If the handler took any action to log or deal with the error, set a bit
in mce->kflags so that the default handler on the end of the machine
check chain can see what has been done.
Get rid of NOTIFY_STOP returns. Make the EDAC and dev-mcelog handlers
skip over errors already processed by CEC.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Tested-by: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20200214222720.13168-5-tony.luck@intel.com
... because no one should be interested in spurious MCEs anyway. Make
the filtering unconditional and move it to amd_filter_mce().
Signed-off-by: Borislav Petkov <bp@suse.de>
Tested-by: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20200407163414.18058-2-bp@alien8.de
Fix the following gcc warning:
drivers/edac/xgene_edac.c:1486:7: warning: variable ‘address’ set but
not used [-Wunused-but-set-variable]
u32 address;
^~~~~~~
Remove the unused macro RBERRADDR_RD while at it.
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200409093259.20069-1-yanaijie@huawei.com
Fix spelling (s/Aramda/Armada/) in a log message and in a comment. While
at it, add a trailing '\n' in messages.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Jan Luebbe <jlu@pengutronix.de>
Link: https://lkml.kernel.org/r/20200413041556.3514-1-christophe.jaillet@wanadoo.fr
Pull perf updates from Ingo Molnar:
"The main changes in this cycle were:
Kernel side changes:
- A couple of x86/cpu cleanups and changes were grandfathered in due
to patch dependencies. These clean up the set of CPU model/family
matching macros with a consistent namespace and C99 initializer
style.
- A bunch of updates to various low level PMU drivers:
* AMD Family 19h L3 uncore PMU
* Intel Tiger Lake uncore support
* misc fixes to LBR TOS sampling
- optprobe fixes
- perf/cgroup: optimize cgroup event sched-in processing
- misc cleanups and fixes
Tooling side changes are to:
- perf {annotate,expr,record,report,stat,test}
- perl scripting
- libapi, libperf and libtraceevent
- vendor events on Intel and S390, ARM cs-etm
- Intel PT updates
- Documentation changes and updates to core facilities
- misc cleanups, fixes and other enhancements"
* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (89 commits)
cpufreq/intel_pstate: Fix wrong macro conversion
x86/cpu: Cleanup the now unused CPU match macros
hwrng: via_rng: Convert to new X86 CPU match macros
crypto: Convert to new CPU match macros
ASoC: Intel: Convert to new X86 CPU match macros
powercap/intel_rapl: Convert to new X86 CPU match macros
PCI: intel-mid: Convert to new X86 CPU match macros
mmc: sdhci-acpi: Convert to new X86 CPU match macros
intel_idle: Convert to new X86 CPU match macros
extcon: axp288: Convert to new X86 CPU match macros
thermal: Convert to new X86 CPU match macros
hwmon: Convert to new X86 CPU match macros
platform/x86: Convert to new CPU match macros
EDAC: Convert to new X86 CPU match macros
cpufreq: Convert to new X86 CPU match macros
ACPI: Convert to new X86 CPU match macros
x86/platform: Convert to new CPU match macros
x86/kernel: Convert to new CPU match macros
x86/kvm: Convert to new CPU match macros
x86/perf/events: Convert to new CPU match macros
...
The new macro set has a consistent namespace and uses C99 initializers
instead of the grufty C89 ones.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Acked-by: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20200320131509.673579000@linutronix.de
Since snprintf() returns the would-be-output size instead of the actual
output size, the succeeding calls may go beyond the given buffer limit.
Fix it by replacing with scnprintf().
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Jan Luebbe <jlu@pengutronix.de>
Link: https://lkml.kernel.org/r/20200311071728.4541-1-tiwai@suse.de
On the ZynqMP platform, zynqmp_get_error_info() is used to read out
error information. In this function, the pinf->col parameter is not
used (it is only used by the Zynq platform's zynq_get_error_info()). So
there's no need to print pinf->col on ZynqMP.
In order to differentiate on which platform handle_error() is executed,
use DDR_ECC_INTR_SUPPORT as the check condition to distinguish between
Zynq and ZynqMP platforms.
[ bp: Massage. ]
Fixes: b500b4a029 ("EDAC, synopsys: Add ECC support for ZynqMP DDR controller")
Signed-off-by: Sherry Sun <sherry.sun@nxp.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Manish Narani <manish.narani@xilinx.com>
Link: https://lkml.kernel.org/r/1584365679-27443-1-git-send-email-sherry.sun@nxp.com
handle_error() currently calls snprintf() a couple of times in
succession to output the message for a CE/UE, therefore overwriting each
part of the message which was formatted with the previous snprintf()
call. As a result, only the part of the message from the last snprintf()
call will be printed.
The simplest and most effective way to fix this problem is to combine
the whole string into one which to supply to a single snprintf() call.
[ bp: Massage. ]
Fixes: b500b4a029 ("EDAC, synopsys: Add ECC support for ZynqMP DDR controller")
Signed-off-by: Sherry Sun <sherry.sun@nxp.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: James Morse <james.morse@arm.com>
Cc: Manish Narani <manish.narani@xilinx.com>
Link: https://lkml.kernel.org/r/1582792452-32575-1-git-send-email-sherry.sun@nxp.com
The driver supports error detection and correction on devices with an
ARM DMC-520 memory controller.
Signed-off-by: Lei Wang <leiwang_git@outlook.com>
Signed-off-by: Shiping Ji <shiping.linux@gmail.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: James Morse <james.morse@arm.com>
Link: https://lkml.kernel.org/r/83b48c70-dc06-d0d4-cae9-a2187fca628b@gmail.com
This warning is output for every virtual CPU in a guest on an EPYC 2
system because kvm doesn't enable SMCA. Once is enough too.
[ bp: Massage. ]
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200217134627.19765-1-prarit@redhat.com
Looking at how mci->{ue,ce}_per_layer[EDAC_MAX_LAYERS] is used, it
turns out that only the leaves in the memory hierarchy are consumed
(in sysfs), but not the intermediate layers, e.g.:
count = dimm->mci->ce_per_layer[dimm->mci->n_layers-1][dimm->idx];
These unused counters only add complexity, remove them. The error
counter values are directly stored in struct dimm_info now.
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200123090210.26933-11-rrichter@marvell.com
The error descriptor is passed to the error reporting functions, so
the error details can be directly generated there. Move string
generation from edac_raw_mc_handle_error() to edac_ce_error() and
edac_ue_error(). The intermediate detail[] string can be removed then.
Also, cleanup the string generation by switching to a single variant
only using the ternary operator.
[ bp: put ternary operators on a separate line for better readability
and use the short-form "inline if" in edac_mc_handle_error(). ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200123090210.26933-10-rrichter@marvell.com
Most arguments of error reporting functions are already stored in the
struct edac_raw_error_desc error descriptor. Pass the error descriptor
to the functions and reduce the functions' argument list.
[ bp: Sort function args in reverse fir tree order. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200123090210.26933-9-rrichter@marvell.com
Many functions carry the enable_per_layer_report argument. This is a
bool value indicating the error information contains some location
data where the error occurred. This can easily being determined by
checking the pos[] array for values. Negative values indicate there is
no location available. So if the top layer is negative, the error
location is unknown.
