smb: client: fix "df: Resource temporarily unavailable" on 5.10 stable kernel
cover smb: client: fix "df: Resource temporarily unavailable" on 5.10 stable kernel
After mounting a remote cifs resource, it becomes unavailable:
df: /mnt/sambashare: Resource temporarily unavailable
It was tested on the following Linux kernels:
Linux altlinux 5.10.208-std-def-alt1
Linux fedora 5.10.208-200.el8.x86_64
The error appeared starting from kernel 5.10.206 after adding the commit [1] "smb: client: fix OOB in SMB2_query_info_init()", in which the buffer length increases by 1 as a result of changes:
...
- iov[0].iov_len = total_len - 1 + input_len;
+ iov[0].iov_len = len;
...
[1] https://patchwork.kernel.org/project/cifs-client/patch/[email protected]/
Error fixed by backported commits in next two patches adapted for the 5.10 kernel:
[PATCH 1/2] stddef: Introduce DECLARE_FLEX_ARRAY() helper [PATCH 2/2] smb3: Replace smb2pdu 1-element arrays with flex-arrays
PATCH 1/2 stddef: Introduce DECLARE_FLEX_ARRAY() helper
From: Kees Cook <[email protected]>
commit 3080ea5553cc909b000d1f1d964a9041962f2c5b upstream.
There are many places where kernel code wants to have several different typed trailing flexible arrays. This would normally be done with multiple flexible arrays in a union, but since GCC and Clang don't (on the surface) allow this, there have been many open-coded workarounds, usually involving neighboring 0-element arrays at the end of a structure. For example, instead of something like this:
struct thing {
...
union {
struct type1 foo[];
struct type2 bar[];
};
};code works around the compiler with:
struct thing {
...
struct type1 foo[0];
struct type2 bar[];
};Another case is when a flexible array is wanted as the single member within a struct (which itself is usually in a union). For example, this would be worked around as:
union many {
...
struct {
struct type3 baz[0];
};
};These kinds of work-arounds cause problems with size checks against such zero-element arrays (for example when building with -Warray-bounds and -Wzero-length-bounds, and with the coming FORTIFY_SOURCE improvements), so they must all be converted to "real" flexible arrays, avoiding warnings like this:
fs/hpfs/anode.c: In function 'hpfs_add_sector_to_btree':fs/hpfs/anode.c:209:27: warning: array subscript 0 is outside the bounds of an interior zero-length array 'struct bplus_internal_node[0]' [-Wzero-length-bounds]
209 | anode->btree.u.internal[0].down = cpu_to_le32(a);
| ~~~~~~~~~~~~~~~~~~~~~~~^~~
In file included from fs/hpfs/hpfs_fn.h:26,
from fs/hpfs/anode.c:10:fs/hpfs/hpfs.h:412:32: note: while referencing 'internal'
412 | struct bplus_internal_node internal[0]; /* (internal) 2-word entries giving
| ^~~~~~~~
drivers/net/can/usb/etas_es58x/es58x_fd.c: In function 'es58x_fd_tx_can_msg':drivers/net/can/usb/etas_es58x/es58x_fd.c:360:35: warning: array subscript 65535 is outside the bounds of an interior zero-length array 'u8[0]' {aka 'unsigned char[]'} [-Wzero-length-bounds]
360 | tx_can_msg = (typeof(tx_can_msg))&es58x_fd_urb_cmd->raw_msg[msg_len];
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from drivers/net/can/usb/etas_es58x/es58x_core.h:22,
from drivers/net/can/usb/etas_es58x/es58x_fd.c:17:
drivers/net/can/usb/etas_es58x/es58x_fd.h:231:6: note: while referencing 'raw_msg'
231 | u8 raw_msg[0];
| ^~~~~~~However, it _is_ entirely possible to have one or more flexible arrays in a struct or union: it just has to be in another struct. And since it cannot be alone in a struct, such a struct must have at least 1 other named member -- but that member can be zero sized. Wrap all this nonsense into the new DECLARE_FLEX_ARRAY() in support of having flexible arrays in unions (or alone in a struct).
As with struct_group(), since this is needed in UAPI headers as well, implement the core there, with a non-UAPI wrapper.
