- Nov 04, 2021
-
-
Luis Chamberlain authored
Now that device_add_disk() supports returning an error, use that. We must unwind alloc_dax() on error. Reviewed-by:
Christoph Hellwig <hch@lst.de> Signed-off-by:
Luis Chamberlain <mcgrof@kernel.org> Link: https://lore.kernel.org/r/20211103230437.1639990-7-mcgrof@kernel.org Signed-off-by:
Jens Axboe <axboe@kernel.dk>
-
Luis Chamberlain authored
Prior to devm being able to use pmem_release_disk() there are other failure which can occur for which we must account for and release the disk for. Address those few cases. Fixes: 3dd60fb9 ("nvdimm/pmem: stop using q_usage_count as external pgmap refcount") Reviewed-by:
-
Luis Chamberlain authored
We never checked for errors on add_disk() as this function returned void. Now that this is fixed, use the shiny new error handling. Since nvdimm/blk uses devm we just need to move the devm registration towards the end. And in hindsight, that seems to also provide a fix given del_gendisk() should not be called unless the disk was already added via add_disk(). The probably of that issue happening is low though, like OOM while calling devm_add_action(), so the fix is minor. We manually unwind in case of add_disk() failure prior to the devm registration. Reviewed-by:
Dan Williams <dan.j.williams@intel.com> Reviewed-by:
-
Luis Chamberlain authored
If nd_integrity_init() fails we'd get del_gendisk() called, but that's not correct as we should only call that if we're done with device_add_disk(). Fix this by providing unwinding prior to the devm call being registered and moving the devm registration to the very end. This should fix calling del_gendisk() if nd_integrity_init() fails. I only spotted this issue through code inspection. It does not fix any real world bug. Reviewed-by:
-
Luis Chamberlain authored
We never checked for errors on add_disk() as this function returned void. Now that this is fixed, use the shiny new error handling. Reviewed-by:
-
Luis Chamberlain authored
This will make it easier to share common error paths. Reviewed-by:
-
- Nov 03, 2021
-
-
Luis Chamberlain authored
del_gendisk() should not called if the disk has not been added. Fix this. Fixes: 41cd8b70 ("libnvdimm, btt: add support for blk integrity") Reviewed-by:
-
- Oct 25, 2021
-
-
Christoph Hellwig authored
Originally all DAX access when through block_device operations and thus needed a queue reference. But since commit cccbce67 ("filesystem-dax: convert to dax_direct_access()") all this happens at the DAX device level which uses its own refcounting. Having the external refcount thus wasn't needed but has otherwise been harmless for long time. But now that "block: drain file system I/O on del_gendisk" waits for q_usage_count to reach 0 in del_gendisk this whole scheme can't work anymore (and pmem is the only driver abusing q_usage_count like that). So switch to the internal reference and remove the unbalanced blk_freeze_queue_start that is taken care of by del_gendisk. Fixes: 8e141f9e ("block: drain file system I/O on del_gendisk") Reported-by:
Yi Zhang <yi.zhang@redhat.com> Signed-off-by:
-
- Oct 18, 2021
-
-
Christoph Hellwig authored
Replace the blk_poll interface that requires the caller to keep a queue and cookie from the submissions with polling based on the bio. Polling for the bio itself leads to a few advantages: - the cookie construction can made entirely private in blk-mq.c - the caller does not need to remember the request_queue and cookie separately and thus sidesteps their lifetime issues - keeping the device and the cookie inside the bio allows to trivially support polling BIOs remapping by stacking drivers - a lot of code to propagate the cookie back up the submission path can be removed entirely. Signed-off-by:
Mark Wunderlich <mark.wunderlich@intel.com> Link: https://lore.kernel.org/r/20211012111226.760968-15-hch@lst.de Signed-off-by:
-