Just check if the top layer is negative and remove
enable_per_layer_report as function argument and also from struct
edac_raw_error_desc.
[ bp: Reflow comments to 80 columns, while at it. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200123090210.26933-8-rrichter@marvell.com
There is a limitation to report only EDAC_MAX_LABELS in e->label of
the error descriptor. This is to prevent a potential string overflow.
The current implementation falls back to "any memory" in this case and
also stops all further processing to find a unique row and channel of
the possible error location.
Reporting "any memory" is wrong as the memory controller reported an
error location for one of the layers. Instead, report "unknown memory"
and also do not break early in the loop to further check row and channel
for uniqueness.
[ bp: Massage commit message. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200123090210.26933-7-rrichter@marvell.com
Carve out the error_count increment into a separate function
edac_inc_csrow(). This better separates code and reduces the indentation
level.
Implementation note: The function edac_inc_csrow() counts the same
as before, ->ce_count is only incremented if row >= 0. This is esp.
true for the case of (!e->enable_per_layer_report). Here, a DIMM was
not found, variable row still has a value of -1 and ->ce_count is not
incremented.
[ bp: Massage commit message. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200214141757.8976-1-rrichter@marvell.com
Each struct mci has its own error descriptor. Create a function
error_desc_to_mci() to determine the corresponding mci from an
error descriptor. This removes @mci from the parameter list of
edac_raw_mc_handle_error() as the mci pointer does not need to be passed
any longer.
[ bp: Massage commit message. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200123090210.26933-5-rrichter@marvell.com
Store the error type in struct edac_raw_error_desc. This makes the
type parameter of edac_raw_mc_handle_error() obsolete.
[ kernel-doc typo ]
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200123090210.26933-4-rrichter@marvell.com
Reorder the new created functions edac_mc_alloc_csrows() and
edac_mc_alloc_dimms() and move them before edac_mc_alloc(). No further
code changes.
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200123090210.26933-3-rrichter@marvell.com
edac_mc_alloc() is huge. Factor out code by moving it to the two new
functions edac_mc_alloc_csrows() and edac_mc_alloc_dimms(). Do not
move code yet for better review.
[ bp: sort local args in reversed fir tree order. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200123090210.26933-2-rrichter@marvell.com
There are dimm and csrow devices linked to the mci device esp. to show
up in sysfs. It must be granted that children devices are removed before
its mci parent. Thus, the release functions must be called in the
correct order and may not miss any child before releasing its parent. In
the current implementation this is only granted by the correct order of
release functions.
A much better approach is to use put_device() that releases the device
only after all users are gone. It is the recommended way to release a
device and free its memory. The function uses the device's refcount and
only frees it if there are no users of it anymore such as children.
So implement a mci_release() function to remove mci devices, use
put_device() to free them and early initialize the mci device right
after its struct has been allocated.
Change the release function so that it can be universally used no
matter if the device is registered or not. Since subsequent dimm
and csrow sysfs links are implemented as children devices, their
refcounts will keep the parent mci device from being removed as long
as sysfs entries exist and until all users have been unregistered in
edac_remove_sysfs_mci_device().
Remove edac_unregister_sysfs() and merge mci sysfs removal into
edac_remove_sysfs_mci_device(). There is only a single instance now that
removes the sysfs entries. The function can now be used in the error
paths for cleanup.
Also, create device release functions for all involved devices
(dev->release), remove device_type release functions (dev_type->
release) and also use dev->init_name instead of dev_set_name().
[ bp: Massage commit message and comments. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Link: https://lkml.kernel.org/r/20200212120340.4764-5-rrichter@marvell.com
All created csrow objects must be removed in the error path of
edac_create_csrow_objects(). The objects have been added as devices.
They need to be removed by doing a device_del() *and* put_device() call
to also free their memory. The missing put_device() leaves a memory
leak. Use device_unregister() instead of device_del() which properly
unregisters the device doing both.
Fixes: 7adc05d2dc ("EDAC/sysfs: Drop device references properly")
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Tested-by: John Garry <john.garry@huawei.com>
Cc: <stable@vger.kernel.org>
Link: https://lkml.kernel.org/r/20200212120340.4764-4-rrichter@marvell.com
A test kernel with the options DEBUG_TEST_DRIVER_REMOVE, KASAN and
DEBUG_KMEMLEAK set, revealed several issues when removing an mci device:
1) Use-after-free:
On 27.11.19 17:07:33, John Garry wrote:
> [ 22.104498] BUG: KASAN: use-after-free in
> edac_remove_sysfs_mci_device+0x148/0x180
The use-after-free is caused by the mci_for_each_dimm() macro called in
edac_remove_sysfs_mci_device(). The iterator was introduced with
c498afaf7d ("EDAC: Introduce an mci_for_each_dimm() iterator").
The iterator loop calls device_unregister(&dimm->dev), which removes
the sysfs entry of the device, but also frees the dimm struct in
dimm_attr_release(). When incrementing the loop in mci_for_each_dimm(),
the dimm struct is accessed again, after having been freed already.
The fix is to free all the mci device's subsequent dimm and csrow
objects at a later point, in _edac_mc_free(), when the mci device itself
is being freed.
This keeps the data structures intact and the mci device can be
fully used until its removal. The change allows the safe usage of
mci_for_each_dimm() to release dimm devices from sysfs.
2) Memory leaks:
Following memory leaks have been detected:
# grep edac /sys/kernel/debug/kmemleak | sort | uniq -c
1 [<000000003c0f58f9>] edac_mc_alloc+0x3bc/0x9d0 # mci->csrows
16 [<00000000bb932dc0>] edac_mc_alloc+0x49c/0x9d0 # csr->channels
16 [<00000000e2734dba>] edac_mc_alloc+0x518/0x9d0 # csr->channels[chn]
1 [<00000000eb040168>] edac_mc_alloc+0x5c8/0x9d0 # mci->dimms
34 [<00000000ef737c29>] ghes_edac_register+0x1c8/0x3f8 # see edac_mc_alloc()
All leaks are from memory allocated by edac_mc_alloc().
Note: The test above shows that edac_mc_alloc() was called here from
ghes_edac_register(), thus both functions show up in the stack trace
but the module causing the leaks is edac_mc. The comments with the data
structures involved were made manually by analyzing the objdump.
The data structures listed above and created by edac_mc_alloc() are
not properly removed during device removal, which is done in
edac_mc_free().
There are two paths implemented to remove the device depending on device
registration, _edac_mc_free() is called if the device is not registered
and edac_unregister_sysfs() otherwise.
The implemenations differ. For the sysfs case, the mci device removal
lacks the removal of subsequent data structures (csrows, channels,
dimms). This causes the memory leaks (see mci_attr_release()).
[ bp: Massage commit message. ]
Fixes: c498afaf7d ("EDAC: Introduce an mci_for_each_dimm() iterator")
Fixes: faa2ad09c0 ("edac_mc: edac_mc_free() cannot assume mem_ctl_info is registered in sysfs.")