Additionally update kernel-doc to understand its existence.
https://github.com/KSPP/linux/issues/137
--- include/linux/stddef.h | 13 +++++++++++++ include/uapi/linux/stddef.h | 16 ++++++++++++++++ scripts/kernel-doc | 3 ++- 3 files changed, 31 insertions(+), 1 deletion(-)
include/linux/stddef.h +13 -0
@@ -84,4 +84,17 @@ enum { #define struct_group_tagged(TAG, NAME, MEMBERS...) \ __struct_group(TAG, NAME, /* no attrs */, MEMBERS) +/** +* DECLARE_FLEX_ARRAY() - Declare a flexible array usable in a union +* +* @TYPE: The type of each flexible array element +* @NAME: The name of the flexible array member +* +* In order to have a flexible array member in a union or alone in a +* struct, it needs to be wrapped in an anonymous struct with at least 1 +* named member, but that member can be empty. +*/ +#define DECLARE_FLEX_ARRAY (TYPE ,NAME ) \ + __DECLARE_FLEX_ARRAY(TYPE, NAME) +#endif
include/uapi/linux/stddef.h +16 -0
@@ -28,4 +28,20 @@ struct { MEMBERS } ATTRS; \ struct TAG { MEMBERS } ATTRS NAME; \ } + +/** + * __DECLARE_FLEX_ARRAY() - Declare a flexible array usable in a union + * + * @TYPE: The type of each flexible array element + * @NAME: The name of the flexible array member + * + * In order to have a flexible array member in a union or alone in a + * struct, it needs to be wrapped in an anonymous struct with at least 1 + * named member, but that member can be empty. + */ +#define __DECLARE_FLEX_ARRAY(TYPE, NAME) \ + struct { \ + struct { } __empty_ ## NAME; \ + TYPE NAME[]; \ + } #endif
scripts/kernel-doc +2 -1
@@ -1232,7 +1232,8 @@ sub dump_struct($$) { $members =~ s/DECLARE_KFIFO\s*\(([^,)]+),\s*([^,)]+),\s*([^,)]+)\)/$2 \*$1/gos; # replace DECLARE_KFIFO_PTR $members =~ s/DECLARE_KFIFO_PTR\s*\(([^,)]+),\s*([^,)]+)\)/$2 \*$1/gos; - + # replace DECLARE_FLEX_ARRAY + $members =~ s/(?:__)?DECLARE_FLEX_ARRAY\s*\($args,\s*$args\)/$1 $2\[\]/gos; my $declaration = $members; # Split nested struct/union elements as newer ones
Signature · 1 line
2.33.8
PATCH 2/2 smb3: Replace smb2pdu 1-element arrays with flex-arrays
In reply to [email protected]From: Kees Cook <[email protected]>
commit eb3e28c1e89b4984308777231887e41aa8a0151f upstream.
The kernel is globally removing the ambiguous 0-length and 1-element arrays in favor of flexible arrays, so that we can gain both compile-time and run-time array bounds checking[1].
Replace the trailing 1-element array with a flexible array in the following structures:
struct smb2_err_rsp
struct smb2_tree_connect_req
struct smb2_negotiate_rsp
struct smb2_sess_setup_req
struct smb2_sess_setup_rsp
struct smb2_read_req
struct smb2_read_rsp
struct smb2_write_req
struct smb2_write_rsp
struct smb2_query_directory_req
struct smb2_query_directory_rsp
struct smb2_set_info_req
struct smb2_change_notify_rsp
struct smb2_create_rsp
struct smb2_query_info_req
struct smb2_query_info_rsp
Replace the trailing 1-element array with a flexible array, but leave the existing structure padding:
struct smb2_file_all_info
struct smb2_lock_req
Adjust all related size calculations to match the changes to sizeof().
No machine code output or .data section differences are produced after these changes.
[1] For lots of details, see both:
https://docs.kernel.org/process/deprecated.html#zero-length-and-one-element-arrays
https://people.kernel.org/kees/bounded-flexible-arrays-in-c
--- fs/cifs/smb2misc.c | 2 +- fs/cifs/smb2ops.c | 14 +++++++------- fs/cifs/smb2pdu.c | 13 ++++++------- fs/cifs/smb2pdu.h | 42 ++++++++++++++++++++++++------------------ 4 files changed, 38 insertions(+), 33 deletions(-)
fs/cifs/smb2misc.c +1 -1