Christoph Hellwig authored
Split the integrity/metadata handling definitions out into a new header. Signed-off-by:
Johannes Thumshirn <johannes.thumshirn@wdc.com> Link: https://lore.kernel.org/r/20210920123328.1399408-17-hch@lst.de Signed-off-by:
-
- Sep 27, 2021
-
-
Christoph Hellwig authored
The recent block layer refactoring broke the way how the pmem driver abused device_add_disk. Fix this by properly passing the attribute groups to device_add_disk. Fixes: 52b85909 ("block: fold register_disk into device_add_disk") Signed-off-by:
Ira Weiny <ira.weiny@intel.com> Reviewed-by:
Darrick J. Wong <djwong@kernel.org> Link: https://lore.kernel.org/r/20210922173431.2454024-2-hch@lst.de Signed-off-by:
-
Christoph Hellwig authored
dax_attribute_group is only used by the pmem driver, and can avoid the completely pointless lookup by the disk name if moved there. This leaves just a single caller of dax_get_by_host, so move dax_get_by_host into the same ifdef block as that caller. Signed-off-by:
-
- Sep 21, 2021
-
-
Dan Williams authored
Now that all of use sites of label data have been converted to nsl_* helpers, introduce the CXL label format. The ->cxl flag in nvdimm_drvdata indicates the label format the device expects. A follow-on patch allows a bus provider to select the label style. Note that the EFI definition of the labels represents the Linux "claim class" with a GUID. The CXL definition of the labels stores the same identifier in UUID byte order. Reviewed-by:
Jonathan Cameron <Jonathan.Cameron@huawei.com> Link: https://lore.kernel.org/r/163116432405.2460985.5547867384570123403.stgit@dwillia2-desk3.amr.corp.intel.com Signed-off-by:
-
Dan Williams authored
Add a definition of the CXL 2.0 region label format. Note this is done as a separate patch to make the next patch that adds namespace label support easier to read. Reported-by:
-
Dan Williams authored
Clean up existing kernel-doc warnings before adding new CXL label data structures. drivers/nvdimm/label.h:66: warning: Function parameter or member 'labelsize' not described in 'nd_namespace_index' drivers/nvdimm/label.h:66: warning: Function parameter or member 'free' not described in 'nd_namespace_index' drivers/nvdimm/label.h:103: warning: Function parameter or member 'align' not described in 'nd_namespace_label' drivers/nvdimm/label.h:103: warning: Function parameter or member 'reserved' not described in 'nd_namespace_label' drivers/nvdimm/label.h:103: warning: Function parameter or member 'type_guid' not described in 'nd_namespace_label' drivers/nvdimm/label.h:103: warning: Function parameter or member 'abstraction_guid' not described in 'nd_namespace_label' drivers/nvdimm/label.h:103: warning: Function parameter or member 'reserved2' not described in 'nd_namespace_label' drivers/nvdimm/label.h:103: warning: Function parameter or member 'checksum' not described in 'nd_namespace_label' Reported-by:
-
Dan Williams authored
The CXL specification defines a mechanism for namespaces to be comprised of multiple dis-contiguous ranges. Introduce that concept to the legacy NVDIMM namespace implementation with a new nsl_set_nrange() helper, that sets the number of ranges to 1. Once the NVDIMM subsystem supports CXL labels and updates its namespace capacity provisioning for dis-contiguous support nsl_set_nrange() can be updated, but in the meantime CXL label validation requires nrange be non-zero. Reported-by:
-
Dan Williams authored
In the CXL namespace label there is no need for nlabel since that is inferred from the region. Add a helper that moves nsl_get_label() behind a helper that validates the number of labels relative to the region. Reviewed-by:
-
Dan Williams authored