Fixes: 7a623c0390 ("edac: rewrite the sysfs code to use struct device")
Reported-by: John Garry <john.garry@huawei.com>
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Tested-by: John Garry <john.garry@huawei.com>
Cc: <stable@vger.kernel.org>
Link: https://lkml.kernel.org/r/20200212120340.4764-3-rrichter@marvell.com
- remove ioremap_nocache given that is is equivalent to
ioremap everywhere
-----BEGIN PGP SIGNATURE-----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=TUCJ
-----END PGP SIGNATURE-----
Merge tag 'ioremap-5.6' of git://git.infradead.org/users/hch/ioremap
Pull ioremap updates from Christoph Hellwig:
"Remove the ioremap_nocache API (plus wrappers) that are always
identical to ioremap"
* tag 'ioremap-5.6' of git://git.infradead.org/users/hch/ioremap:
remove ioremap_nocache and devm_ioremap_nocache
MIPS: define ioremap_nocache to ioremap
Pull RAS updates from Borislav Petkov:
- Misc fixes to the MCE code all over the place, by Jan H. Schönherr.
- Initial support for AMD F19h and other cleanups to amd64_edac, by
Yazen Ghannam.
- Other small cleanups.
* 'ras-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
EDAC/mce_amd: Make fam_ops static global
EDAC/amd64: Drop some family checks for newer systems
EDAC/amd64: Add family ops for Family 19h Models 00h-0Fh
x86/amd_nb: Add Family 19h PCI IDs
EDAC/mce_amd: Always load on SMCA systems
x86/MCE/AMD, EDAC/mce_amd: Add new Load Store unit McaType
x86/mce: Fix use of uninitialized MCE message string
x86/mce: Fix mce=nobootlog
x86/mce: Take action on UCNA/Deferred errors again
x86/mce: Remove mce_inject_log() in favor of mce_log()
x86/mce: Pass MCE message to mce_panic() on failed kernel recovery
x86/mce/therm_throt: Mark throttle_active_work() as __maybe_unused
-----BEGIN PGP SIGNATURE-----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=fR3M
-----END PGP SIGNATURE-----
Merge tag 'edac_for_5.6' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras
Pull EDAC updates from Borislav Petkov:
"A totally boring branch this time around: a garden variety of small
fixes all over the place"
* tag 'edac_for_5.6' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras:
EDAC/amd64: Do not warn when removing instances
EDAC/sifive: Fix return value check in ecc_register()
EDAC/aspeed: Remove unneeded semicolon
EDAC: remove set but not used variable 'ecc_loc'
EDAC: skx_common: downgrade message importance on missing PCI device
EDAC/Kconfig: Fix Kconfig indentation
On machines which do not populate all nodes with DIMMs, the driver
doesn't initialize an instance there. However, the instance removal
remove_one_instance() path will warn unconditionally, which is wrong.
Remove the WARN_ON() even if the warning is innocent because it causes a
splat in dmesg.
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200117115939.5524-1-bp@alien8.de
In case of error, the function edac_device_alloc_ctl_info() returns a
NULL pointer, not ERR_PTR(). Replace the IS_ERR() test in the return
value check with a NULL test.
Fixes: 91abaeaaff ("EDAC/sifive: Add EDAC platform driver for SiFive SoCs")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200115150303.112627-1-weiyongjun1@huawei.com
In general, "pvt->umc != NULL" is used to check if the system is Family
17h+. However, there are a few places that are using direct family
checks.
Replace the remaining family checks with a check for "pvt->umc != NULL".
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200110015651.14887-6-Yazen.Ghannam@amd.com
Add family ops to support AMD Family 19h systems. Existing Family 17h
functions can be used. Also, add Family 19h to the list of families to
automatically load the module.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200110015651.14887-5-Yazen.Ghannam@amd.com
MCA error decoding on SMCA systems is not dependent on family. Return
success early if the system supports the SMCA feature.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200110015651.14887-3-Yazen.Ghannam@amd.com
Add support for a new version of the Load Store unit bank type as
indicated by its McaType value, which will be present in future SMCA
systems.
Add the new (HWID, MCATYPE) tuple. Reuse the same name, since this is
logically the same to the user.
Also, add the new error descriptions to edac_mce_amd.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200110015651.14887-2-Yazen.Ghannam@amd.com
The commit 9209fb5189 ("riscv: move sifive_l2_cache.c to drivers/soc")
moves the sifive L2 cache driver to driver/soc. It did not move the
header file along with the driver. Therefore this patch moves the header
file to driver/soc
Signed-off-by: Yash Shah <yash.shah@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
[paul.walmsley@sifive.com: updated to fix the include guard]
Fixes: 9209fb5189 ("riscv: move sifive_l2_cache.c to drivers/soc")
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
ioremap has provided non-cached semantics by default since the Linux 2.6
days, so remove the additional ioremap_nocache interface.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Arnd Bergmann <arnd@arndb.de>
The sifive_l2_cache.c is in no way related to RISC-V architecture
memory management. It is a little stub driver working around the fact
that the EDAC maintainers prefer their drivers to be structured in a
certain way that doesn't fit the SiFive SOCs.
Move the file to drivers/soc and add a Kconfig option for it, as well
as the whole drivers/soc boilerplate for CONFIG_SOC_SIFIVE.
Fixes: a967a289f1 ("RISC-V: sifive_l2_cache: Add L2 cache controller driver for SiFive SoCs")
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
[paul.walmsley@sifive.com: keep the MAINTAINERS change specific to the L2$ controller code]
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
Fixes gcc '-Wunused-but-set-variable' warning:
drivers/edac/i5100_edac.c: In function ‘i5100_read_log’:
drivers/edac/i5100_edac.c:489:11: warning: variable ‘ecc_loc’
set but not used [-Wunused-but-set-variable]
It is never used, and so can be removed.
Signed-off-by: yu kuai <yukuai3@huawei.com>
Reviewed-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Link: https://lore.kernel.org/r/20191216110121.46698-1-yukuai3@huawei.com
Both skx_edac and i10nm_edac drivers are loaded based on the matching CPU being
available which leads the module to be automatically loaded in virtual machines
as well. That will fail due the missing PCI devices. In both drivers the first
function to make use of the PCI devices is skx_get_hi_lo() will simply print
EDAC skx: Can't get tolm/tohm
for each CPU core, which is noisy. This patch makes it a debug message.
Signed-off-by: Aristeu Rozanski <aris@redhat.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Link: https://lore.kernel.org/r/20191204212325.c4k47p5hrnn3vpb5@redhat.com
Adjust indentation from spaces to tab (+optional two spaces) as in
coding style with a command like:
$ sed -e 's/^ /\t/' -i */Kconfig
[ bp: make it a single line. ]
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: James Morse <james.morse@arm.com>
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191120134206.15588-1-krzk@kernel.org
Simplify by using the Altera System Manager driver that abstracts the
differences between ARM32 and ARM64. Also allows the removal of the
Arria10 test function since this is handled by the System Manager
driver.
