
> > > per-memcg, etc). These stats give insights to users to how much memory
> > > is used by the kernel and for what purposes.
> > >
> > > Currently, memory used by kvm mmu is not accounted in any of those
> > > kernel memory stats. This patch series accounts the memory pages
> > > used by KVM for page tables in those stats in a new
> > > NR_SECONDARY_PAGETABLE stat. This stat can be later extended to account
> > > for other types of secondary pages tables (e.g. iommu page tables).
> > >
> > > KVM has a decent number of large allocations that aren't for page
> > > tables, but for most of them, the number/size of those allocations
> > > scales linearly with either the number of vCPUs or the amount of memory
> > > assigned to the VM. KVM's secondary page table allocations do not scale
> > > linearly, especially when nested virtualization is in use.
> > >
> > > >From a KVM perspective, NR_SECONDARY_PAGETABLE will scale with KVM's
> > > per-VM pages_{4k,2m,1g} stats unless the guest is doing something
> > > bizarre (e.g. accessing only 4kb chunks of 2mb pages so that KVM is
> > > forced to allocate a large number of page tables even though the guest
> > > isn't accessing that much memory). However, someone would need to either
> > > understand how KVM works to make that connection, or know (or be told) to
> > > go look at KVM's stats if they're running VMs to better decipher the stats.
> > >
> > > Furthermore, having NR_PAGETABLE side-by-side with NR_SECONDARY_PAGETABLE
> > > is informative. For example, when backing a VM with THP vs. HugeTLB,
> > > NR_SECONDARY_PAGETABLE is roughly the same, but NR_PAGETABLE is an order
> > > of magnitude higher with THP. So having this stat will at the very least
> > > prove to be useful for understanding tradeoffs between VM backing types,
> > > and likely even steer folks towards potential optimizations.
> > >
> > > The original discussion with more details about the rationale:
> > > https://lore.kernel.org/all/87ilqoi77b.wl-maz@xxxxxxxxxx
> > >
> > > This stat will be used by subsequent patches to count KVM mmu
> > > memory usage.
> >
> > Nits and triviata:
> >
> > > --- a/Documentation/filesystems/proc.rst
> > > +++ b/Documentation/filesystems/proc.rst
> > > @@ -977,6 +977,7 @@ Example output. You may not have all of these fields.
> > > SUnreclaim: 142336 kB
> > > KernelStack: 11168 kB
> > > PageTables: 20540 kB
> > > + SecPageTables: 0 kB
> > > NFS_Unstable: 0 kB
> > > Bounce: 0 kB
> > > WritebackTmp: 0 kB
> > > @@ -1085,6 +1086,9 @@ KernelStack
> > > Memory consumed by the kernel stacks of all tasks
> > > PageTables
> > > Memory consumed by userspace page tables
> > > +SecPageTables
> > > + Memory consumed by secondary page tables, this currently
> > > + currently includes KVM mmu allocations on x86 and arm64.
> >
> > Something happened to the whitespace there.
>
> Yeah I have the fix for this queued for v7. Thanks!
>
> >
> > > + "Node %d SecPageTables: %8lu kB\n"
> > > ...
> > > + nid, K(node_page_state(pgdat, NR_SECONDARY_PAGETABLE)),
> >
> > The use of "sec" in the user-facing changes and "secondary" in the
> > programmer-facing changes is irksome. Can we be consistent? I'd
> > prefer "secondary" throughout.
> >
>
> SecondaryPageTables is too long (unfortunately), it messes up the
> formatting in node_read_meminfo() and meminfo_proc_show(). I would
> prefer "secondary" as well, but I don't know if breaking the format in
> this way is okay.

Any thoughts here Andrew? Change to SecondaryPageTables anyway? Change
all to use "sec" instead of "secondary"? Leave as-is?


>
> This is what I mean by breaking the format btw (the numbers become misaligned):
>
> diff --git a/drivers/base/node.c b/drivers/base/node.c
> index 5ad56a0cd593..4f85750a0f8e 100644
> --- a/drivers/base/node.c
> +++ b/drivers/base/node.c
> @@ -433,7 +433,7 @@ static ssize_t node_read_meminfo(struct device *dev,
> "Node %d ShadowCallStack:%8lu kB\n"
> #endif
> "Node %d PageTables: %8lu kB\n"
> - "Node %d SecPageTables: %8lu kB\n"
> + "Node %d SecondaryPageTables: %8lu kB\n"
> "Node %d NFS_Unstable: %8lu kB\n"
> "Node %d Bounce: %8lu kB\n"
> "Node %d WritebackTmp: %8lu kB\n"
> diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
> index 208efd4fa52c..b7166d09a38f 100644
> --- a/fs/proc/meminfo.c
> +++ b/fs/proc/meminfo.c
> @@ -115,7 +115,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
> #endif
> show_val_kb(m, "PageTables: ",
> global_node_page_state(NR_PAGETABLE));
> - show_val_kb(m, "SecPageTables: ",
> + show_val_kb(m, "SecondaryPageTables: ",
> global_node_page_state(NR_SECONDARY_PAGETABLE));
>
> show_val_kb(m, "NFS_Unstable: ", 0);