In preparation for CXL labels that move the uuid to a different offset in the label, add nsl_{ref,get,validate}_uuid(). These helpers use the proper uuid_t type. That type definition predated the libnvdimm subsystem, so now is as a good a time as any to convert all the uuid handling in the subsystem to uuid_t to match the helpers. Note that the uuid fields in the label data and superblocks is not replaced per Andy's expectation that uuid_t is a kernel internal type not to appear in external ABI interfaces. So, in those case {import,export}_uuid() is used to go between the 2 types. Also note that this rework uncovered some unnecessary copies for label comparisons, those are cleaned up with nsl_uuid_equal(). As for the whitespace changes, all new code is clang-format compliant. Reported-by:Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by:
-
- Aug 24, 2021
-
-
sumiyawang authored
There is a use after free crash when the pmem driver tears down its mapping while I/O is still inbound. This is triggered by driver unbind, "ndctl destroy-namespace", while I/O is in flight. Fix the sequence of blk_cleanup_queue() vs memunmap(). The crash signature is of the form: BUG: unable to handle page fault for address: ffffc90080200000 CPU: 36 PID: 9606 Comm: systemd-udevd Call Trace: ? pmem_do_bvec+0xf9/0x3a0 ? xas_alloc+0x55/0xd0 pmem_rw_page+0x4b/0x80 bdev_read_page+0x86/0xb0 do_mpage_readpage+0x5d4/0x7a0 ? lru_cache_add+0xe/0x10 mpage_readpages+0xf9/0x1c0 ? bd_link_disk_holder+0x1a0/0x1a0 blkdev_readpages+0x1d/0x20 read_pages+0x67/0x1a0 ndctl Call Trace in vmcore: PID: 23473 TASK: ffff88c4fbbe8000 CPU: 1 COMMAND: "ndctl" __schedule schedule blk_mq_freeze_queue_wait blk_freeze_queue blk_cleanup_queue pmem_release_queue devm_action_release release_nodes devres_release_all device_release_driver_internal device_driver_detach unbind_store Cc: <stable@vger.kernel.org> Signed-off-by:
sumiyawang <sumiyawang@tencent.com> Reviewed-by:
yongduan <yongduan@tencent.com> Link: https://lore.kernel.org/r/1629632949-14749-1-git-send-email-sumiyawang@tencent.com Fixes: 50f44ee7 ("mm/devm_memremap_pages: fix final page put race") Signed-off-by:
-
Dan Williams authored
In preparation for LIBNVDIMM to manage labels on CXL devices deploy helpers that abstract the label type from the implementation. The CXL label format is mostly similar to the EFI label format with concepts / fields added, like dynamic region creation and label type guids, and other concepts removed like BLK-mode and interleave-set-cookie ids. CXL labels do have the concept of a claim class represented by an "abstraction" identifier. It turns out both label implementations use the same ids, but EFI encodes them as GUIDs and CXL labels encode them as UUIDs. For now abstract out the claim class such that the UUID vs GUID distinction can later be hidden in the helper. Reviewed-by:
-
Dan Williams authored
In preparation for CXL label support, which does not have the type-guid concept, wrap the existing users with nsl_set_type_guid, and nsl_validate_type_guid. Recall that the type-guid is a value in the ACPI NFIT table to indicate how the memory range is used / should be presented to upper layers. Reviewed-by:
-
Dan Williams authored
In preparation for LIBNVDIMM to manage labels on CXL devices deploy helpers that abstract the label type from the implementation. The CXL label format is mostly similar to the EFI label format with concepts / fields added, like dynamic region creation and label type guids, and other concepts removed like BLK-mode and interleave-set-cookie ids. Finish off the BLK-mode specific helper conversion with the nlabel and position behaviour that is specific to EFI v1.2 labels and not the original v1.1 definition. Reviewed-by:
-
Dan Williams authored
In preparation for LIBNVDIMM to manage labels on CXL devices deploy helpers that abstract the label type from the implementation. The CXL label format is mostly similar to the EFI label format with concepts / fields added, like dynamic region creation and label type guids, and other concepts removed like BLK-mode and interleave-set-cookie ids. Given BLK-mode is not even supported on CXL push hide the BLK-mode specific details inside the helpers. Reviewed-by:
