Trying VMAP_STACK with KVM, vmlinux was not starting.
This was due to SRR0 and SRR1 clobbered by an ISI due to
the rfi being in a different page than the mtsrr0/1:
c0003fe0 <mmu_off>:
c0003fe0: 38 83 00 54 addi r4,r3,84
c0003fe4: 7c 60 00 a6 mfmsr r3
c0003fe8: 70 60 00 30 andi. r0,r3,48
c0003fec: 4d 82 00 20 beqlr
c0003ff0: 7c 63 00 78 andc r3,r3,r0
c0003ff4: 7c 9a 03 a6 mtsrr0 r4
c0003ff8: 7c 7b 03 a6 mtsrr1 r3
c0003ffc: 7c 00 04 ac hwsync
c0004000: 4c 00 00 64 rfi
Align the 4 instruction block used to deactivate MMU to order 4,
so that the block never crosses a page boundary.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/30d2cda111b7977227fff067fa7e358440e2b3a4.1576916812.git.christophe.leroy@c-s.fr
The part decidated to handling hash_page() is fully unneeded for
processors not having real hash pages like the 603.
Lets enlarge the content of the feature fixup, and provide
an alternative which jumps directly instead of getting NIPs.
Also, in preparation of VMAP stacks, the end of DSI handler has moved
to later in the code as it won't fit anymore once VMAP stacks
are there.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/c31b22c91af8b011d0a4fd9e52ad6afb4b593f71.1576916812.git.christophe.leroy@c-s.fr
When we enable VMAP_STACK there will not be enough room for the
alignment handler at 0x600 in head_8xx.S. For now move the tail of the
alignment handler out of line, and branch to it.
Suggested-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
To support CONFIG_VMAP_STACK, the kernel has to activate Data MMU
Translation for accessing the stack. Before doing that it must save
SRR0, SRR1 and also DAR and DSISR when relevant, in order to not
loose them in case there is a Data TLB Miss once the translation is
reactivated.
This patch adds fields in thread struct for saving those registers.
It prepares entry_32.S to handle exception entry with
Data MMU Translation enabled and alters EXCEPTION_PROLOG macros to
save SRR0, SRR1, DAR and DSISR then reenables Data MMU.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/a775a1fea60f190e0f63503463fb775310a2009b.1576916812.git.christophe.leroy@c-s.fr
Commit a25bd72bad ("powerpc/mm/radix: Workaround prefetch issue with
KVM") introduced a number of workarounds as coming out of a guest with
the mmu enabled would make the cpu would start running in hypervisor
state with the PID value from the guest. The cpu will then start
prefetching for the hypervisor with that PID value.
In Power9 DD2.2 the cpu behaviour was modified to fix this. When
accessing Quadrant 0 in hypervisor mode with LPID != 0 prefetching will
not be performed. This means that we can get rid of the workarounds for
Power9 DD2.2 and later revisions. Add a new cpu feature
CPU_FTR_P9_RADIX_PREFETCH_BUG to indicate if the workarounds are needed.
Signed-off-by: Jordan Niethe <jniethe5@gmail.com>
Acked-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20191206031722.25781-1-jniethe5@gmail.com
Many drivers don't check for errors when they get a 0xFFs response from an
MMIO load. As a result after an EEH event occurs a driver can get stuck in
a polling loop unless it some kind of internal timeout logic.
Currently EEH tries to detect and report stuck drivers by dumping a stack
trace after eeh_dev_check_failure() is called EEH_MAX_FAILS times on an
already frozen PE. The value of EEH_MAX_FAILS was chosen so that a dump
would occur every few seconds if the driver was spinning in a loop. This
results in a lot of spurious stack traces in the kernel log.
Fix this by limiting it to printing one stack trace for each PE freeze. If
the driver is truely stuck the kernel's hung task detector is better suited
to reporting the probelm anyway.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Reviewed-by: Sam Bobroff <sbobroff@linux.ibm.com>
Tested-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20191016012536.22588-1-oohall@gmail.com
The powerpc PCI code requires that a pci_dn structure exists for all
devices in the system. This is fine for real devices since at boot a pci_dn
is created for each PCI device in the DT and it's fine for hotplugged devices
since the hotplug slot driver will manage the pci_dn's devices in hotplug
slots. For SR-IOV, we need the platform / pcibios to manage the pci_dn for
virtual functions since firmware is unaware of VFs, and they aren't
"hot plugged" in the traditional sense.
