[PATCH v2 0/8] powernv/memtrace: don't abuse memory hot(un)plug infrastructure for memory allocations

From: David Hildenbrand <david@redhat.com>
To: linux-kernel@vger.kernel.org
Cc: linux-mm@kvack.org, linuxppc-dev@lists.ozlabs.org,
	David Hildenbrand <david@redhat.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	"Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>,
	Benjamin Herrenschmidt <benh@kernel.crashing.org>,
	Michael Ellerman <mpe@ellerman.id.au>,
	Michal Hocko <mhocko@kernel.org>, Michal Hocko <mhocko@suse.com>,
	Mike Rapoport <rppt@kernel.org>,
	Nicholas Piggin <npiggin@gmail.com>,
	Oscar Salvador <osalvador@suse.de>,
	Paul Mackerras <paulus@samba.org>,
	Rashmica Gupta <rashmica.g@gmail.com>,
	Wei Yang <richard.weiyang@linux.alibaba.com>
Subject: [PATCH v2 0/8] powernv/memtrace: don't abuse memory hot(un)plug infrastructure for memory allocations
Date: Wed, 11 Nov 2020 15:53:14 +0100	[thread overview]
Message-ID: <20201111145322.15793-1-david@redhat.com> (raw)

Based on latest linux/master

powernv/memtrace is the only in-kernel user that rips out random memory
it never added (doesn't own) in order to allocate memory without a
linear mapping. Let's stop abusing memory hot(un)plug infrastructure for
that - use alloc_contig_pages() for allocating memory and remove the
linear mapping manually.

The original idea was discussed in:
 https://lkml.kernel.org/r/48340e96-7e6b-736f-9e23-d3111b915b6e@redhat.com

I only tested via QEMU TCG with a single NUMA node- see patch #8 for more
details.

Error handling and cleanup handling in memtrace code is a mess - that
should definitely get cleaned up sooner or later. Once we have __GFP_ZERO
support for alloc_contig_pages(), we can drop manual clearing. I added
a TODO for now, so this series can go via the powerpc tree - the __GFP_ZERO
change is then better suited via the mm tree, along with support for
__GFP_ZERO.

v1 -> v2:
- Tweaks to patch descriptions
- "powernv/memtrace: don't leak kernel memory to user space"
-- Added. Reported by Michael.
- "powernv/memtrace: fix crashing the kernel when enabling concurrently"
-- Added, discovered while testing.
- "powerpc/mm: protect linear mapping modifications by a mutex"
-- Added. Although we currently won't have concurrency, this is cleaner and
   future-proof.
- "powerepc/book3s64/hash: drop WARN_ON in hash__remove_section_mapping"
-- Added. Suggested by Oscar
- "powernv/memtrace: don't abuse memory hot(un)plug infrastructure for
   memory allocations"
-- Reshuffle the code to make review easier.
-- Add a TODO regarding __GFP_ZERO. Adapt to changed page clearing code.
-- Use GFP_KERNEL | __GFP_THISNODE | __GFP_NOWARN for allocations.

David Hildenbrand (8):
  powernv/memtrace: don't leak kernel memory to user space
  powernv/memtrace: fix crashing the kernel when enabling concurrently
  powerpc/mm: factor out creating/removing linear mapping
  powerpc/mm: protect linear mapping modifications by a mutex
  powerpc/mm: print warning in arch_remove_linear_mapping()
  powerepc/book3s64/hash: drop WARN_ON in hash__remove_section_mapping
  powerpc/mm: remove linear mapping if __add_pages() fails in
    arch_add_memory()
  powernv/memtrace: don't abuse memory hot(un)plug infrastructure for
    memory allocations

 arch/powerpc/mm/book3s64/hash_utils.c     |   1 -
 arch/powerpc/mm/mem.c                     |  53 +++++--
 arch/powerpc/platforms/powernv/Kconfig    |   8 +-
 arch/powerpc/platforms/powernv/memtrace.c | 175 ++++++++++------------
 include/linux/memory_hotplug.h            |   3 +
 5 files changed, 125 insertions(+), 115 deletions(-)

-- 
2.26.2