mm: fix race between __split_huge_pmd_locked() and GUP-fast

commit 3a5a8d343e1cf96eb9971b17cbd4b832ab19b8e7 upstream.

__split_huge_pmd_locked() can be called for a present THP, devmap or
(non-prese

mm: fix race between __split_huge_pmd_locked() and GUP-fast

commit 3a5a8d343e1cf96eb9971b17cbd4b832ab19b8e7 upstream.

__split_huge_pmd_locked() can be called for a present THP, devmap or
(non-present) migration entry. It calls pmdp_invalidate() unconditionally
on the pmdp and only determines if it is present or not based on the
returned old pmd. This is a problem for the migration entry case because
pmd_mkinvalid(), called by pmdp_invalidate() must only be called for a
present pmd.

On arm64 at least, pmd_mkinvalid() will mark the pmd such that any future
call to pmd_present() will return true. And therefore any lockless
pgtable walker could see the migration entry pmd in this state and start
interpretting the fields as if it were present, leading to BadThings (TM).
GUP-fast appears to be one such lockless pgtable walker.

x86 does not suffer the above problem, but instead pmd_mkinvalid() will
corrupt the offset field of the swap entry within the swap pte. See link
below for discussion of that problem.

Fix all of this by only calling pmdp_invalidate() for a present pmd. And
for good measure let's add a warning to all implementations of
pmdp_invalidate[_ad](). I've manually reviewed all other
pmdp_invalidate[_ad]() call sites and believe all others to be conformant.

This is a theoretical bug found during code review. I don't have any test
case to trigger it in practice.

Link: https://lkml.kernel.org/r/20240501143310.1381675-1-ryan.roberts@arm.com
Link: https://lore.kernel.org/all/0dd7827a-6334-439a-8fd0-43c98e6af22b@arm.com/
Fixes: 84c3fc4e9c56 ("mm: thp: check pmd migration entry in common path")
Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Andreas Larsson <andreas@gaisler.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Aneesh Kumar K.V <aneesh.kumar@kernel.org>
Cc: Borislav Petkov (AMD) <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Naveen N. Rao <naveen.n.rao@linux.ibm.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

show more ...

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guest

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to
avoid postcopy hangs") introduced an undesired side effect when combined
with memory ballooning and VM migration: memory part of the inflated
memory balloon will consume memory.

Assuming we have a 100GiB VM and inflated the balloon to 40GiB. Our VM
will consume ~60GiB of memory. If we now trigger a VM migration,
hypervisors like QEMU will read all VM memory. As s390x does not support
the shared zeropage, we'll end up allocating for all previously-inflated
memory part of the memory balloon: 50 GiB. So we might easily
(unexpectedly) crash the VM on the migration source.

Even worse, hypervisors like QEMU optimize for zeropage migration to not
consume memory on the migration destination: when migrating a
"page full of zeroes", on the migration destination they check whether the
target memory is already zero (by reading the destination memory) and avoid
writing to the memory to not allocate memory: however, s390x will also
allocate memory here, implying that also on the migration destination, we
will end up allocating all previously-inflated memory part of the memory
balloon.

This is especially bad if actual memory overcommit was not desired, when
memory ballooning is used for dynamic VM memory resizing, setting aside
some memory during boot that can be added later on demand. Alternatives
like virtio-mem that would avoid this issue are not yet available on
s390x.

There could be ways to optimize some cases in user space: before reading
memory in an anonymous private mapping on the migration source, check via
/proc/self/pagemap if anything is already populated. Similarly check on
the migration destination before reading. While that would avoid
populating tables full of shared zeropages on all architectures, it's
harder to get right and performant, and requires user space changes.

Further, with posctopy live migration we must place a page, so there,
"avoid touching memory to avoid allocating memory" is not really
possible. (Note that a previously we would have falsely inserted
shared zeropages into processes using UFFDIO_ZEROPAGE where
mm_forbids_zeropage() would have actually forbidden it)

PV is currently incompatible with memory ballooning, and in the common
case, KVM guests don't make use of storage keys. Instead of zapping
zeropages when enabling storage keys / PV, that turned out to be
problematic in the past, let's do exactly the same we do with KSM pages:
trigger unsharing faults to replace the shared zeropages by proper
anonymous folios.

What about added latency when enabling storage kes? Having a lot of
zeropages in applicable environments (PV, legacy guests, unittests) is
unexpected. Further, KSM could today already unshare the zeropages
and unmerging KSM pages when enabling storage kets would unshare the
KSM-placed zeropages in the same way, resulting in the same latency.

[ agordeev: Fixed sparse and checkpatch complaints and error handling ]

Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Fixes: fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs")
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20240411161441.910170-3-david@redhat.com
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

show more ...

mm: fix race between __split_huge_pmd_locked() and GUP-fast

commit 3a5a8d343e1cf96eb9971b17cbd4b832ab19b8e7 upstream.

__split_huge_pmd_locked() can be called for a present THP, devmap or
(non-prese

mm: fix race between __split_huge_pmd_locked() and GUP-fast

commit 3a5a8d343e1cf96eb9971b17cbd4b832ab19b8e7 upstream.

__split_huge_pmd_locked() can be called for a present THP, devmap or
(non-present) migration entry. It calls pmdp_invalidate() unconditionally
on the pmdp and only determines if it is present or not based on the
returned old pmd. This is a problem for the migration entry case because
pmd_mkinvalid(), called by pmdp_invalidate() must only be called for a
present pmd.