Management of the pci_dn is handled by the, poorly named, functions:
add_pci_dev_data() and remove_pci_dev_data(). The entire body of these
functions is #ifdef`ed around CONFIG_PCI_IOV and they cannot be used
in any other context, so make them only available when CONFIG_PCI_IOV
is selected, and rename them to reflect their actual usage rather than
having them masquerade as generic code.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Reviewed-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190821062655.19735-2-oohall@gmail.com
When disabling virtual functions on an SR-IOV adapter we currently do not
correctly remove the EEH state for the now-dead virtual functions. When
removing the pci_dn that was created for the VF when SR-IOV was enabled
we free the corresponding eeh_dev without removing it from the child device
list of the eeh_pe that contained it. This can result in crashes due to the
use-after-free.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Reviewed-by: Sam Bobroff <sbobroff@linux.ibm.com>
Tested-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190821062655.19735-1-oohall@gmail.com
The eeh_sysfs_remove_device() function is supposed to clear the
EEH_DEV_SYSFS flag since it indicates the EEH sysfs entries have been added
for a pci_dev.
When the sysfs files are removed eeh_remove_device() the eeh_dev and the
pci_dev have already been de-associated. This then causes the
pci_dev_to_eeh_dev() call in eeh_sysfs_remove_device() to return NULL so
the flag can't be cleared from the still-live eeh_dev. This problem is
worked around in the caller by clearing the flag manually. However, this
behaviour doesn't make a whole lot of sense, so this patch fixes it by:
a) Re-ordering eeh_remove_device() so that eeh_sysfs_remove_device() is
called before de-associating the pci_dev and eeh_dev.
b) Making eeh_sysfs_remove_device() emit a warning if there's no
corresponding eeh_dev for a pci_dev. The paths where the sysfs
files are only reachable if EEH was setup for the device
for the device in the first place so hitting this warning
indicates a programming error.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Reviewed-by: Sam Bobroff <sbobroff@linux.ibm.com>
Tested-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190715085612.8802-6-oohall@gmail.com
There are several EEH sysfs properties that only exists when the
"ibm,is-open-sriov-pf" property appears in the device tree node of the PCI
device. This used on pseries to indicate to the guest that the hypervisor
allows the guest to configure the SR-IOV capability. Doing this requires
some handshaking between the guest, hypervisor and userspace when a VF is
EEH frozen which is why these properties exist.
This is all dead code on non-pseries platforms so wrap it in an #ifdef
CONFIG_PPC_PSERIES to make the dependency clearer.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Tested-by: Sam Bobroff <sbobroff@linux.ibm.com>
Reviewed-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190715085612.8802-4-oohall@gmail.com
__get_datapage() is only a few instructions to retrieve the
address of the page where the kernel stores data to the VDSO.
By inlining this function into its users, a bl/blr pair and
a mflr/mtlr pair is avoided, plus a few reg moves.
The improvement is noticeable (about 55 nsec/call on an 8xx)
vdsotest before the patch:
gettimeofday: vdso: 731 nsec/call
clock-gettime-realtime-coarse: vdso: 668 nsec/call
clock-gettime-monotonic-coarse: vdso: 745 nsec/call
vdsotest after the patch:
gettimeofday: vdso: 677 nsec/call
clock-gettime-realtime-coarse: vdso: 613 nsec/call
clock-gettime-monotonic-coarse: vdso: 690 nsec/call
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/c39ef7f3dfa25356b01e211d539671f279086c09.1575273217.git.christophe.leroy@c-s.fr
Commit 18ad51dd34 ("powerpc: Add VDSO version of getcpu") added
getcpu() for PPC64 only, by making use of a user readable general
purpose SPR.
PPC32 doesn't have any such SPR.
