diff options
author | Christian Borntraeger <borntraeger@de.ibm.com> | 2008-01-26 16:11:16 +0300 |
---|---|---|
committer | Martin Schwidefsky <schwidefsky@de.ibm.com> | 2008-01-26 16:11:23 +0300 |
commit | a2fd64d6aaf498756f700eb1d07818efee046733 (patch) | |
tree | f89c7119bcb00826a7ed99b14d2bbe7d2edcc043 /arch/s390/mm | |
parent | d09293aee9fd0f9f291fa3a0e4ca5012f514d083 (diff) | |
download | linux-a2fd64d6aaf498756f700eb1d07818efee046733.tar.xz |
[S390] vmemmap: allocate struct pages before 1:1 mapping
We have seen an oops in an OOM situation, where show_mem tried to
access the struct page of a dcss segment. The vmemmap code has
already created the 1:1 mapping but failed allocating the struct
pages. In the OOM case, show_mem now walks the memory. It uses
pfn_valid to detect if it may access the struct page. In the case
described above, the mapping was established and pfn_valid returned
true. As the struct pages were not allocated, the kernel oopsed.
We have to ensure that we have created the struct pages, before we
add a mapping pointing to the pages.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Diffstat (limited to 'arch/s390/mm')
-rw-r--r-- | arch/s390/mm/vmem.c | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c index 1fc99891c4eb..79d13a166a3d 100644 --- a/arch/s390/mm/vmem.c +++ b/arch/s390/mm/vmem.c @@ -236,10 +236,10 @@ static int vmem_add_mem(unsigned long start, unsigned long size) { int ret; - ret = vmem_add_range(start, size); + ret = vmem_add_mem_map(start, size); if (ret) return ret; - return vmem_add_mem_map(start, size); + return vmem_add_range(start, size); } /* |