On arm64 at least, pmd_mkinvalid() will mark the pmd such that any future
call to pmd_present() will return true. And therefore any lockless
pgtable walker could see the migration entry pmd in this state and start
interpretting the fields as if it were present, leading to BadThings (TM).
GUP-fast appears to be one such lockless pgtable walker.

x86 does not suffer the above problem, but instead pmd_mkinvalid() will
corrupt the offset field of the swap entry within the swap pte. See link
below for discussion of that problem.

Fix all of this by only calling pmdp_invalidate() for a present pmd. And
for good measure let's add a warning to all implementations of
pmdp_invalidate[_ad](). I've manually reviewed all other
pmdp_invalidate[_ad]() call sites and believe all others to be conformant.

This is a theoretical bug found during code review. I don't have any test
case to trigger it in practice.

Link: https://lkml.kernel.org/r/20240501143310.1381675-1-ryan.roberts@arm.com
Link: https://lore.kernel.org/all/0dd7827a-6334-439a-8fd0-43c98e6af22b@arm.com/
Fixes: 84c3fc4e9c56 ("mm: thp: check pmd migration entry in common path")
Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Andreas Larsson <andreas@gaisler.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Aneesh Kumar K.V <aneesh.kumar@kernel.org>
Cc: Borislav Petkov (AMD) <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Naveen N. Rao <naveen.n.rao@linux.ibm.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

show more ...

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guest

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to
avoid postcopy hangs") introduced an undesired side effect when combined
with memory ballooning and VM migration: memory part of the inflated
memory balloon will consume memory.

Assuming we have a 100GiB VM and inflated the balloon to 40GiB. Our VM
will consume ~60GiB of memory. If we now trigger a VM migration,
hypervisors like QEMU will read all VM memory. As s390x does not support
the shared zeropage, we'll end up allocating for all previously-inflated
memory part of the memory balloon: 50 GiB. So we might easily
(unexpectedly) crash the VM on the migration source.

Even worse, hypervisors like QEMU optimize for zeropage migration to not
consume memory on the migration destination: when migrating a
"page full of zeroes", on the migration destination they check whether the
target memory is already zero (by reading the destination memory) and avoid
writing to the memory to not allocate memory: however, s390x will also
allocate memory here, implying that also on the migration destination, we
will end up allocating all previously-inflated memory part of the memory
balloon.

This is especially bad if actual memory overcommit was not desired, when
memory ballooning is used for dynamic VM memory resizing, setting aside
some memory during boot that can be added later on demand. Alternatives
like virtio-mem that would avoid this issue are not yet available on
s390x.

There could be ways to optimize some cases in user space: before reading
memory in an anonymous private mapping on the migration source, check via
/proc/self/pagemap if anything is already populated. Similarly check on
the migration destination before reading. While that would avoid
populating tables full of shared zeropages on all architectures, it's
harder to get right and performant, and requires user space changes.

Further, with posctopy live migration we must place a page, so there,
"avoid touching memory to avoid allocating memory" is not really
possible. (Note that a previously we would have falsely inserted
shared zeropages into processes using UFFDIO_ZEROPAGE where
mm_forbids_zeropage() would have actually forbidden it)

PV is currently incompatible with memory ballooning, and in the common
case, KVM guests don't make use of storage keys. Instead of zapping
zeropages when enabling storage keys / PV, that turned out to be
problematic in the past, let's do exactly the same we do with KSM pages:
trigger unsharing faults to replace the shared zeropages by proper
anonymous folios.

What about added latency when enabling storage kes? Having a lot of
zeropages in applicable environments (PV, legacy guests, unittests) is
unexpected. Further, KSM could today already unshare the zeropages
and unmerging KSM pages when enabling storage kets would unshare the
KSM-placed zeropages in the same way, resulting in the same latency.

[ agordeev: Fixed sparse and checkpatch complaints and error handling ]

Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Fixes: fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs")
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20240411161441.910170-3-david@redhat.com
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

show more ...

mm: fix race between __split_huge_pmd_locked() and GUP-fast

commit 3a5a8d343e1cf96eb9971b17cbd4b832ab19b8e7 upstream.

__split_huge_pmd_locked() can be called for a present THP, devmap or
(non-prese

mm: fix race between __split_huge_pmd_locked() and GUP-fast

commit 3a5a8d343e1cf96eb9971b17cbd4b832ab19b8e7 upstream.

__split_huge_pmd_locked() can be called for a present THP, devmap or
(non-present) migration entry. It calls pmdp_invalidate() unconditionally
on the pmdp and only determines if it is present or not based on the
returned old pmd. This is a problem for the migration entry case because
pmd_mkinvalid(), called by pmdp_invalidate() must only be called for a
present pmd.

On arm64 at least, pmd_mkinvalid() will mark the pmd such that any future
call to pmd_present() will return true. And therefore any lockless
pgtable walker could see the migration entry pmd in this state and start
interpretting the fields as if it were present, leading to BadThings (TM).
GUP-fast appears to be one such lockless pgtable walker.

x86 does not suffer the above problem, but instead pmd_mkinvalid() will
corrupt the offset field of the swap entry within the swap pte. See link
below for discussion of that problem.