Signed-off-by: Thor Thayer <thor.thayer@linux.intel.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: James Morse <james.morse@arm.com>
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Meng.Li@windriver.com
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/1574361048-17572-4-git-send-email-thor.thayer@linux.intel.com
Cleanup the ECC Manager peripheral test in probe function as suggested
by James. Remove the check for Stratix10.
Suggested-by: James Morse <james.morse@arm.com>
Signed-off-by: Thor Thayer <thor.thayer@linux.intel.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/1573156890-26891-2-git-send-email-thor.thayer@linux.intel.com
When an IRQ occurs, regmap_{read,write,...}() is invoked in atomic
context. Regmap must indicate register IO is fast so that a spinlock is
used instead of a mutex to avoid sleeping in atomic context:
lock_acquire
__mutex_lock
mutex_lock_nested
regmap_lock_mutex
regmap_write
a10_eccmgr_irq_unmask
unmask_irq.part.0
irq_enable
__irq_startup
irq_startup
__setup_irq
request_threaded_irq
devm_request_threaded_irq
altr_sdram_probe
Mark it so.
[ bp: Massage. ]
Fixes: 3dab6bd526 ("EDAC, altera: Add support for Stratix10 SDRAM EDAC")
Reported-by: Meng Li <Meng.Li@windriver.com>
Signed-off-by: Meng Li <Meng.Li@windriver.com>
Signed-off-by: Thor Thayer <thor.thayer@linux.intel.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: James Morse <james.morse@arm.com>
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: stable <stable@vger.kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/1574361048-17572-2-git-send-email-thor.thayer@linux.intel.com
The following warning from the refcount framework is seen during ghes
initialization:
EDAC MC0: Giving out device to module ghes_edac.c controller ghes_edac: DEV ghes (INTERRUPT)
------------[ cut here ]------------
refcount_t: increment on 0; use-after-free.
WARNING: CPU: 36 PID: 1 at lib/refcount.c:156 refcount_inc_checked
[...]
Call trace:
refcount_inc_checked
ghes_edac_register
ghes_probe
...
It warns if the refcount is incremented from zero. This warning is
reasonable as a kernel object is typically created with a refcount of
one and freed once the refcount is zero. Afterwards the object would be
"used-after-free".
For GHES, the refcount is initialized with zero, and that is why this
message is seen when initializing the first instance. However, whenever
the refcount is zero, the device will be allocated and registered. Since
the ghes_reg_mutex protects the refcount and serializes allocation and
freeing of ghes devices, a use-after-free cannot happen here.
Instead of using refcount_inc() for the first instance, use
refcount_set(). This can be used here because the refcount is zero at
this point and can not change due to its protection by the mutex.
Fixes: 23f61b9fc5 ("EDAC/ghes: Fix locking and memory barrier issues")
Reported-by: John Garry <john.garry@huawei.com>
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Tested-by: John Garry <john.garry@huawei.com>
Cc: <huangming23@huawei.com>
Cc: James Morse <james.morse@arm.com>
Cc: <linuxarm@huawei.com>
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: <tanxiaofei@huawei.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: <wanghuiqiang@huawei.com>
Link: https://lkml.kernel.org/r/20191121213628.21244-1-rrichter@marvell.com
The code in ghes_edac.c and edac_mc.c for grain_bits calculation and
calling trace_mc_event() is now the same. Move it to a single location
in edac_raw_mc_handle_error().
The only difference is the missing IS_ENABLED(CONFIG_RAS) switch, but
this is needed for ghes too.
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106093239.25517-13-rrichter@marvell.com
detail_location[] is used to collect two location strings so they can
be passed as one to trace_mc_event(). Instead of having an extra copy
step, assemble the location string in other_detail[] from the
beginning.
Using other_detail[] to call trace_mc_event() is now the same as in
edac_mc.c and code can be unified.
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: James Morse <james.morse@arm.com>
Reviewed-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106093239.25517-12-rrichter@marvell.com
The current code to convert a physical address mask to a grain
(defined as granularity in bytes) is:
e->grain = ~(mem_err->physical_addr_mask & ~PAGE_MASK);
This is broken in several ways:
1) It calculates to wrong grain values. E.g., a physical address mask
of ~0xfff should give a grain of 0x1000. Without considering
PAGE_MASK, there is an off-by-one. Things are worse when also
filtering it with ~PAGE_MASK. This will calculate to a grain with the
upper bits set. In the example it even calculates to ~0.
2) The grain does not depend on and is unrelated to the kernel's
page-size. The page-size only matters when unmapping memory in
memory_failure(). Smaller grains are wrongly rounded up to the
page-size, on architectures with a configurable page-size (e.g. arm64)
this could round up to the even bigger page-size of the hypervisor.
Fix this with:
e->grain = ~mem_err->physical_addr_mask + 1;
The grain_bits are defined as:
grain = 1 << grain_bits;
Change also the grain_bits calculation accordingly, it is the same
formula as in edac_mc.c now and the code can be unified.
The value in ->physical_addr_mask coming from firmware is assumed to
be contiguous, but this is not sanity-checked. However, in case the
mask is non-contiguous, a conversion to grain_bits effectively
converts the grain bit mask to a power of 2 by rounding it up.
Suggested-by: James Morse <james.morse@arm.com>
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106093239.25517-11-rrichter@marvell.com
The e string to which this is pointing to has already been cleared
earlier in the function so remove the needless zero string termination.
[ bp: Correct the commit message. ]
Suggested-by: Joe Perches <joe@perches.com>
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106093239.25517-6-rrichter@marvell.com
No need to crash the system in case edac_mc_alloc() is called with
invalid arguments, just warn and return. This would cause a checkpatch
warning when touching the code later, so just fix it.
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106093239.25517-5-rrichter@marvell.com
Introduce an mci_for_each_dimm() iterator. It returns a pointer to
a struct dimm_info. This makes the declaration and use of an index
obsolete and avoids access to internal data of struct mci (direct array
access etc).
[ bp: push the struct dimm_info *dimm; declaration into the
CONFIG_EDAC_DEBUG block. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106093239.25517-4-rrichter@marvell.com
The EDAC_DIMM_OFF() macro takes 5 arguments to get the DIMM's index.
Simplify this by storing the index in struct dimm_info to avoid its
calculation and remove the EDAC_DIMM_OFF() macro. The index can be
directly used then.
Another advantage is that edac_mc_alloc() could be used even if the
exact size of the layers is unknown. Only the number of DIMMs would be
needed.
Rename iterator variable to idx, while at it. The name is more handy,
esp. when searching for it in the code.
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106093239.25517-3-rrichter@marvell.com
The EDAC_DIMM_PTR() macro takes 3 arguments from struct mem_ctl_info.
Clean up this interface to only pass the mci struct and replace this
macro with a new function edac_get_dimm().
Also introduce an edac_get_dimm_by_index() function for later use.