For non SMP, just return CPU id 0 from the VDSO directly.
PPC32 doesn't support CONFIG_NUMA so NUMA node is always 0.
Before the patch, vdsotest reported:
getcpu: syscall: 1572 nsec/call
getcpu: libc: 1787 nsec/call
getcpu: vdso: not tested
Now, vdsotest reports:
getcpu: syscall: 1582 nsec/call
getcpu: libc: 502 nsec/call
getcpu: vdso: 187 nsec/call
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/eaac4b6494ecff1811220fccc895bf282aab884a.1575273217.git.christophe.leroy@c-s.fr
Unlike standard powerpc, Powerpc 8xx doesn't have SPRN_DABR, but
it has a breakpoint support based on a set of comparators which
allow more flexibility.
Commit 4ad8622dc5 ("powerpc/8xx: Implement hw_breakpoint")
implemented breakpoints by emulating the DABR behaviour. It did
this by setting one comparator the match 4 bytes at breakpoint address
and the other comparator to match 4 bytes at breakpoint address + 4.
Rewrite 8xx hw_breakpoint to make breakpoints match all addresses
defined by the breakpoint address and length by making full use of
comparators.
Now, comparator E is set to match any address greater than breakpoint
address minus one. Comparator F is set to match any address lower than
breakpoint address plus breakpoint length. Addresses are aligned
to 32 bits.
When the breakpoint range starts at address 0, the breakpoint is set
to match comparator F only. When the breakpoint range end at address
0xffffffff, the breakpoint is set to match comparator E only.
Otherwise the breakpoint is set to match comparator E and F.
At the same time, use registers bit names instead of hardcoded values.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/05105deeaf63bc02151aea2cdeaf525534e0e9d4.1574790198.git.christophe.leroy@c-s.fr
/* Background. */
For a very long time, extending openat(2) with new features has been
incredibly frustrating. This stems from the fact that openat(2) is
possibly the most famous counter-example to the mantra "don't silently
accept garbage from userspace" -- it doesn't check whether unknown flags
are present[1].
This means that (generally) the addition of new flags to openat(2) has
been fraught with backwards-compatibility issues (O_TMPFILE has to be
defined as __O_TMPFILE|O_DIRECTORY|[O_RDWR or O_WRONLY] to ensure old
kernels gave errors, since it's insecure to silently ignore the
flag[2]). All new security-related flags therefore have a tough road to
being added to openat(2).
Userspace also has a hard time figuring out whether a particular flag is
supported on a particular kernel. While it is now possible with
contemporary kernels (thanks to [3]), older kernels will expose unknown
flag bits through fcntl(F_GETFL). Giving a clear -EINVAL during
openat(2) time matches modern syscall designs and is far more
fool-proof.
In addition, the newly-added path resolution restriction LOOKUP flags
(which we would like to expose to user-space) don't feel related to the
pre-existing O_* flag set -- they affect all components of path lookup.
We'd therefore like to add a new flag argument.
Adding a new syscall allows us to finally fix the flag-ignoring problem,
and we can make it extensible enough so that we will hopefully never
need an openat3(2).
/* Syscall Prototype. */
/*
* open_how is an extensible structure (similar in interface to
* clone3(2) or sched_setattr(2)). The size parameter must be set to
* sizeof(struct open_how), to allow for future extensions. All future
* extensions will be appended to open_how, with their zero value
* acting as a no-op default.
*/
struct open_how { /* ... */ };
int openat2(int dfd, const char *pathname,
struct open_how *how, size_t size);
/* Description. */
The initial version of 'struct open_how' contains the following fields:
flags
Used to specify openat(2)-style flags. However, any unknown flag
bits or otherwise incorrect flag combinations (like O_PATH|O_RDWR)
will result in -EINVAL. In addition, this field is 64-bits wide to
allow for more O_ flags than currently permitted with openat(2).
mode
The file mode for O_CREAT or O_TMPFILE.
Must be set to zero if flags does not contain O_CREAT or O_TMPFILE.
resolve
Restrict path resolution (in contrast to O_* flags they affect all
path components). The current set of flags are as follows (at the
moment, all of the RESOLVE_ flags are implemented as just passing
the corresponding LOOKUP_ flag).