Fix all of this by only calling pmdp_invalidate() for a present pmd. And
for good measure let's add a warning to all implementations of
pmdp_invalidate[_ad](). I've manually reviewed all other
pmdp_invalidate[_ad]() call sites and believe all others to be conformant.

This is a theoretical bug found during code review. I don't have any test
case to trigger it in practice.

Link: https://lkml.kernel.org/r/20240501143310.1381675-1-ryan.roberts@arm.com
Link: https://lore.kernel.org/all/0dd7827a-6334-439a-8fd0-43c98e6af22b@arm.com/
Fixes: 84c3fc4e9c56 ("mm: thp: check pmd migration entry in common path")
Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Andreas Larsson <andreas@gaisler.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Aneesh Kumar K.V <aneesh.kumar@kernel.org>
Cc: Borislav Petkov (AMD) <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Naveen N. Rao <naveen.n.rao@linux.ibm.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

show more ...

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guest

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to
avoid postcopy hangs") introduced an undesired side effect when combined
with memory ballooning and VM migration: memory part of the inflated
memory balloon will consume memory.

Assuming we have a 100GiB VM and inflated the balloon to 40GiB. Our VM
will consume ~60GiB of memory. If we now trigger a VM migration,
hypervisors like QEMU will read all VM memory. As s390x does not support
the shared zeropage, we'll end up allocating for all previously-inflated
memory part of the memory balloon: 50 GiB. So we might easily
(unexpectedly) crash the VM on the migration source.

Even worse, hypervisors like QEMU optimize for zeropage migration to not
consume memory on the migration destination: when migrating a
"page full of zeroes", on the migration destination they check whether the
target memory is already zero (by reading the destination memory) and avoid
writing to the memory to not allocate memory: however, s390x will also
allocate memory here, implying that also on the migration destination, we
will end up allocating all previously-inflated memory part of the memory
balloon.

This is especially bad if actual memory overcommit was not desired, when
memory ballooning is used for dynamic VM memory resizing, setting aside
some memory during boot that can be added later on demand. Alternatives
like virtio-mem that would avoid this issue are not yet available on
s390x.

There could be ways to optimize some cases in user space: before reading
memory in an anonymous private mapping on the migration source, check via
/proc/self/pagemap if anything is already populated. Similarly check on
the migration destination before reading. While that would avoid
populating tables full of shared zeropages on all architectures, it's
harder to get right and performant, and requires user space changes.

Further, with posctopy live migration we must place a page, so there,
"avoid touching memory to avoid allocating memory" is not really
possible. (Note that a previously we would have falsely inserted
shared zeropages into processes using UFFDIO_ZEROPAGE where
mm_forbids_zeropage() would have actually forbidden it)

PV is currently incompatible with memory ballooning, and in the common
case, KVM guests don't make use of storage keys. Instead of zapping
zeropages when enabling storage keys / PV, that turned out to be
problematic in the past, let's do exactly the same we do with KSM pages:
trigger unsharing faults to replace the shared zeropages by proper
anonymous folios.

What about added latency when enabling storage kes? Having a lot of
zeropages in applicable environments (PV, legacy guests, unittests) is
unexpected. Further, KSM could today already unshare the zeropages
and unmerging KSM pages when enabling storage kets would unshare the
KSM-placed zeropages in the same way, resulting in the same latency.

[ agordeev: Fixed sparse and checkpatch complaints and error handling ]

Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Fixes: fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs")
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20240411161441.910170-3-david@redhat.com
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

show more ...

mm: fix race between __split_huge_pmd_locked() and GUP-fast

commit 3a5a8d343e1cf96eb9971b17cbd4b832ab19b8e7 upstream.

__split_huge_pmd_locked() can be called for a present THP, devmap or
(non-prese

mm: fix race between __split_huge_pmd_locked() and GUP-fast

commit 3a5a8d343e1cf96eb9971b17cbd4b832ab19b8e7 upstream.

__split_huge_pmd_locked() can be called for a present THP, devmap or
(non-present) migration entry. It calls pmdp_invalidate() unconditionally
on the pmdp and only determines if it is present or not based on the
returned old pmd. This is a problem for the migration entry case because
pmd_mkinvalid(), called by pmdp_invalidate() must only be called for a
present pmd.

On arm64 at least, pmd_mkinvalid() will mark the pmd such that any future
call to pmd_present() will return true. And therefore any lockless
pgtable walker could see the migration entry pmd in this state and start
interpretting the fields as if it were present, leading to BadThings (TM).
GUP-fast appears to be one such lockless pgtable walker.

x86 does not suffer the above problem, but instead pmd_mkinvalid() will
corrupt the offset field of the swap entry within the swap pte. See link
below for discussion of that problem.

Fix all of this by only calling pmdp_invalidate() for a present pmd. And
for good measure let's add a warning to all implementations of
pmdp_invalidate[_ad](). I've manually reviewed all other
pmdp_invalidate[_ad]() call sites and believe all others to be conformant.

This is a theoretical bug found during code review. I don't have any test
case to trigger it in practice.

Link: https://lkml.kernel.org/r/20240501143310.1381675-1-ryan.roberts@arm.com
Link: https://lore.kernel.org/all/0dd7827a-6334-439a-8fd0-43c98e6af22b@arm.com/
Fixes: 84c3fc4e9c56 ("mm: thp: check pmd migration entry in common path")
Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Andreas Larsson <andreas@gaisler.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Aneesh Kumar K.V <aneesh.kumar@kernel.org>
Cc: Borislav Petkov (AMD) <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Naveen N. Rao <naveen.n.rao@linux.ibm.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

show more ...

