hdf5/Report-error-if-dimensions-of-chunked-storage-in-data-layout-2.patch
Atri Bhattacharya a0635c9ad7 Accepting request 1030627 from home:eeich:branches:science
- Fix CVEs:
  * CVE-2021-46244 (bsc#1195215)
    Compound-datatypes-may-not-have-members-of-size-0.patch
  * CVE-2018-13867 (bsc#1101906)
    Validate-location-offset-of-the-accumulated-metadata-when-comparing.patch
  * CVE-2018-16438 (bsc#1107069)
    Make-sure-info-block-for-external-links-has-at-least-3-bytes.patch
  * CVE-2020-10812 (bsc#1167400)
    Hot-fix-for-CVE-2020-10812.patch
  * CVE-2021-45830 (bsc#1194375)
    H5O_fsinfo_decode-Make-more-resilient-to-out-of-bounds-read.patch
  * CVE-2019-8396 (bsc#1125882)
    H5O__pline_decode-Make-more-resilient-to-out-of-bounds-read.patch
  * CVE-2018-11205 (bsc#1093663)
    Pass-compact-chunk-size-info-to-ensure-requested-elements-are-within-bounds.patch
  * CVE-2021-46242 (bsc#1195212)
    When-evicting-driver-info-block-NULL-the-corresponding-entry.patch
  * CVE-2021-45833 (bsc#1194366)
    Report-error-if-dimensions-of-chunked-storage-in-data-layout-2.patch
  * CVE-2018-14031 (bsc#1101475)
    H5O_dtype_decode_helper-Parent-of-enum-needs-to-have-same-size-as-enum-itself.patch
  * CVE-2018-17439 (bsc#1111598)
    H5IMget_image_info-H5Sget_simple_extent_dims-does-not-exceed-array-size.patch
- Fix an error message:
    Fix-error-message-not-the-name-but-the-link-information-is-parsed.patch

OBS-URL: https://build.opensuse.org/request/show/1030627
OBS-URL: https://build.opensuse.org/package/show/science/hdf5?expand=0&rev=159
2022-10-24 06:53:22 +00:00

49 lines
2.1 KiB
Diff

From: Egbert Eich <eich@suse.com>
Date: Wed Sep 28 19:11:16 2022 +0200
Subject: Report error if dimensions of chunked storage in data layout < 2
Patch-mainline: Not yet
Git-repo: ssh://eich@192.168.122.1:/home/eich/sources/HPC/hdf5
Git-commit: 34b621424504265cff3c33cf634a70efb52db180
References:
For Data Layout Messages version 1 & 2 the specification state
that the value stored in the data field is 1 greater than the
number of dimensions in the dataspace. For version 3 this is
not explicitly stated but the implementation suggests it to be
the case.
Thus the set value needs to be at least 2. For dimensionality
< 2 an out-of-bounds access occurs as in CVE-2021-45833.
This fixes CVE-2021-45833.
Signed-off-by: Egbert Eich <eich@suse.com>
Signed-off-by: Egbert Eich <eich@suse.de>
---
src/H5Olayout.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/src/H5Olayout.c b/src/H5Olayout.c
index c939e72744..9fa9e36e8c 100644
--- a/src/H5Olayout.c
+++ b/src/H5Olayout.c
@@ -168,6 +168,10 @@ H5O__layout_decode(H5F_t *f, H5O_t H5_ATTR_UNUSED *open_oh, unsigned H5_ATTR_UNU
p += ndims * 4; /* Skip over dimension sizes (32-bit quantities) */
} /* end if */
else {
+ if (ndims < 2)
+ HGOTO_ERROR(H5E_OHDR, H5E_CANTLOAD, NULL,
+ "bad dimensions for chunked storage")
+
mesg->u.chunk.ndims = ndims;
for (u = 0; u < ndims; u++)
UINT32DECODE(p, mesg->u.chunk.dim[u]);
@@ -241,6 +245,9 @@ H5O__layout_decode(H5F_t *f, H5O_t H5_ATTR_UNUSED *open_oh, unsigned H5_ATTR_UNU
mesg->u.chunk.ndims = *p++;
if (mesg->u.chunk.ndims > H5O_LAYOUT_NDIMS)
HGOTO_ERROR(H5E_OHDR, H5E_CANTLOAD, NULL, "dimensionality is too large")
+ if (mesg->u.chunk.ndims < 2)
+ HGOTO_ERROR(H5E_OHDR, H5E_CANTLOAD, NULL,
+ "bad dimensions for chunked storage")
/* B-tree address */
H5F_addr_decode(f, &p, &(mesg->storage.u.chunk.idx_addr));