This allows it to get a DIMM pointer only by a given index. This can
be useful if the DIMM's position within the layers of the memory
controller or the exact size of the layers are unknown.
Small style changes made for some hunks after applying the semantic
patch.
Semantic patch used:
@@ expression mci, a, b,c; @@
-EDAC_DIMM_PTR(mci->layers, mci->dimms, mci->n_layers, a, b, c)
+edac_get_dimm(mci, a, b, c)
[ bp: Touchups. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Jason Baron <jbaron@akamai.com>
Cc: Qiuxu Zhuo <qiuxu.zhuo@intel.com>
Cc: Tero Kristo <t-kristo@ti.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106093239.25517-2-rrichter@marvell.com
This message keeps flooding dmesg on boxes where ECC is disabled or the
DIMMs do not support ECC but the module gets auto-probed. What's even
worse is that autoprobing happens on every CPU due to the CPU-family
matching the driver does and uevent being generated for each CPU device.
What is more, this message is becoming even more useless on newer
systems where forcing ECC is not recommended and it should be done in
the BIOS so the BIOS can do all the necessary work, i.e., just setting a
bit in an MSR is not enough anymore.
So get rid of it.
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: Yazen Ghannam <yazen.ghannam@amd.com>
Cc: linux-edac@vger.kernel.org
Link: https://lkml.kernel.org/r/20191106160607.GC28380@zn.tnic
The ghes registration and refcount is broken in several ways:
* ghes_edac_register() returns with success for a 2nd instance
even if a first instance's registration is still running. This is
not correct as the first instance may fail later. A subsequent
registration may not finish before the first. Parallel registrations
must be avoided.
* The refcount was increased even if a registration failed. This
leads to stale counters preventing the device from being released.
* The ghes refcount may not be decremented properly on unregistration.
Always decrement the refcount once ghes_edac_unregister() is called to
keep the refcount sane.
* The ghes_pvt pointer is handed to the irq handler before registration
finished.
* The mci structure could be freed while the irq handler is running.
Fix this by adding a mutex to ghes_edac_register(). This mutex
serializes instances to register and unregister. The refcount is only
increased if the registration succeeded. This makes sure the refcount is
in a consistent state after registering or unregistering a device.
Note: A spinlock cannot be used here as the code section may sleep.
The ghes_pvt is protected by ghes_lock now. This ensures the pointer is
not updated before registration was finished or while the irq handler is
running. It is unset before unregistering the device including necessary
(implicit) memory barriers making the changes visible to other CPUs.
Thus, the device can not be used anymore by an interrupt.
Also, rename ghes_init to ghes_refcount for better readability and
switch to refcount API.
A refcount is needed because there can be multiple GHES structures being
defined (see ACPI 6.3 specification, 18.3.2.7 Generic Hardware Error
Source, "Some platforms may describe multiple Generic Hardware Error
Source structures with different notification types, ...").
Another approach to use the mci's device refcount (get_device()) and
have a release function does not work here. A release function will be
called only for device_release() with the last put_device() call. The
device must be deleted *before* that with device_del(). This is only
possible by maintaining an own refcount.
[ bp: touchups. ]
Fixes: 0fe5f281f7 ("EDAC, ghes: Model a single, logical memory controller")
Fixes: 1e72e673b9 ("EDAC/ghes: Fix Use after free in ghes_edac remove path")
Co-developed-by: James Morse <james.morse@arm.com>
Signed-off-by: James Morse <james.morse@arm.com>
Co-developed-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191105200732.3053-1-rrichter@marvell.com
Return early before checking for ECC if the node does not have any
populated memory.
Free any cached hardware data before returning. Also, return 0 in this
case since this is not a failure. Other nodes may have memory and the
module should attempt to load an instance for them.
Move printing of hardware information to after the instance is
initialized, so that the information is only printed for nodes with
memory.
Return an error code when ECC is disabled. This check happens after
checking for memory. The module should explicitly fail to load if memory
is populated on a node and ECC is disabled.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106012448.243970-6-Yazen.Ghannam@amd.com
...now that the data is available earlier.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106012448.243970-5-Yazen.Ghannam@amd.com
The maximum number of memory controllers is fixed within a family/model
group. In most cases, this has been fixed at 2, but some systems may
have up to 8.
The struct amd64_family_type already contains family/model-specific
information, and this can be used rather than adding model checks to
various functions.
Create a new field in struct amd64_family_type for max_mcs.
Set this when setting other family type information, and use this when
needing the maximum number of memory controllers possible for a system.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106012448.243970-4-Yazen.Ghannam@amd.com
Split out gathering hardware information from init_one_instance()
into a separate function hw_info_get(). This is necessary so that
the information can be cached earlier and used to check if memory is
populated and if ECC is enabled on a node.
Also, define a function hw_info_put() to back out changes made in
hw_info_get().
Check for an allocated PCI device (Function 0 for Family 17h or Function
1 for pre-Family 17h) before freeing, since hw_info_put() may be called
before PCI siblings are reserved.
Drop the family check when freeing pvt->umc. This will be NULL on
pre-Family 17h systems. However, kfree() is safe and will check for a
NULL pointer before freeing.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106012448.243970-3-Yazen.Ghannam@amd.com
The struct amd64_family_type doesn't change between multiple nodes and
instances of the module, so make it global.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191106012448.243970-2-Yazen.Ghannam@amd.com
The following commit introduced a warning on error reports without a
non-zero grain value.
3724ace582 ("EDAC/mc: Fix grain_bits calculation")
The amd64_edac_mod module does not provide a value, so the warning will
be given on the first reported memory error.
Set the grain per DIMM to cacheline size (64 bytes). This is the current
recommendation.
Fixes: 3724ace582 ("EDAC/mc: Fix grain_bits calculation")
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20191022203448.13962-7-Yazen.Ghannam@amd.com
Simplify this function implementation by using a known wrapper function.
This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Joel Stanley <joel@jms.id.au>
Cc: Andrew Jeffery <andrew@aj.id.au>
Cc: James Morse <james.morse@arm.com>
Cc: kernel-janitors@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-aspeed@lists.ozlabs.org
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Stefan Schaeckeler <sschaeck@cisco.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/baabb9e9-a1b2-3a04-9fb6-aa632de5f722@web.de
Skylake logs some additional useful information in per-channel
registers in addition the the architectural status/addr/misc
logged in the machine check bank.
Pick up this information and add it to the EDAC log:
retry_rd_err_[five 32-bit register values]
Sorry, no definitions for these registers. OEMs and DIMM vendors
will be able to use them to isolate which cells in the DIMM are
causing problems.
correrrcnt[per rank corrected error counts]
Note that if additional errors are logged while these registers are
being read, you may see a jumble of values some from earlier errors,
others from later errors (since the registers report the most recent
logged error). The correrrcnt registers provide error counts per possible
rank. If these counts only change by one since the previous error logged
for this channel, then it is safe to assume that the registers logged
provide a coherent view of one error.