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guest

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to
avoid postcopy hangs") introduced an undesired side effect when combined
with memory ballooning and VM migration: memory part of the inflated
memory balloon will consume memory.

Assuming we have a 100GiB VM and inflated the balloon to 40GiB. Our VM
will consume ~60GiB of memory. If we now trigger a VM migration,
hypervisors like QEMU will read all VM memory. As s390x does not support
the shared zeropage, we'll end up allocating for all previously-inflated
memory part of the memory balloon: 50 GiB. So we might easily
(unexpectedly) crash the VM on the migration source.

Even worse, hypervisors like QEMU optimize for zeropage migration to not
consume memory on the migration destination: when migrating a
"page full of zeroes", on the migration destination they check whether the
target memory is already zero (by reading the destination memory) and avoid
writing to the memory to not allocate memory: however, s390x will also
allocate memory here, implying that also on the migration destination, we
will end up allocating all previously-inflated memory part of the memory
balloon.

This is especially bad if actual memory overcommit was not desired, when
memory ballooning is used for dynamic VM memory resizing, setting aside
some memory during boot that can be added later on demand. Alternatives
like virtio-mem that would avoid this issue are not yet available on
s390x.

There could be ways to optimize some cases in user space: before reading
memory in an anonymous private mapping on the migration source, check via
/proc/self/pagemap if anything is already populated. Similarly check on
the migration destination before reading. While that would avoid
populating tables full of shared zeropages on all architectures, it's
harder to get right and performant, and requires user space changes.

Further, with posctopy live migration we must place a page, so there,
"avoid touching memory to avoid allocating memory" is not really
possible. (Note that a previously we would have falsely inserted
shared zeropages into processes using UFFDIO_ZEROPAGE where
mm_forbids_zeropage() would have actually forbidden it)

PV is currently incompatible with memory ballooning, and in the common
case, KVM guests don't make use of storage keys. Instead of zapping
zeropages when enabling storage keys / PV, that turned out to be
problematic in the past, let's do exactly the same we do with KSM pages:
trigger unsharing faults to replace the shared zeropages by proper
anonymous folios.

What about added latency when enabling storage kes? Having a lot of
zeropages in applicable environments (PV, legacy guests, unittests) is
unexpected. Further, KSM could today already unshare the zeropages
and unmerging KSM pages when enabling storage kets would unshare the
KSM-placed zeropages in the same way, resulting in the same latency.

[ agordeev: Fixed sparse and checkpatch complaints and error handling ]

Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Fixes: fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs")
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20240411161441.910170-3-david@redhat.com
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

show more ...

mm: fix race between __split_huge_pmd_locked() and GUP-fast

commit 3a5a8d343e1cf96eb9971b17cbd4b832ab19b8e7 upstream.

__split_huge_pmd_locked() can be called for a present THP, devmap or
(non-prese

mm: fix race between __split_huge_pmd_locked() and GUP-fast

commit 3a5a8d343e1cf96eb9971b17cbd4b832ab19b8e7 upstream.

__split_huge_pmd_locked() can be called for a present THP, devmap or
(non-present) migration entry. It calls pmdp_invalidate() unconditionally
on the pmdp and only determines if it is present or not based on the
returned old pmd. This is a problem for the migration entry case because
pmd_mkinvalid(), called by pmdp_invalidate() must only be called for a
present pmd.

On arm64 at least, pmd_mkinvalid() will mark the pmd such that any future
call to pmd_present() will return true. And therefore any lockless
pgtable walker could see the migration entry pmd in this state and start
interpretting the fields as if it were present, leading to BadThings (TM).
GUP-fast appears to be one such lockless pgtable walker.

x86 does not suffer the above problem, but instead pmd_mkinvalid() will
corrupt the offset field of the swap entry within the swap pte. See link
below for discussion of that problem.

Fix all of this by only calling pmdp_invalidate() for a present pmd. And
for good measure let's add a warning to all implementations of
pmdp_invalidate[_ad](). I've manually reviewed all other
pmdp_invalidate[_ad]() call sites and believe all others to be conformant.

This is a theoretical bug found during code review. I don't have any test
case to trigger it in practice.

Link: https://lkml.kernel.org/r/20240501143310.1381675-1-ryan.roberts@arm.com
Link: https://lore.kernel.org/all/0dd7827a-6334-439a-8fd0-43c98e6af22b@arm.com/
Fixes: 84c3fc4e9c56 ("mm: thp: check pmd migration entry in common path")
Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Andreas Larsson <andreas@gaisler.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Aneesh Kumar K.V <aneesh.kumar@kernel.org>
Cc: Borislav Petkov (AMD) <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Naveen N. Rao <naveen.n.rao@linux.ibm.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

show more ...

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guest

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to
avoid postcopy hangs") introduced an undesired side effect when combined
with memory ballooning and VM migration: memory part of the inflated
memory balloon will consume memory.

Assuming we have a 100GiB VM and inflated the balloon to 40GiB. Our VM
will consume ~60GiB of memory. If we now trigger a VM migration,
hypervisors like QEMU will read all VM memory. As s390x does not support
the shared zeropage, we'll end up allocating for all previously-inflated
memory part of the memory balloon: 50 GiB. So we might easily
(unexpectedly) crash the VM on the migration source.

