Revision e2093926a098a8ccf0f1d10f6df8dad452cb28d3 authored by Ross Zwisler on 02 June 2017, 21:46:37 UTC, committed by Linus Torvalds on 02 June 2017, 22:07:37 UTC
We currently have two related PMD vs PTE races in the DAX code. These can both be easily triggered by having two threads reading and writing simultaneously to the same private mapping, with the key being that private mapping reads can be handled with PMDs but private mapping writes are always handled with PTEs so that we can COW. Here is the first race: CPU 0 CPU 1 (private mapping write) __handle_mm_fault() create_huge_pmd() - FALLBACK handle_pte_fault() passes check for pmd_devmap() (private mapping read) __handle_mm_fault() create_huge_pmd() dax_iomap_pmd_fault() inserts PMD dax_iomap_pte_fault() does a PTE fault, but we already have a DAX PMD installed in our page tables at this spot. Here's the second race: CPU 0 CPU 1 (private mapping read) __handle_mm_fault() passes check for pmd_none() create_huge_pmd() dax_iomap_pmd_fault() inserts PMD (private mapping write) __handle_mm_fault() create_huge_pmd() - FALLBACK (private mapping read) __handle_mm_fault() passes check for pmd_none() create_huge_pmd() handle_pte_fault() dax_iomap_pte_fault() inserts PTE dax_iomap_pmd_fault() inserts PMD, but we already have a PTE at this spot. The core of the issue is that while there is isolation between faults to the same range in the DAX fault handlers via our DAX entry locking, there is no isolation between faults in the code in mm/memory.c. This means for instance that this code in __handle_mm_fault() can run: if (pmd_none(*vmf.pmd) && transparent_hugepage_enabled(vma)) { ret = create_huge_pmd(&vmf); But by the time we actually get to run the fault handler called by create_huge_pmd(), the PMD is no longer pmd_none() because a racing PTE fault has installed a normal PMD here as a parent. This is the cause of the 2nd race. The first race is similar - there is the following check in handle_pte_fault(): } else { /* See comment in pte_alloc_one_map() */ if (pmd_devmap(*vmf->pmd) || pmd_trans_unstable(vmf->pmd)) return 0; So if a pmd_devmap() PMD (a DAX PMD) has been installed at vmf->pmd, we will bail and retry the fault. This is correct, but there is nothing preventing the PMD from being installed after this check but before we actually get to the DAX PTE fault handlers. In my testing these races result in the following types of errors: BUG: Bad rss-counter state mm:ffff8800a817d280 idx:1 val:1 BUG: non-zero nr_ptes on freeing mm: 15 Fix this issue by having the DAX fault handlers verify that it is safe to continue their fault after they have taken an entry lock to block other racing faults. [ross.zwisler@linux.intel.com: improve fix for colliding PMD & PTE entries] Link: http://lkml.kernel.org/r/20170526195932.32178-1-ross.zwisler@linux.intel.com Link: http://lkml.kernel.org/r/20170522215749.23516-2-ross.zwisler@linux.intel.com Signed-off-by: Ross Zwisler <ross.zwisler@linux.intel.com> Reported-by: Pawel Lebioda <pawel.lebioda@intel.com> Reviewed-by: Jan Kara <jack@suse.cz> Cc: "Darrick J. Wong" <darrick.wong@oracle.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@lst.de> Cc: Dan Williams <dan.j.williams@intel.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Matthew Wilcox <mawilcox@microsoft.com> Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com> Cc: Pawel Lebioda <pawel.lebioda@intel.com> Cc: Dave Jiang <dave.jiang@intel.com> Cc: Xiong Zhou <xzhou@redhat.com> Cc: Eryu Guan <eguan@redhat.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent d0f0931
diffconfig
#!/usr/bin/python
#
# diffconfig - a tool to compare .config files.
#
# originally written in 2006 by Matt Mackall
# (at least, this was in his bloatwatch source code)
# last worked on 2008 by Tim Bird
#
import sys, os
def usage():
print("""Usage: diffconfig [-h] [-m] [<config1> <config2>]
Diffconfig is a simple utility for comparing two .config files.
Using standard diff to compare .config files often includes extraneous and
distracting information. This utility produces sorted output with only the
changes in configuration values between the two files.
Added and removed items are shown with a leading plus or minus, respectively.
Changed items show the old and new values on a single line.
If -m is specified, then output will be in "merge" style, which has the
changed and new values in kernel config option format.
If no config files are specified, .config and .config.old are used.
Example usage:
$ diffconfig .config config-with-some-changes
-EXT2_FS_XATTR n
CRAMFS n -> y
EXT2_FS y -> n
LOG_BUF_SHIFT 14 -> 16
PRINTK_TIME n -> y
""")
sys.exit(0)
# returns a dictionary of name/value pairs for config items in the file
def readconfig(config_file):
d = {}
for line in config_file:
line = line[:-1]
if line[:7] == "CONFIG_":
name, val = line[7:].split("=", 1)
d[name] = val
if line[-11:] == " is not set":
d[line[9:-11]] = "n"
return d
def print_config(op, config, value, new_value):
global merge_style
if merge_style:
if new_value:
if new_value=="n":
print("# CONFIG_%s is not set" % config)
else:
print("CONFIG_%s=%s" % (config, new_value))
else:
if op=="-":
print("-%s %s" % (config, value))
elif op=="+":
print("+%s %s" % (config, new_value))
else:
print(" %s %s -> %s" % (config, value, new_value))
def main():
global merge_style
# parse command line args
if ("-h" in sys.argv or "--help" in sys.argv):
usage()
merge_style = 0
if "-m" in sys.argv:
merge_style = 1
sys.argv.remove("-m")
argc = len(sys.argv)
if not (argc==1 or argc == 3):
print("Error: incorrect number of arguments or unrecognized option")
usage()
if argc == 1:
# if no filenames given, assume .config and .config.old
build_dir=""
if "KBUILD_OUTPUT" in os.environ:
build_dir = os.environ["KBUILD_OUTPUT"]+"/"
configa_filename = build_dir + ".config.old"
configb_filename = build_dir + ".config"
else:
configa_filename = sys.argv[1]
configb_filename = sys.argv[2]
try:
a = readconfig(open(configa_filename))
b = readconfig(open(configb_filename))
except (IOError):
e = sys.exc_info()[1]
print("I/O error[%s]: %s\n" % (e.args[0],e.args[1]))
usage()
# print items in a but not b (accumulate, sort and print)
old = []
for config in a:
if config not in b:
old.append(config)
old.sort()
for config in old:
print_config("-", config, a[config], None)
del a[config]
# print items that changed (accumulate, sort, and print)
changed = []
for config in a:
if a[config] != b[config]:
changed.append(config)
else:
del b[config]
changed.sort()
for config in changed:
print_config("->", config, a[config], b[config])
del b[config]
# now print items in b but not in a
# (items from b that were in a were removed above)
new = sorted(b.keys())
for config in new:
print_config("+", config, None, b[config])
main()
![swh spinner](/static/img/swh-spinner.gif)
Computing file changes ...