[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <152532351517.17218.3583455156840230837.stgit@dwillia2-desk3.amr.corp.intel.com>
Date: Wed, 02 May 2018 21:58:35 -0700
From: Dan Williams <dan.j.williams@...el.com>
To: linux-nvdimm@...ts.01.org
Cc: Tony Luck <tony.luck@...el.com>,
Peter Zijlstra <peterz@...radead.org>,
Linus Torvalds <torvalds@...ux-foundation.org>,
Borislav Petkov <bp@...en8.de>,
Mike Snitzer <snitzer@...hat.com>, x86@...nel.org,
Thomas Gleixner <tglx@...utronix.de>,
Christoph Hellwig <hch@....de>,
Ross Zwisler <ross.zwisler@...ux.intel.com>,
Andy Lutomirski <luto@...capital.net>,
Ingo Molnar <mingo@...hat.com>,
Al Viro <viro@...iv.linux.org.uk>,
Andrew Morton <akpm@...ux-foundation.org>,
linux-kernel@...r.kernel.org, tony.luck@...el.com,
linux-fsdevel@...r.kernel.org
Subject: [PATCH v2 0/9] use memcpy_mcsafe() for copy_to_iter()
Changes since v1 [1]:
* Remove the loop unrolling in the assembly implementation since it
significantly complicates the exception handling (Linus)
* Introduce a ->copy_to_iter() dax operation for symmetry with the
existing ->copy_from_iter() operation to allow platform /
device-specific implementations.
[1]: https://lists.01.org/pipermail/linux-nvdimm/2018-May/015548.html
---
Currently memcpy_mcsafe() is only deployed in the pmem driver when
reading through a /dev/pmemX block device. However, a filesystem in dax
mode mounted on a /dev/pmemX block device will bypass the block layer
and the driver for reads. The filesystem-dax (fsdax) read case uses
dax_direct_access() and copy_to_iter() to bypass the block layer.
The result of the bypass is that the kernel treats machine checks during
read as system fatal (reboot) when they could simply be flagged as an
I/O error, similar to performing reads through the pmem driver. Prevent
this fatal condition by deploying memcpy_mcsafe() in the fsdax read
path.
The main differences between this copy_to_user_mcsafe() and
copy_user_generic_unrolled() are:
* Typical tail/residue handling after a fault retries the copy
byte-by-byte until the fault happens again. Re-triggering machine
checks is potentially fatal so the implementation uses source alignment
and poison alignment assumptions to avoid re-triggering machine
checks.
* SMAP coordination is handled external to the assembly with
__uaccess_begin() and __uaccess_end().
* ITER_KVEC and ITER_BVEC can now end prematurely with an error.
The new MCSAFE_DEBUG facility is proposed as a way to unit test the
exception handling without requiring an ACPI EINJ capable platform.
---
Dan Williams (9):
x86, memcpy_mcsafe: remove loop unrolling
x86, memcpy_mcsafe: add labels for write fault handling
x86, memcpy_mcsafe: return bytes remaining
x86, memcpy_mcsafe: add write-protection-fault handling
x86, memcpy_mcsafe: define copy_to_iter_mcsafe()
dax: introduce a ->copy_to_iter dax operation
dax: report bytes remaining in dax_iomap_actor()
pmem: switch to copy_to_iter_mcsafe()
x86, nfit_test: unit test for memcpy_mcsafe()
arch/x86/Kconfig | 1
arch/x86/Kconfig.debug | 3 +
arch/x86/include/asm/mcsafe_debug.h | 50 ++++++++++++++++
arch/x86/include/asm/string_64.h | 10 ++-
arch/x86/include/asm/uaccess_64.h | 14 ++++
arch/x86/lib/memcpy_64.S | 109 ++++++++++++++++-------------------
arch/x86/lib/usercopy_64.c | 17 +++++
drivers/dax/super.c | 10 +++
drivers/md/dm-linear.c | 16 +++++
drivers/md/dm-log-writes.c | 15 +++++
drivers/md/dm-stripe.c | 21 +++++++
drivers/md/dm.c | 25 ++++++++
drivers/nvdimm/claim.c | 3 +
drivers/nvdimm/pmem.c | 13 +++-
drivers/s390/block/dcssblk.c | 7 ++
fs/dax.c | 21 ++++---
include/linux/dax.h | 5 ++
include/linux/device-mapper.h | 5 +-
include/linux/string.h | 4 +
include/linux/uio.h | 15 +++++
lib/iov_iter.c | 61 ++++++++++++++++++++
tools/testing/nvdimm/test/nfit.c | 48 +++++++++++++++
22 files changed, 394 insertions(+), 79 deletions(-)
create mode 100644 arch/x86/include/asm/mcsafe_debug.h
Powered by blists - more mailing lists