Even worse, hypervisors like QEMU optimize for zeropage migration to not
consume memory on the migration destination: when migrating a
"page full of zeroes", on the migration destination they check whether the
target memory is already zero (by reading the destination memory) and avoid
writing to the memory to not allocate memory: however, s390x will also
allocate memory here, implying that also on the migration destination, we
will end up allocating all previously-inflated memory part of the memory
balloon.

This is especially bad if actual memory overcommit was not desired, when
memory ballooning is used for dynamic VM memory resizing, setting aside
some memory during boot that can be added later on demand. Alternatives
like virtio-mem that would avoid this issue are not yet available on
s390x.

There could be ways to optimize some cases in user space: before reading
memory in an anonymous private mapping on the migration source, check via
/proc/self/pagemap if anything is already populated. Similarly check on
the migration destination before reading. While that would avoid
populating tables full of shared zeropages on all architectures, it's
harder to get right and performant, and requires user space changes.

Further, with posctopy live migration we must place a page, so there,
"avoid touching memory to avoid allocating memory" is not really
possible. (Note that a previously we would have falsely inserted
shared zeropages into processes using UFFDIO_ZEROPAGE where
mm_forbids_zeropage() would have actually forbidden it)

PV is currently incompatible with memory ballooning, and in the common
case, KVM guests don't make use of storage keys. Instead of zapping
zeropages when enabling storage keys / PV, that turned out to be
problematic in the past, let's do exactly the same we do with KSM pages:
trigger unsharing faults to replace the shared zeropages by proper
anonymous folios.

What about added latency when enabling storage kes? Having a lot of
zeropages in applicable environments (PV, legacy guests, unittests) is
unexpected. Further, KSM could today already unshare the zeropages
and unmerging KSM pages when enabling storage kets would unshare the
KSM-placed zeropages in the same way, resulting in the same latency.

[ agordeev: Fixed sparse and checkpatch complaints and error handling ]

Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Fixes: fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs")
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20240411161441.910170-3-david@redhat.com
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

show more ...

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guest

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to
avoid postcopy hangs") introduced an undesired side effect when combined
with memory ballooning and VM migration: memory part of the inflated
memory balloon will consume memory.

Assuming we have a 100GiB VM and inflated the balloon to 40GiB. Our VM
will consume ~60GiB of memory. If we now trigger a VM migration,
hypervisors like QEMU will read all VM memory. As s390x does not support
the shared zeropage, we'll end up allocating for all previously-inflated
memory part of the memory balloon: 50 GiB. So we might easily
(unexpectedly) crash the VM on the migration source.

Even worse, hypervisors like QEMU optimize for zeropage migration to not
consume memory on the migration destination: when migrating a
"page full of zeroes", on the migration destination they check whether the
target memory is already zero (by reading the destination memory) and avoid
writing to the memory to not allocate memory: however, s390x will also
allocate memory here, implying that also on the migration destination, we
will end up allocating all previously-inflated memory part of the memory
balloon.

This is especially bad if actual memory overcommit was not desired, when
memory ballooning is used for dynamic VM memory resizing, setting aside
some memory during boot that can be added later on demand. Alternatives
like virtio-mem that would avoid this issue are not yet available on
s390x.

There could be ways to optimize some cases in user space: before reading
memory in an anonymous private mapping on the migration source, check via
/proc/self/pagemap if anything is already populated. Similarly check on
the migration destination before reading. While that would avoid
populating tables full of shared zeropages on all architectures, it's
harder to get right and performant, and requires user space changes.

Further, with posctopy live migration we must place a page, so there,
"avoid touching memory to avoid allocating memory" is not really
possible. (Note that a previously we would have falsely inserted
shared zeropages into processes using UFFDIO_ZEROPAGE where
mm_forbids_zeropage() would have actually forbidden it)

PV is currently incompatible with memory ballooning, and in the common
case, KVM guests don't make use of storage keys. Instead of zapping
zeropages when enabling storage keys / PV, that turned out to be
problematic in the past, let's do exactly the same we do with KSM pages:
trigger unsharing faults to replace the shared zeropages by proper
anonymous folios.

What about added latency when enabling storage kes? Having a lot of
zeropages in applicable environments (PV, legacy guests, unittests) is
unexpected. Further, KSM could today already unshare the zeropages
and unmerging KSM pages when enabling storage kets would unshare the
KSM-placed zeropages in the same way, resulting in the same latency.

[ agordeev: Fixed sparse and checkpatch complaints and error handling ]

Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Fixes: fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs")
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20240411161441.910170-3-david@redhat.com
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

show more ...

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guest

s390/mm: Re-enable the shared zeropage for !PV and !skeys KVM guests

[ Upstream commit 06201e00ee3e4beacac48aab2b83eff64ebf0bc0 ]

commit fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to
avoid postcopy hangs") introduced an undesired side effect when combined
with memory ballooning and VM migration: memory part of the inflated
memory balloon will consume memory.

Assuming we have a 100GiB VM and inflated the balloon to 40GiB. Our VM
will consume ~60GiB of memory. If we now trigger a VM migration,
hypervisors like QEMU will read all VM memory. As s390x does not support
the shared zeropage, we'll end up allocating for all previously-inflated
memory part of the memory balloon: 50 GiB. So we might easily
(unexpectedly) crash the VM on the migration source.