-
Dan Williams authored
In preparation for LIBNVDIMM to manage labels on CXL devices deploy helpers that abstract the label type from the implementation. The CXL label format is mostly similar to the EFI label format with concepts / fields added, like dynamic region creation and label type guids, and other concepts removed like BLK-mode and interleave-set-cookie ids. CXL labels support checksums by default, but early versions of the EFI labels did not. Add a validate function that can return true in the case the label format does not implement a checksum. Reviewed-by:
-
Dan Williams authored
In preparation for LIBNVDIMM to manage labels on CXL devices deploy helpers that abstract the label type from the implementation. The CXL label format is mostly similar to the EFI label format with concepts / fields added, like dynamic region creation and label type guids, and other concepts removed like BLK-mode and interleave-set-cookie ids. Reported-by:
kernel test robot <lkp@intel.com> Reviewed-by:
-
Dan Williams authored
In preparation to handle CXL labels with the same code that handles EFI labels, add a specific interleave-set-cookie validation helper rather than a getter since the CXL label type does not support this concept. The answer for CXL labels will always be true. Reviewed-by:
-
Dan Williams authored
In preparation for LIBNVDIMM to manage labels on CXL devices deploy helpers that abstract the label type from the implementation. The CXL label format is mostly similar to the EFI label format with concepts / fields added, like dynamic region creation and label type guids, and other concepts removed like BLK-mode and interleave-set-cookie ids. In addition to nsl_get_* helpers there is the nsl_ref_name() helper that returns a pointer to a label field rather than copying the data. Where changes touch the old whitespace style, update to clang-format expectations. Reviewed-by:
-
- Aug 11, 2021
-
-
Dan Williams authored
There are a few scenarios where init_active_labels() can return without registering deactivate_labels() to run when the region is disabled. In particular label error injection creates scenarios where a DIMM is disabled, but labels on other DIMMs in the region become activated. Arrange for init_active_labels() to always register deactivate_labels(). Reported-by:
Krzysztof Kensicki <krzysztof.kensicki@intel.com> Cc: <stable@vger.kernel.org> Fixes: bf9bccc1 ("libnvdimm: pmem label sets and namespace instantiation.") Reviewed-by:
Jeff Moyer <jmoyer@redhat.com> Link: https://lore.kernel.org/r/162766356450.3223041.1183118139023841447.stgit@dwillia2-desk3.amr.corp.intel.com Signed-off-by:
-
- Jul 21, 2021
-
-
Uwe Kleine-König authored
The driver core ignores the return value of this callback because there is only little it can do when a device disappears. This is the final bit of a long lasting cleanup quest where several buses were converted to also return void from their remove callback. Additionally some resource leaks were fixed that were caused by drivers returning an error code in the expectation that the driver won't go away. With struct bus_type::remove returning void it's prevented that newly implemented buses return an ignored error code and so don't anticipate wrong expectations for driver authors. Reviewed-by: Tom Rix <trix@redhat.com> (For fpga) Reviewed-by:
Mathieu Poirier <mathieu.poirier@linaro.org> Reviewed-by: Cornelia Huck <cohuck@redhat.com> (For drivers/s390 and drivers/vfio) Acked-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> (For ARM, Amba and related parts) Acked-by:
Mark Brown <broonie@kernel.org> Acked-by: Chen-Yu Tsai <wens@csie.org> (for sunxi-rsb) Acked-by:
Pali Rohár <pali@kernel.org> Acked-by: Mauro Carvalho Chehab <mchehab@kernel.org> (for media) Acked-by: Hans de Goede <hdegoede@redhat.com> (For drivers/platform) Acked-by:
Alexandre Belloni <alexandre.belloni@bootlin.com> Acked-By:
Vinod Koul <vkoul@kernel.org> Acked-by: Juergen Gross <jgross@suse.com> (For xen) Acked-by: Lee Jones <lee.jones@linaro.org> (For mfd) Acked-by: Johannes Thumshirn <jth@kernel.org> (For mcb) Acked-by:
Johan Hovold <johan@kernel.org> Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> (For slimbus) Acked-by: Kirti Wankhede <kwankhede@nvidia.com> (For vfio) Acked-by:
Maximilian Luz <luzmaximilian@gmail.com> Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com> (For ulpi and typec) Acked-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com> (For ipack) Acked-by: Geoff Levand <geoff@infradead.org> (For ps3) Acked-by: Yehezkel Bernat <YehezkelShB@gmail.com> (For thunderbolt) Acked-by: Alexander Shishkin <alexander.shishkin@linux.intel.com> (For intel_th) Acked-by: Dominik Brodowski <linux@dominikbrodowski.net> (For pcmcia) Acked-by: Rafael J. Wysocki <rafael@kernel.org> (For ACPI) Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org> (rpmsg and apr) Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com> (For intel-ish-hid) Acked-by: Dan Williams <dan.j.williams@intel.com> (For CXL, DAX, and NVDIMM) Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com> (For isa) Acked-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (For firewire) Acked-by: Benjamin Tissoires <benjamin.tissoires@redhat.com> (For hid) Acked-by: Thorsten Scherer <t.scherer@eckelmann.de> (For siox) Acked-by: Sven Van Asbroeck <TheSven73@gmail.com> (For anybuss) Acked-by: Ulf Hansson <ulf.hansson@linaro.org> (For MMC) Acked-by: Wolfram Sang <wsa@kernel.org> # for I2C Acked-by:
Sudeep Holla <sudeep.holla@arm.com> Acked-by:
Geert Uytterhoeven <geert@linux-m68k.org> Acked-by:
Dmitry Torokhov <dmitry.torokhov@gmail.com> Acked-by:
Finn Thain <fthain@linux-m68k.org> Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Link: https://lore.kernel.org/r/20210713193522.1770306-6-u.kleine-koenig@pengutronix.de Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
- Jun 15, 2021
-
-
Dan Williams authored
LIBNVDIMM device objects register sysfs power attributes despite nothing requiring that support. Clean up sysfs remove the power/ attribute group. This requires a device_create() and a device_register() usage to be converted to the device_initialize() + device_add() pattern. Reviewed-by:
-
Dan Williams authored
CXL is a hotplug bus and arranges for nvdimm devices to be dynamically discovered and removed. The libnvdimm core manages shutdown of nvdimm security operations when the device is unregistered. That functionality is moved to nvdimm_delete() and invoked by the CXL-to-nvdimm glue code. Reviewed-by:
-
- Jun 09, 2021
-
-
Dan Williams authored
The queue_to_disk() helper can not be used after del_gendisk() communicate @disk via the pgmap->owner. Otherwise, queue_to_disk() returns NULL resulting in the splat below. Kernel attempted to read user page (330) - exploit attempt? (uid: 0) BUG: Kernel NULL pointer dereference on read at 0x00000330 Faulting instruction address: 0xc000000000906344 Oops: Kernel access of bad area, sig: 11 [#1] [..] NIP [c000000000906344] pmem_pagemap_cleanup+0x24/0x40 LR [c0000000004701d4] memunmap_pages+0x1b4/0x4b0 Call Trace: [c000000022cbb9c0] [c0000000009063c8] pmem_pagemap_kill+0x28/0x40 (unreliable) [c000000022cbb9e0] [c0000000004701d4] memunmap_pages+0x1b4/0x4b0 [c000000022cbba90] [c0000000008b28a0] devm_action_release+0x30/0x50 [c000000022cbbab0] [c0000000008b39c8] release_nodes+0x2f8/0x3e0 [c000000022cbbb60] [c0000000008ac440] device_release_driver_internal+0x190/0x2b0 [c000000022cbbba0] [c0000000008a8450] unbind_store+0x130/0x170 Reported-by:
Sachin Sant <sachinp@linux.vnet.ibm.com> Fixes: 87eb73b2 ("nvdimm-pmem: convert to blk_alloc_disk/blk_cleanup_disk") Link: http://lore.kernel.org/r/DFB75BA8-603F-4A35-880B-C5B23EF8FA7D@linux.vnet.ibm.com Cc: Christoph Hellwig <hch@lst.de> Cc: Ulf Hansson <ulf.hansson@linaro.org> Cc: Jens Axboe <axboe@kernel.dk> Signed-off-by:
-
- Jun 01, 2021
-
-
Christoph Hellwig authored
Convert the nvme-multipath driver to use the blk_alloc_disk and blk_cleanup_disk helpers to simplify gendisk and request_queue allocation. Signed-off-by:
Ulf Hansson <ulf.hansson@linaro.org> Link: https://lore.kernel.org/r/20210521055116.1053587-19-hch@lst.de Signed-off-by:
-
Christoph Hellwig authored
Convert the nvdimm-pmem driver to use the blk_alloc_disk and blk_cleanup_disk helpers to simplify gendisk and request_queue allocation. Signed-off-by:
-
Christoph Hellwig authored
Convert the nvdimm-btt driver to use the blk_alloc_disk and blk_cleanup_disk helpers to simplify gendisk and request_queue allocation. Signed-off-by:
Hannes Reinecke <hare@suse.de> Reviewed-by:
-
Christoph Hellwig authored
Convert the nvdimm-blk driver to use the blk_alloc_disk and blk_cleanup_disk helpers to simplify gendisk and request_queue allocation. Signed-off-by:
-
Christoph Hellwig authored
Automatically set the GENHD_FL_EXT_DEVT flag for all disks allocated without an explicit number of minors. This is what all new block drivers should do, so make sure it is the default without boilerplate code. Signed-off-by:
-
- May 07, 2021
-
-
Matthew Wilcox (Oracle) authored
My UEK-derived config has 1030 files depending on pagemap.h before this change. Afterwards, just 326 files need to be rebuilt when I touch pagemap.h. I think blkdev.h is probably included too widely, but untangling that dependency is harder and this solves my problem. x86 allmodconfig builds, but there may be implicit include problems on other architectures. Link: https://lkml.kernel.org/r/20210309195747.283796-1-willy@infradead.org Signed-off-by:
Matthew Wilcox (Oracle) <willy@infradead.org> Acked-by: Dan Williams <dan.j.williams@intel.com> [nvdimm] Acked-by: Jens Axboe <axboe@kernel.dk> [block] Reviewed-by:
William Kucharski <william.kucharski@oracle.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
- Apr 10, 2021
-
-
Vaibhav Jain authored
In case a platform doesn't provide explicit flush-hints but provides an explicit flush callback via ND_REGION_ASYNC region flag, then nvdimm_has_flush() still returns '0' indicating that writes do not require flushing. This happens on PPC64 with patch at [1] applied, where 'deep_flush' of a region was denied even though an explicit flush function was provided. Fix this by adding a condition to nvdimm_has_flush() to test for the ND_REGION_ASYNC flag on the region and see if a 'region->flush' callback is assigned. Link: http://lore.kernel.org/r/161703936121.36.7260632399582101498.stgit@e1fbed493c87 [1] Fixes: c5d4355d ("libnvdimm: nd_region flush callback support") Reported-by:
Shivaprasad G Bhat <sbhat@linux.ibm.com> Signed-off-by:
Vaibhav Jain <vaibhav@linux.ibm.com> Link: https://lore.kernel.org/r/20210402092555.208590-1-vaibhav@linux.ibm.com Signed-off-by:
-
- Apr 09, 2021
-
-
Dan Williams authored
Previous kernels allowed the BLKROSET to override the disk's read-only status. With that situation fixed the pmem driver needs to rely on notification events to reevaluate the disk read-only status after the host region has been marked read-write. Recall that when libnvdimm determines that the persistent memory has lost persistence (for example lack of energy to flush from DRAM to FLASH on an NVDIMM-N device) it marks the region read-only, but that state can be overridden by the user via: echo 0 > /sys/bus/nd/devices/regionX/read_only ...to date there is no notification that the region has restored persistence, so the user override is the only recovery. Fixes: 52f019d4 ("block: add a hard-readonly flag to struct gendisk") Reported-by:
Vishal Verma <vishal.l.verma@intel.com> Tested-by:
-