RESOLVE_NO_XDEV => LOOKUP_NO_XDEV
RESOLVE_NO_SYMLINKS => LOOKUP_NO_SYMLINKS
RESOLVE_NO_MAGICLINKS => LOOKUP_NO_MAGICLINKS
RESOLVE_BENEATH => LOOKUP_BENEATH
RESOLVE_IN_ROOT => LOOKUP_IN_ROOT
open_how does not contain an embedded size field, because it is of
little benefit (userspace can figure out the kernel open_how size at
runtime fairly easily without it). It also only contains u64s (even
though ->mode arguably should be a u16) to avoid having padding fields
which are never used in the future.
Note that as a result of the new how->flags handling, O_PATH|O_TMPFILE
is no longer permitted for openat(2). As far as I can tell, this has
always been a bug and appears to not be used by userspace (and I've not
seen any problems on my machines by disallowing it). If it turns out
this breaks something, we can special-case it and only permit it for
openat(2) but not openat2(2).
After input from Florian Weimer, the new open_how and flag definitions
are inside a separate header from uapi/linux/fcntl.h, to avoid problems
that glibc has with importing that header.
/* Testing. */
In a follow-up patch there are over 200 selftests which ensure that this
syscall has the correct semantics and will correctly handle several
attack scenarios.
In addition, I've written a userspace library[4] which provides
convenient wrappers around openat2(RESOLVE_IN_ROOT) (this is necessary
because no other syscalls support RESOLVE_IN_ROOT, and thus lots of care
must be taken when using RESOLVE_IN_ROOT'd file descriptors with other
syscalls). During the development of this patch, I've run numerous
verification tests using libpathrs (showing that the API is reasonably
usable by userspace).
/* Future Work. */
Additional RESOLVE_ flags have been suggested during the review period.
These can be easily implemented separately (such as blocking auto-mount
during resolution).
Furthermore, there are some other proposed changes to the openat(2)
interface (the most obvious example is magic-link hardening[5]) which
would be a good opportunity to add a way for userspace to restrict how
O_PATH file descriptors can be re-opened.
Another possible avenue of future work would be some kind of
CHECK_FIELDS[6] flag which causes the kernel to indicate to userspace
which openat2(2) flags and fields are supported by the current kernel
(to avoid userspace having to go through several guesses to figure it
out).
[1]: https://lwn.net/Articles/588444/
[2]: https://lore.kernel.org/lkml/CA+55aFyyxJL1LyXZeBsf2ypriraj5ut1XkNDsunRBqgVjZU_6Q@mail.gmail.com
[3]: commit 629e014bb8 ("fs: completely ignore unknown open flags")
[4]: https://sourceware.org/bugzilla/show_bug.cgi?id=17523
[5]: https://lore.kernel.org/lkml/20190930183316.10190-2-cyphar@cyphar.com/
[6]: https://youtu.be/ggD-eb3yPVs
Suggested-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Aleksa Sarai <cyphar@cyphar.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This implements the tricky tracing and soft irq handling bits in C,
leaving the low level bit to asm.
A functional difference is that this redirects the interrupt exit to
a return stub to execute blr, rather than the lr address itself. This
is probably barely measurable on real hardware, but it keeps the link
stack balanced.
Tested with QEMU.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
[mpe: Move power4_fixup_nap back into exceptions-64s.S]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190711022404.18132-1-npiggin@gmail.com
With the previous patch applied pcibios_setup_device() will always be run
when pcibios_bus_add_device() is called. There are several code paths where
pcibios_setup_bus_device() is still called (the PowerPC specific PCI
hotplug support is one) so with just the previous patch applied the setup
can be run multiple times on a device, once before the device is added
to the bus and once after.