Even worse, hypervisors like QEMU optimize for zeropage migration to not
consume memory on the migration destination: when migrating a
"page full of zeroes", on the migration destination they check whether the
target memory is already zero (by reading the destination memory) and avoid
writing to the memory to not allocate memory: however, s390x will also
allocate memory here, implying that also on the migration destination, we
will end up allocating all previously-inflated memory part of the memory
balloon.

This is especially bad if actual memory overcommit was not desired, when
memory ballooning is used for dynamic VM memory resizing, setting aside
some memory during boot that can be added later on demand. Alternatives
like virtio-mem that would avoid this issue are not yet available on
s390x.

There could be ways to optimize some cases in user space: before reading
memory in an anonymous private mapping on the migration source, check via
/proc/self/pagemap if anything is already populated. Similarly check on
the migration destination before reading. While that would avoid
populating tables full of shared zeropages on all architectures, it's
harder to get right and performant, and requires user space changes.

Further, with posctopy live migration we must place a page, so there,
"avoid touching memory to avoid allocating memory" is not really
possible. (Note that a previously we would have falsely inserted
shared zeropages into processes using UFFDIO_ZEROPAGE where
mm_forbids_zeropage() would have actually forbidden it)

PV is currently incompatible with memory ballooning, and in the common
case, KVM guests don't make use of storage keys. Instead of zapping
zeropages when enabling storage keys / PV, that turned out to be
problematic in the past, let's do exactly the same we do with KSM pages:
trigger unsharing faults to replace the shared zeropages by proper
anonymous folios.

What about added latency when enabling storage kes? Having a lot of
zeropages in applicable environments (PV, legacy guests, unittests) is
unexpected. Further, KSM could today already unshare the zeropages
and unmerging KSM pages when enabling storage kets would unshare the
KSM-placed zeropages in the same way, resulting in the same latency.

[ agordeev: Fixed sparse and checkpatch complaints and error handling ]

Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Fixes: fa41ba0d08de ("s390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs")
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20240411161441.910170-3-david@redhat.com
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

show more ...

mm/treewide: replace pud_large() with pud_leaf()

[ Upstream commit 0a845e0f6348ccfa2dcc8c450ffd1c9ffe8c4add ]

pud_large() is always defined as pud_leaf(). Merge their usages. Chose
pud_leaf() bec

mm/treewide: replace pud_large() with pud_leaf()

[ Upstream commit 0a845e0f6348ccfa2dcc8c450ffd1c9ffe8c4add ]

pud_large() is always defined as pud_leaf(). Merge their usages. Chose
pud_leaf() because pud_leaf() is a global API, while pud_large() is not.

Link: https://lkml.kernel.org/r/20240305043750.93762-9-peterx@redhat.com
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Kirill A. Shutemov <kirill@shutemov.name>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: "Naveen N. Rao" <naveen.n.rao@linux.ibm.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Cc: Yang Shi <shy828301@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Stable-dep-of: c567f2948f57 ("Revert "x86/mm/ident_map: Use gbpages only where full GB page should be mapped."")
Signed-off-by: Sasha Levin <sashal@kernel.org>

show more ...

s390: implement the new page table range API

Add set_ptes() and update_mmu_cache_range().

Link: https://lkml.kernel.org/r/20230802151406.3735276-24-willy@infradead.org
Signed-off-by: Matthew Wilcox

s390: implement the new page table range API

Add set_ptes() and update_mmu_cache_range().

Link: https://lkml.kernel.org/r/20230802151406.3735276-24-willy@infradead.org
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
Acked-by: Mike Rapoport (IBM) <rppt@kernel.org>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
Cc: Sven Schnelle <svens@linux.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

show more ...

s390/mm: get rid of VMEM_MAX_PHYS macro

There are no users of VMEM_MAX_PHYS macro left, remove it.

Reviewed-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@linux.i

s390/mm: get rid of VMEM_MAX_PHYS macro

There are no users of VMEM_MAX_PHYS macro left, remove it.

Reviewed-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

show more ...

mm: Rename arch pte_mkwrite()'s to pte_mkwrite_novma()

The x86 Shadow stack feature includes a new type of memory called shadow
stack. This shadow stack memory has some unusual properties, which req

mm: Rename arch pte_mkwrite()'s to pte_mkwrite_novma()

The x86 Shadow stack feature includes a new type of memory called shadow
stack. This shadow stack memory has some unusual properties, which requires
some core mm changes to function properly.

One of these unusual properties is that shadow stack memory is writable,
but only in limited ways. These limits are applied via a specific PTE
bit combination. Nevertheless, the memory is writable, and core mm code
will need to apply the writable permissions in the typical paths that
call pte_mkwrite(). The goal is to make pte_mkwrite() take a VMA, so
that the x86 implementation of it can know whether to create regular
writable or shadow stack mappings.

But there are a couple of challenges to this. Modifying the signatures of
each arch pte_mkwrite() implementation would be error prone because some
are generated with macros and would need to be re-implemented. Also, some
pte_mkwrite() callers operate on kernel memory without a VMA.