With this change EDAC logs look like this:
EDAC MC4: 1 CE memory read error on CPU_SrcID#2_MC#0_Chan#1_DIMM#0 (channel:1 slot:0 page:0x8f26018 offset:0x0 grain:32 syndrome:0x0 - err_code:0x0101:0x0091 socket:2 imc:0 rank:0 bg:0 ba:0 row:0x1f880 col:0x200 retry_rd_err_log[0001a209 00000000 00000001 04800001 0001f880] correrrcnt[0001 0000 0000 0000 0000 0000 0000 0000])
Acked-by: Aristeu Rozanski <aris@redhat.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
ghes_edac models a single logical memory controller, and uses a global
ghes_init variable to ensure only the first ghes_edac_register() will
do anything.
ghes_edac is registered the first time a GHES entry in the HEST is
probed. There may be multiple entries, so subsequent attempts to
register ghes_edac are silently ignored as the work has already been
done.
When a GHES entry is unregistered, it calls ghes_edac_unregister(),
which free()s the memory behind the global variables in ghes_edac.
But there may be multiple GHES entries, the next call to
ghes_edac_unregister() will dereference the free()d memory, and attempt
to free it a second time.
This may also be triggered on a platform with one GHES entry, if the
driver is unbound/re-bound and unbound. The re-bind step will do
nothing because of ghes_init, the second unbind will then do the same
work as the first.
Doing the unregister work on the first call is unsafe, as another
CPU may be processing a notification in ghes_edac_report_mem_error(),
using the memory we are about to free.
ghes_init is already half of the reference counting. We only need
to do the register work for the first call, and the unregister work
for the last. Add the unregister check.
This means we no longer free ghes_edac's memory while there are
GHES entries that may receive a notification.
This was detected by KASAN and DEBUG_TEST_DRIVER_REMOVE.
[ bp: merge into a single patch. ]
Fixes: 0fe5f281f7 ("EDAC, ghes: Model a single, logical memory controller")
Reported-by: John Garry <john.garry@huawei.com>
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: <stable@vger.kernel.org>
Link: https://lkml.kernel.org/r/20191014171919.85044-2-james.morse@arm.com
Link: https://lkml.kernel.org/r/304df85b-8b56-b77e-1a11-aa23769f2e7c@huawei.com
Make the main workhorse the "count" functions which can log a @count
of errors. Have the current APIs edac_device_handle_{ce,ue}() call
the _count() variants and this way keep the exported symbols number
unchanged.
[ bp: Rewrite. ]
Signed-off-by: Hanna Hawa <hhhawa@amazon.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: benh@amazon.com
Cc: dwmw@amazon.co.uk
Cc: hanochu@amazon.com
Cc: James Morse <james.morse@arm.com>
Cc: jonnyc@amazon.com
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: ronenk@amazon.com
Cc: talel@amazon.com
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190923191741.29322-2-hhhawa@amazon.com
drivers/edac/skx_common.c: In function ‘skx_mce_output_error’:
drivers/edac/skx_common.c:478:8: warning: variable ‘type’ set but not used [-Wunused-but-set-variable]
478 | char *type, *optype;
| ^~~~
Acked-by: Borislav Petkov <bp@alien8.de>
Acked-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
There are several vars unused on this driver, probably because
it was a modified copy of another driver. Get rid of them.
drivers/edac/sb_edac.c: In function ‘knl_get_dimm_capacity’:
drivers/edac/sb_edac.c:1343:16: warning: variable ‘sad_size’ set but not used [-Wunused-but-set-variable]
1343 | u64 sad_base, sad_size, sad_limit = 0;
| ^~~~~~~~
drivers/edac/sb_edac.c: In function ‘sbridge_mce_output_error’:
drivers/edac/sb_edac.c:2955:8: warning: variable ‘type’ set but not used [-Wunused-but-set-variable]
2955 | char *type, *optype, msg[256];
| ^~~~
drivers/edac/sb_edac.c: In function ‘sbridge_unregister_mci’:
drivers/edac/sb_edac.c:3203:22: warning: variable ‘pvt’ set but not used [-Wunused-but-set-variable]
3203 | struct sbridge_pvt *pvt;
| ^~~
At top level:
drivers/edac/sb_edac.c:266:18: warning: ‘correrrthrsld’ defined but not used [-Wunused-const-variable=]
266 | static const u32 correrrthrsld[] = {
| ^~~~~~~~~~~~~
drivers/edac/sb_edac.c:257:18: warning: ‘correrrcnt’ defined but not used [-Wunused-const-variable=]
257 | static const u32 correrrcnt[] = {
| ^~~~~~~~~~
Acked-by: Borislav Petkov <bp@alien8.de>
Acked-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
There are several temporary unused vars:
drivers/edac/i5400_edac.c: In function ‘i5400_get_mc_regs’:
drivers/edac/i5400_edac.c:1058:6: warning: variable ‘maxdimmperch’ set but not used [-Wunused-but-set-variable]
1058 | int maxdimmperch;
| ^~~~~~~~~~~~
drivers/edac/i5400_edac.c:1057:6: warning: variable ‘maxch’ set but not used [-Wunused-but-set-variable]
1057 | int maxch;
| ^~~~~
drivers/edac/i5400_edac.c: In function ‘i5400_init_dimms’:
drivers/edac/i5400_edac.c:1174:6: warning: variable ‘max_dimms’ set but not used [-Wunused-but-set-variable]
1174 | int max_dimms;
| ^~~~~~~~~
drivers/edac/i5400_edac.c:1173:14: warning: variable ‘channel_count’ set but not used [-Wunused-but-set-variable]
1173 | int ndimms, channel_count;
| ^~~~~~~~~~~~~
Get rid of them.
Acked-by: Borislav Petkov <bp@alien8.de>
Acked-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
There are 3 types of non-recoverable errors that the MC reports:
- Fatal;
- Non-fatal uncorrected
- Non-fatal correctable
While we don't add it to the log itself, it could be useful to
have this at least for debug messages.
This shuts up this warning:
drivers/edac/i5400_edac.c: In function ‘i5400_proccess_non_recoverable_info’:
drivers/edac/i5400_edac.c:524:8: warning: variable ‘type’ set but not used [-Wunused-but-set-variable]
524 | char *type = NULL;
| ^~~~
Acked-by: Borislav Petkov <bp@alien8.de>
Acked-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The declaration of the kerneldoc entry is wrong, causing this
warning:
drivers/edac/i7300_edac.c:824: warning: Function parameter or member 'mir_no' not described in 'decode_mir'
Acked-by: Borislav Petkov <bp@alien8.de>
Acked-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
One var was renamed, but the associated kernel-doc markup still
points to the old name.
Acked-by: Borislav Petkov <bp@alien8.de>
Acked-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
As reported by GCC with W=1:
drivers/edac/i5100_edac.c:714:16: warning: variable ‘et’ set but not used [-Wunused-but-set-variable]
714 | unsigned long et;
| ^~
It sounds some left over from some code before the addition of
an udelay().