There's no need to run the setup in the early case any more so just
remove it entirely.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Tested-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20191028085424.12006-3-oohall@gmail.com
Move PCI device setup from pcibios_add_device() and pcibios_fixup_bus() to
pcibios_bus_add_device(). This ensures that platform-specific DMA and IOMMU
setup occurs after the device has been registered in sysfs, which is a
requirement for IOMMU group assignment to work
This fixes IOMMU group assignment for hotplugged devices on pseries, where
the existing behavior results in IOMMU assignment before registration.
Thanks to Lukas Wunner <lukas@wunner.de> for the suggestion.
Signed-off-by: Shawn Anastasio <shawn@anastas.io>
Tested-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20191028085424.12006-2-oohall@gmail.com
On pseries there is a bug with adding hotplugged devices to an IOMMU
group. For a number of dumb reasons fixing that bug first requires
re-working how VFs are configured on PowerNV. For background, on
PowerNV we use the pcibios_sriov_enable() hook to do two things:
1. Create a pci_dn structure for each of the VFs, and
2. Configure the PHB's internal BARs so the MMIO range for each VF
maps to a unique PE.
Roughly speaking a PE is the hardware counterpart to a Linux IOMMU
group since all the devices in a PE share the same IOMMU table. A PE
also defines the set of devices that should be isolated in response to
a PCI error (i.e. bad DMA, UR/CA, AER events, etc). When isolated all
MMIO and DMA traffic to and from devicein the PE is blocked by the
root complex until the PE is recovered by the OS.
The requirement to block MMIO causes a giant headache because the P8
PHB generally uses a fixed mapping between MMIO addresses and PEs. As
a result we need to delay configuring the IOMMU groups for device
until after MMIO resources are assigned. For physical devices (i.e.
non-VFs) the PE assignment is done in pcibios_setup_bridge() which is
called immediately after the MMIO resources for downstream
devices (and the bridge's windows) are assigned. For VFs the setup is
more complicated because:
a) pcibios_setup_bridge() is not called again when VFs are activated, and
b) The pci_dev for VFs are created by generic code which runs after
pcibios_sriov_enable() is called.
The work around for this is a two step process:
1. A fixup in pcibios_add_device() is used to initialised the cached
pe_number in pci_dn, then
2. A bus notifier then adds the device to the IOMMU group for the PE
specified in pci_dn->pe_number.
A side effect fixing the pseries bug mentioned in the first paragraph
is moving the fixup out of pcibios_add_device() and into
pcibios_bus_add_device(), which is called much later. This results in
step 2. failing because pci_dn->pe_number won't be initialised when
the bus notifier is run.
We can fix this by removing the need for the fixup. The PE for a VF is
known before the VF is even scanned so we can initialise
pci_dn->pe_number pcibios_sriov_enable() instead. Unfortunately,
moving the initialisation causes two problems:
1. We trip the WARN_ON() in the current fixup code, and
2. The EEH core clears pdn->pe_number when recovering a VF and
relies on the fixup to correctly re-set it.
The only justification for either of these is a comment in
eeh_rmv_device() suggesting that pdn->pe_number *must* be set to
IODA_INVALID_PE in order for the VF to be scanned. However, this
comment appears to have no basis in reality. Both bugs can be fixed by
just deleting the code.
Tested-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20191028085424.12006-1-oohall@gmail.com
The SUPPORT_SYSRQ ifdeffery is not nice as:
- May create misunderstanding about sizeof(struct uart_port) between
different objects
- Prevents moving functions from serial_core.h
- Reduces readability (well, it's ifdeffery - it's hard to follow)
In order to remove SUPPORT_SYSRQ, has_sysrq variable has been added.
Initialise it in driver's probe and remove ifdeffery.
In contrast to 8250/8250_of, legacy_serial on powerpc does fill
(struct plat_serial8250_port). The reason is likely that it's done on
device_initcall(), not on probe. So, 8250_core is not yet probed.
Propagate value from platform_device on 8250 probe - in case powepc
legacy driver it's initialized on initcall, in case 8250_of it will be
initialized later on of_platform_serial_setup().
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: linuxppc-dev@lists.ozlabs.org
Signed-off-by: Dmitry Safonov <dima@arista.com>
Link: https://lore.kernel.org/r/20191213000657.931618-6-dima@arista.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