So this can be done in a three step process. First pte_mkwrite() can be
renamed to pte_mkwrite_novma() in each arch, with a generic pte_mkwrite()
added that just calls pte_mkwrite_novma(). Next callers without a VMA can
be moved to pte_mkwrite_novma(). And lastly, pte_mkwrite() and all callers
can be changed to take/pass a VMA.

Start the process by renaming pte_mkwrite() to pte_mkwrite_novma() and
adding the pte_mkwrite() wrapper in linux/pgtable.h. Apply the same
pattern for pmd_mkwrite(). Since not all archs have a pmd_mkwrite_novma(),
create a new arch config HAS_HUGE_PAGE that can be used to tell if
pmd_mkwrite() should be defined. Otherwise in the !HAS_HUGE_PAGE cases the
compiler would not be able to find pmd_mkwrite_novma().

No functional change.

Suggested-by: Linus Torvalds <torvalds@linuxfoundation.org>
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Mike Rapoport (IBM) <rppt@kernel.org>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/lkml/CAHk-=wiZjSu7c9sFYZb3q04108stgHff2wfbokGCCgW7riz+8Q@mail.gmail.com/
Link: https://lore.kernel.org/all/20230613001108.3040476-2-rick.p.edgecombe%40intel.com

show more ...

Revert "s390/mm: get rid of VMEM_MAX_PHYS macro"

This reverts commit 456be42aa713e7f83b467db66ceae779431c7d9d.

The assumption VMEM_MAX_PHYS should match ident_map_size
is wrong. At least discontigu

Revert "s390/mm: get rid of VMEM_MAX_PHYS macro"

This reverts commit 456be42aa713e7f83b467db66ceae779431c7d9d.

The assumption VMEM_MAX_PHYS should match ident_map_size
is wrong. At least discontiguous saved segments (DCSS)
could be loaded at addresses beyond ident_map_size and
dcssblk device driver might fail as result.

Reported-by: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>

show more ...

s390/mm: get rid of VMEM_MAX_PHYS macro

VMEM_MAX_PHYS is supposed to be the highest physical
address that can be added to the identity mapping.
It should match ident_map_size, which has the same
mea

s390/mm: get rid of VMEM_MAX_PHYS macro

VMEM_MAX_PHYS is supposed to be the highest physical
address that can be added to the identity mapping.
It should match ident_map_size, which has the same
meaning. However, unlike ident_map_size it is not
adjusted against various limiting factors (see the
comment to setup_ident_map_size() function). That
renders all checks against VMEM_MAX_PHYS invalid.

Further, VMEM_MAX_PHYS is currently set to vmemmap,
which is an address in virtual memory space. However,
it gets compared against physical addresses in various
locations. That works, because both address spaces
are the same on s390, but otherwise it is wrong.

Instead of fixing VMEM_MAX_PHYS misuse and semantics
just remove it.

Acked-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>

show more ...

procfs: consolidate arch_report_meminfo declaration

The arch_report_meminfo() function is provided by four architectures,
with a __weak fallback in procfs itself. On architectures that don't
have a

procfs: consolidate arch_report_meminfo declaration

The arch_report_meminfo() function is provided by four architectures,
with a __weak fallback in procfs itself. On architectures that don't
have a custom version, the __weak version causes a warning because
of the missing prototype.

Remove the architecture specific prototypes and instead add one
in linux/proc_fs.h.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com> # for arch/x86
Acked-by: Helge Deller <deller@gmx.de> # parisc
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Message-Id: <20230516195834.551901-1-arnd@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>

show more ...

s390/mm: fix direct map accounting

Commit bb1520d581a3 ("s390/mm: start kernel with DAT enabled") did not
implement direct map accounting in the early page table setup code. In
result the reported v

s390/mm: fix direct map accounting

Commit bb1520d581a3 ("s390/mm: start kernel with DAT enabled") did not
implement direct map accounting in the early page table setup code. In
result the reported values are bogus now:

$cat /proc/meminfo
...
DirectMap4k: 5120 kB
DirectMap1M: 18446744073709546496 kB
DirectMap2G: 0 kB

Fix this by adding the missing accounting. The result looks sane again:

$cat /proc/meminfo
...
DirectMap4k: 6156 kB
DirectMap1M: 2091008 kB
DirectMap2G: 6291456 kB

Fixes: bb1520d581a3 ("s390/mm: start kernel with DAT enabled")
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>

show more ...

mm: add PTE pointer parameter to flush_tlb_fix_spurious_fault()

s390 can do more fine-grained handling of spurious TLB protection faults,
when there also is the PTE pointer available.

Therefore, pa

mm: add PTE pointer parameter to flush_tlb_fix_spurious_fault()

s390 can do more fine-grained handling of spurious TLB protection faults,
when there also is the PTE pointer available.

Therefore, pass on the PTE pointer to flush_tlb_fix_spurious_fault() as an
additional parameter.

This will add no functional change to other architectures, but those with
private flush_tlb_fix_spurious_fault() implementations need to be made
aware of the new parameter.

Link: https://lkml.kernel.org/r/20230306161548.661740-1-gerald.schaefer@linux.ibm.com
Signed-off-by: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com> [arm64]
Acked-by: Michael Ellerman <mpe@ellerman.id.au> [powerpc]
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Borislav Petkov (AMD) <bp@alien8.de>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

show more ...

s390/mm: add support for RDP (Reset DAT-Protection)

RDP instruction allows to reset DAT-protection bit in a PTE, with less
CPU synchronization overhead than IPTE instruction. In particular, IPTE
can

s390/mm: add support for RDP (Reset DAT-Protection)

RDP instruction allows to reset DAT-protection bit in a PTE, with less
CPU synchronization overhead than IPTE instruction. In particular, IPTE
can cause machine-wide synchronization overhead, and excessive IPTE usage
can negatively impact machine performance.