Acked-by: Borislav Petkov <bp@alien8.de>
Acked-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
- fix various clang build and cppcheck issues
- switch ARM to use new common outgoing-CPU-notification code
- add some additional explanation about the boot code
- kbuild "make clean" fixes
- get rid of another "(____ptrval____)", this time for the VDSO code
- avoid treating cache maintenance faults as a write
- add a frame pointer unwinder implementation for clang
- add EDAC support for Aurora L2 cache
- improve robustness of adjust_lowmem_bounds() finding the bounds of
lowmem.
- add reset control for AMBA primecell devices
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=Evmc
-----END PGP SIGNATURE-----
Merge tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm
Pull ARM updates from Russell King:
- fix various clang build and cppcheck issues
- switch ARM to use new common outgoing-CPU-notification code
- add some additional explanation about the boot code
- kbuild "make clean" fixes
- get rid of another "(____ptrval____)", this time for the VDSO code
- avoid treating cache maintenance faults as a write
- add a frame pointer unwinder implementation for clang
- add EDAC support for Aurora L2 cache
- improve robustness of adjust_lowmem_bounds() finding the bounds of
lowmem.
- add reset control for AMBA primecell devices
* tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm: (24 commits)
ARM: 8906/1: drivers/amba: add reset control to amba bus probe
ARM: 8905/1: Emit __gnu_mcount_nc when using Clang 10.0.0 or newer
ARM: 8904/1: skip nomap memblocks while finding the lowmem/highmem boundary
ARM: 8903/1: ensure that usable memory in bank 0 starts from a PMD-aligned address
ARM: 8891/1: EDAC: armada_xp: Add support for more SoCs
ARM: 8888/1: EDAC: Add driver for the Marvell Armada XP SDRAM and L2 cache ECC
ARM: 8892/1: EDAC: Add missing debugfs_create_x32 wrapper
ARM: 8890/1: l2x0: add marvell,ecc-enable property for aurora
ARM: 8889/1: dt-bindings: document marvell,ecc-enable binding
ARM: 8886/1: l2x0: support parity-enable/disable on aurora
ARM: 8885/1: aurora-l2: add defines for parity and ECC registers
ARM: 8887/1: aurora-l2: add prefix to MAX_RANGE_SIZE
ARM: 8902/1: l2c: move cache-aurora-l2.h to asm/hardware
ARM: 8900/1: UNWINDER_FRAME_POINTER implementation for Clang
ARM: 8898/1: mm: Don't treat faults reported from cache maintenance as writes
ARM: 8896/1: VDSO: Don't leak kernel addresses
ARM: 8895/1: visit mach-* and plat-* directories when cleaning
ARM: 8894/1: boot: Replace open-coded nop with macro
ARM: 8893/1: boot: Explain the 8 nops
ARM: 8876/1: fix O= building with CONFIG_FPE_FASTFPE
...
Pull x86 cpu-feature updates from Ingo Molnar:
- Rework the Intel model names symbols/macros, which were decades of
ad-hoc extensions and added random noise. It's now a coherent, easy
to follow nomenclature.
- Add new Intel CPU model IDs:
- "Tiger Lake" desktop and mobile models
- "Elkhart Lake" model ID
- and the "Lightning Mountain" variant of Airmont, plus support code
- Add the new AVX512_VP2INTERSECT instruction to cpufeatures
- Remove Intel MPX user-visible APIs and the self-tests, because the
toolchain (gcc) is not supporting it going forward. This is the
first, lowest-risk phase of MPX removal.
- Remove X86_FEATURE_MFENCE_RDTSC
- Various smaller cleanups and fixes
* 'x86-cpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (25 commits)
x86/cpu: Update init data for new Airmont CPU model
x86/cpu: Add new Airmont variant to Intel family
x86/cpu: Add Elkhart Lake to Intel family
x86/cpu: Add Tiger Lake to Intel family
x86: Correct misc typos
x86/intel: Add common OPTDIFFs
x86/intel: Aggregate microserver naming
x86/intel: Aggregate big core graphics naming
x86/intel: Aggregate big core mobile naming
x86/intel: Aggregate big core client naming
x86/cpufeature: Explain the macro duplication
x86/ftrace: Remove mcount() declaration
x86/PCI: Remove superfluous returns from void functions
x86/msr-index: Move AMD MSRs where they belong
x86/cpu: Use constant definitions for CPU models
lib: Remove redundant ftrace flag removal
x86/crash: Remove unnecessary comparison
x86/bitops: Use __builtin_constant_p() directly instead of IS_IMMEDIATE()
x86: Remove X86_FEATURE_MFENCE_RDTSC
x86/mpx: Remove MPX APIs
...
Add the new Family 17h Model 70h PCI IDs (device 18h functions 0 and 6)
to the AMD64 EDAC module.
[ bp: s/f17_base_addr_to_cs_size/f17_addr_mask_to_cs_size/g ]
Signed-off-by: Isaac Vaughn <isaac.vaughn@knights.ucf.edu>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: James Morse <james.morse@arm.com>
Cc: linux-edac@vger.kernel.org
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Robert Richter <rrichter@marvell.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190906192131.8ced0ca112146f32d82b6cae@knights.ucf.edu
Debug messages are inconsistently used in the error handlers. Some lack
an error message, some are called regardless of the return status,
messages for the same error are at different locations in the code
depending on the error code. This happens esp. near put_device() calls.
Make those debug messages more consistent. Additionally, unify the error
messages to have the same terms for the same operations of the device.
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190902123216.9809-5-rrichter@marvell.com
Use of 'unsigned int' instead of bare use of 'unsigned'. Fix this for
edac_mc*, ghes and the i5100 driver as reported by checkpatch.pl.
While at it, struct member dev_ch_attribute->channel is always used as
unsigned int. Change type to unsigned int to avoid type casts.
[ bp: Massage. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190902123216.9809-2-rrichter@marvell.com
The Armada 38x and other integrated SoCs use a reduced pin count so the
width of the SDRAM interface is smaller than the Armada XP SoCs. This
means that the definition of "full" and "half" width is reduced from
64/32 to 32/16.
Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Add support for the ECC functionality as found in the DDR RAM and L2
cache controllers on the MV78230/MV78x60 SoCs. This driver has been
tested on the MV78460 (on a custom board with a DDR3 ECC DIMM).
[cp use SPDX license]
Signed-off-by: Jan Luebbe <jlu@pengutronix.de>
Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Reviewed-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
We already have wrappers for x8 and x16, so add the missing x32 one.
Signed-off-by: Jan Luebbe <jlu@pengutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Currently big microservers have _XEON_D while small microservers have
_X, Make it uniformly: _D.
for i in `git grep -l "\(INTEL_FAM6_\|VULNWL_INTEL\|INTEL_CPU_FAM6\).*_\(X\|XEON_D\)"`
do
sed -i -e 's/\(\(INTEL_FAM6_\|VULNWL_INTEL\|INTEL_CPU_FAM6\).*ATOM.*\)_X/\1_D/g' \
-e 's/\(\(INTEL_FAM6_\|VULNWL_INTEL\|INTEL_CPU_FAM6\).*\)_XEON_D/\1_D/g' ${i}
done
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Tony Luck <tony.luck@intel.com>
Cc: x86@kernel.org
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Borislav Petkov <bp@alien8.de>
Link: https://lkml.kernel.org/r/20190827195122.677152989@infradead.org
Future AMD systems will support asymmetric dual-rank DIMMs. These are
DIMMs where the ranks are of different sizes.