RDP can be used instead of IPTE, if the new PTE only differs in SW bits
and _PAGE_PROTECT HW bit, for PTE protection changes from RO to RW.
SW PTE bit changes are allowed, e.g. for dirty and young tracking, but none
of the other HW-defined part of the PTE must change. This is because the
architecture forbids such changes to an active and valid PTE, which
is why invalidation with IPTE is always used first, before writing a new
entry.

The RDP optimization helps mainly for fault-driven SW dirty-bit tracking.
Writable PTEs are initially always mapped with HW _PAGE_PROTECT bit set,
to allow SW dirty-bit accounting on first write protection fault, where
the DAT-protection would then be reset. The reset is now done with RDP
instead of IPTE, if RDP instruction is available.

RDP cannot always guarantee that the DAT-protection reset is propagated
to all CPUs immediately. This means that spurious TLB protection faults
on other CPUs can now occur. For this, common code provides a
flush_tlb_fix_spurious_fault() handler, which will now be used to do a
CPU-local TLB flush. However, this will clear the whole TLB of a CPU, and
not just the affected entry. For more fine-grained flushing, by simply
doing a (local) RDP again, flush_tlb_fix_spurious_fault() would need to
also provide the PTE pointer.

Note that spurious TLB protection faults cannot really be distinguished
from racing pagetable updates, where another thread already installed the
correct PTE. In such a case, the local TLB flush would be unnecessary
overhead, but overall reduction of CPU synchronization overhead by not
using IPTE is still expected to be beneficial.

Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

show more ...

mm: remove __HAVE_ARCH_PTE_SWP_EXCLUSIVE

__HAVE_ARCH_PTE_SWP_EXCLUSIVE is now supported by all architectures that
support swp PTEs, so let's drop it.

Link: https://lkml.kernel.org/r/20230113171026.

mm: remove __HAVE_ARCH_PTE_SWP_EXCLUSIVE

__HAVE_ARCH_PTE_SWP_EXCLUSIVE is now supported by all architectures that
support swp PTEs, so let's drop it.

Link: https://lkml.kernel.org/r/20230113171026.582290-27-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

show more ...

s390/mm: start kernel with DAT enabled

The setup of the kernel virtual address space is spread
throughout the sources, boot stages and config options
like this:

1. The available physical memory reg

s390/mm: start kernel with DAT enabled

The setup of the kernel virtual address space is spread
throughout the sources, boot stages and config options
like this:

1. The available physical memory regions are queried
and stored as mem_detect information for later use
in the decompressor.

2. Based on the physical memory availability the virtual
memory layout is established in the decompressor;

3. If CONFIG_KASAN is disabled the kernel paging setup
code populates kernel pgtables and turns DAT mode on.
It uses the information stored at step [1].

4. If CONFIG_KASAN is enabled the kernel early boot
kasan setup populates kernel pgtables and turns DAT
mode on. It uses the information stored at step [1].

The kasan setup creates early_pg_dir directory and
directly overwrites swapper_pg_dir entries to make
shadow memory pages available.

Move the kernel virtual memory setup to the decompressor
and start the kernel with DAT turned on right from the
very first istruction. That completely eliminates the
boot phase when the kernel runs in DAT-off mode, simplies
the overall design and consolidates pgtables setup.

The identity mapping is created in the decompressor, while
kasan shadow mappings are still created by the early boot
kernel code.

Share with decompressor the existing kasan memory allocator.
It decreases the size of a newly requested memory block from
pgalloc_pos and ensures that kernel image is not overwritten.
pgalloc_low and pgalloc_pos pointers are made preserved boot
variables for that.

Use the bootdata infrastructure to setup swapper_pg_dir
and invalid_pg_dir directories used by the kernel later.
The interim early_pg_dir directory established by the
kasan initialization code gets eliminated as result.

As the kernel runs in DAT-on mode only the PSW_KERNEL_BITS
define gets PSW_MASK_DAT bit by default. Additionally, the
setup_lowcore_dat_off() and setup_lowcore_dat_on() routines
get merged, since there is no DAT-off mode stage anymore.

The memory mappings are created with RW+X protection that
allows the early boot code setting up all necessary data
and services for the kernel being booted. Just before the
paging is enabled the memory protection is changed to
RO+X for text, RO+NX for read-only data and RW+NX for
kernel data and the identity mapping.

Reviewed-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

show more ...

s390/pgtable: add REGION3_KERNEL_EXEC protection

Similar to existing PAGE_KERNEL_EXEC and SEGMENT_KERNEL_EXEC
memory protection add REGION3_KERNEL_EXEC attribute that
could be set on PUD pgtable ent

s390/pgtable: add REGION3_KERNEL_EXEC protection

Similar to existing PAGE_KERNEL_EXEC and SEGMENT_KERNEL_EXEC
memory protection add REGION3_KERNEL_EXEC attribute that
could be set on PUD pgtable entries.

Reviewed-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

show more ...