The even rank will use the Primary Even Chip Select registers and the
odd rank will use the Secondary Odd Chip Select registers.
Recognize if a Secondary Odd Chip Select is being used. Use the
Secondary Odd Address Mask when calculating the chip select size.
[ bp: move csrow_sec_enabled() to the header, fix CS_ODD define and
tone-down the capitalized words spelling. ]
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190821235938.118710-8-Yazen.Ghannam@amd.com
AMD Family 17h systems have a set of secondary Chip Select Base
Addresses and Address Masks. These do not represent unique Chip
Selects, rather they are used in conjunction with the primary
Chip Select registers in certain cases.
Cache these secondary Chip Select registers for future use.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190821235938.118710-7-Yazen.Ghannam@amd.com
AMD Family 17h systems currently require address translation in order to
report the system address of a DRAM ECC error. This is currently done
before decoding the syndrome information. The syndrome information does
not depend on the address translation, so the proper EDAC csrow/channel
reporting can function without the address. However, the syndrome
information will not be decoded if the address translation fails.
Decode the syndrome information before doing the address translation.
The syndrome information is architecturally defined in MCA_SYND and can
be considered robust. The address translation is system-specific and may
fail on newer systems without proper updates to the translation
algorithm.
Fixes: 713ad54675 ("EDAC, amd64: Define and register UMC error decode function")
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190821235938.118710-6-Yazen.Ghannam@amd.com
Chip Select memory size reporting on AMD Family 17h was recently fixed
in order to account for interleaving. However, the current method is not
robust.
The Chip Select Address Mask can be used to find the memory size. There
are a couple of cases.
1) For single-rank and dual-rank non-interleaved, use the address mask
plus 1 as the size.
2) For dual-rank interleaved, do #1 but "de-interleave" the address mask
first.
Always "de-interleave" the address mask in order to simplify the code
flow. Bit mask manipulation is necessary to check for interleaving, so
just go ahead and do the de-interleaving. In the non-interleaved case,
the original and de-interleaved address masks will be the same.
To de-interleave the mask, count the number of zero bits in the middle
of the mask and swap them with the most significant bits.
For example,
Original=0xFFFF9FE, De-interleaved=0x3FFFFFE
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190821235938.118710-5-Yazen.Ghannam@amd.com
Currently, the DIMM info for AMD Family 17h systems is initialized in
init_csrows(). This function is shared with legacy systems, and it has a
limit of two channel support.
This prevents initialization of the DIMM info for a number of ranks, so
there will be missing ranks in the EDAC sysfs.
Create a new init_csrows_df() for Family17h+ and revert init_csrows()
back to pre-Family17h support.
Loop over all channels in the new function in order to support systems
with more than two channels.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190821235938.118710-4-Yazen.Ghannam@amd.com
AMD Family 17h systems support x4 and x16 DRAM devices. However, the
device type is not checked when setting mci.edac_ctl_cap.
Set the appropriate capability flag based on the device type.
Default to x8 DRAM device when neither the x4 or x16 bits are set.
[ bp: reverse cpk_en check to save an indentation level. ]
Fixes: 2d09d8f301 ("EDAC, amd64: Determine EDAC MC capabilities on Fam17h")
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190821235938.118710-3-Yazen.Ghannam@amd.com
The struct chip_select array that's used for saving chip select bases
and masks is fixed at length of two. There should be one struct
chip_select for each controller, so this array should be increased to
support systems that may have more than two controllers.
Increase the size of the struct chip_select array to eight, which is the
largest number of controllers per die currently supported on AMD
systems.
Fix number of DIMMs and Chip Select bases/masks on Family17h, because
AMD Family 17h systems support 2 DIMMs, 4 CS bases, and 2 CS masks per
channel.
Also, carve out the Family 17h+ reading of the bases/masks into a
separate function. This effectively reverts the original bases/masks
reading code to before Family 17h support was added.
Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190821235938.118710-2-Yazen.Ghannam@amd.com
Depending on how BIOS has marked the reserved region containing the 32KB
MCHBAR you can get warnings like:
resource sanity check: requesting [mem 0xfed10000-0xfed1ffff], which spans more than reserved [mem 0xfed10000-0xfed17fff]
caller dnv_rd_reg+0xc8/0x240 [pnd2_edac] mapping multiple BARs
Not all of the mmio regions used in dnv_rd_reg() are the same size. The
MCHBAR window is 32KB and the sideband ports are 64KB. Pass the correct
size to ioremap() depending on which resource we're reading from.
Signed-off-by: Stephen Douthit <stephend@silicom-usa.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Add ECC support for Mellanox BlueField SoC DDR controller.
This requires SMC to the running Arm Trusted Firmware to report
what is the current memory configuration.
Reviewed-by: James Morse <james.morse@arm.com>
Signed-off-by: Shravan Kumar Ramani <sramani@mellanox.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Smatch complains about the cast of a u32 pointer to unsigned long:
drivers/edac/altera_edac.c:1878 altr_edac_a10_irq_handler()
warn: passing casted pointer '&irq_status' to 'find_first_bit()'
This code wouldn't work on a 64 bit big endian system because it would
read past the end of &irq_status.
[ bp: massage. ]
Fixes: 13ab8448d2 ("EDAC, altera: Add ECC Manager IRQ controller support")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Thor Thayer <thor.thayer@linux.intel.com>
Cc: James Morse <james.morse@arm.com>
Cc: kernel-janitors@vger.kernel.org
Cc: linux-edac <linux-edac@vger.kernel.org>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190624134717.GA1754@mwanda
The grain in EDAC is defined as "minimum granularity for an error
report, in bytes". The following calculation of the grain_bits in
edac_mc is wrong:
grain_bits = fls_long(e->grain) + 1;
Where grain_bits is defined as:
grain = 1 << grain_bits
Example:
grain = 8 # 64 bit (8 bytes)
grain_bits = fls_long(8) + 1
grain_bits = 4 + 1 = 5
grain = 1 << grain_bits
grain = 1 << 5 = 32
Replace it with the correct calculation:
grain_bits = fls_long(e->grain - 1);
The example gives now:
grain_bits = fls_long(8 - 1)
grain_bits = fls_long(7)
grain_bits = 3
grain = 1 << 3 = 8
Also, check if the hardware reports a reasonable grain != 0 and fallback
with a warning to 1 byte granularity otherwise.
[ bp: massage a bit. ]
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: "linux-edac@vger.kernel.org" <linux-edac@vger.kernel.org>
Cc: James Morse <james.morse@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Tony Luck <tony.luck@intel.com>
Link: https://lkml.kernel.org/r/20190624150758.6695-2-rrichter@marvell.com