@@ -117,7 +117,7 @@ static __u32 get_neg_ctxt_len(struct smb2_sync_hdr *hdr, __u32 len, }else if (nc_offset +1 == non_ctxlen) {cifs_dbg (FYI ,"no SPNEGO security blob in negprot rsp\n" ); size_of_pad_before_neg_ctxts =0 ; - }else if (non_ctxlen ==SMB311_NEGPROT_BASE_SIZE ) + }else if (non_ctxlen ==SMB311_NEGPROT_BASE_SIZE +1 )/* has padding, but no SPNEGO blob */ size_of_pad_before_neg_ctxts = nc_offset - non_ctxlen +1 ;else
fs/cifs/smb2ops.c +7 -7
@@ -5561,7 +5561,7 @@ struct smb_version_values smb20_values = { .header_size = sizeof(struct smb2_sync_hdr), .header_preamble_size = 0, .max_header_size = MAX_SMB2_HDR_SIZE, - .read_rsp_size = sizeof(struct smb2_read_rsp) - 1, + .read_rsp_size = sizeof(struct smb2_read_rsp), .lock_cmd = SMB2_LOCK, .cap_unix = 0, .cap_nt_find = SMB2_NT_FIND,
@@ -5583,7 +5583,7 @@ struct smb_version_values smb21_values = { .header_size = sizeof(struct smb2_sync_hdr), .header_preamble_size = 0, .max_header_size = MAX_SMB2_HDR_SIZE, - .read_rsp_size = sizeof(struct smb2_read_rsp) - 1, + .read_rsp_size = sizeof(struct smb2_read_rsp), .lock_cmd = SMB2_LOCK, .cap_unix = 0, .cap_nt_find = SMB2_NT_FIND,
@@ -5604,7 +5604,7 @@ struct smb_version_values smb3any_values = { .header_size = sizeof(struct smb2_sync_hdr), .header_preamble_size = 0, .max_header_size = MAX_SMB2_HDR_SIZE, - .read_rsp_size = sizeof(struct smb2_read_rsp) - 1, + .read_rsp_size = sizeof(struct smb2_read_rsp), .lock_cmd = SMB2_LOCK, .cap_unix = 0, .cap_nt_find = SMB2_NT_FIND,
@@ -5625,7 +5625,7 @@ struct smb_version_values smbdefault_values = { .header_size = sizeof(struct smb2_sync_hdr), .header_preamble_size = 0, .max_header_size = MAX_SMB2_HDR_SIZE, - .read_rsp_size = sizeof(struct smb2_read_rsp) - 1, + .read_rsp_size = sizeof(struct smb2_read_rsp), .lock_cmd = SMB2_LOCK, .cap_unix = 0, .cap_nt_find = SMB2_NT_FIND,
@@ -5646,7 +5646,7 @@ struct smb_version_values smb30_values = { .header_size = sizeof(struct smb2_sync_hdr), .header_preamble_size = 0, .max_header_size = MAX_SMB2_HDR_SIZE, - .read_rsp_size = sizeof(struct smb2_read_rsp) - 1, + .read_rsp_size = sizeof(struct smb2_read_rsp), .lock_cmd = SMB2_LOCK, .cap_unix = 0, .cap_nt_find = SMB2_NT_FIND,
@@ -5667,7 +5667,7 @@ struct smb_version_values smb302_values = { .header_size = sizeof(struct smb2_sync_hdr), .header_preamble_size = 0, .max_header_size = MAX_SMB2_HDR_SIZE, - .read_rsp_size = sizeof(struct smb2_read_rsp) - 1, + .read_rsp_size = sizeof(struct smb2_read_rsp), .lock_cmd = SMB2_LOCK, .cap_unix = 0, .cap_nt_find = SMB2_NT_FIND,
@@ -5688,7 +5688,7 @@ struct smb_version_values smb311_values = { .header_size = sizeof(struct smb2_sync_hdr), .header_preamble_size = 0, .max_header_size = MAX_SMB2_HDR_SIZE, - .read_rsp_size = sizeof(struct smb2_read_rsp) - 1, + .read_rsp_size = sizeof(struct smb2_read_rsp), .lock_cmd = SMB2_LOCK, .cap_unix = 0, .cap_nt_find = SMB2_NT_FIND,
fs/cifs/smb2pdu.c +6 -7
@@ -1261,7 +1261,7 @@ SMB2_sess_sendreceive(struct SMB2_sess_data *sess_data)/* Testing shows that buffer offset must be at location of Buffer[0] */ req->SecurityBufferOffset = -cpu_to_le16 (sizeof (struct smb2_sess_setup_req ) -1 /* pad */ ); +cpu_to_le16 (sizeof (struct smb2_sess_setup_req )); req->SecurityBufferLength =cpu_to_le16 (sess_data->iov[1 ].iov_len);memset (&rqst,0 ,sizeof (struct smb_rqst ));
@@ -1760,8 +1760,7 @@ SMB2_tcon(const unsigned int xid, struct cifs_ses *ses, const char *tree, iov[0 ].iov_len = total_len -1 ;/* Testing shows that buffer offset must be at location of Buffer[0] */ - req->PathOffset =cpu_to_le16 (sizeof (struct smb2_tree_connect_req ) - -1 /* pad */ ); + req->PathOffset =cpu_to_le16 (sizeof (struct smb2_tree_connect_req )); req->PathLength =cpu_to_le16 (unc_path_len -2 ); iov[1 ].iov_base = unc_path; iov[1 ].iov_len = unc_path_len;
@@ -4676,7 +4675,7 @@ int SMB2_query_directory_init(const unsigned int xid,memcpy (bufptr, &asteriks, len); req->FileNameOffset = -cpu_to_le16 (sizeof (struct smb2_query_directory_req ) -1 ); +cpu_to_le16 (sizeof (struct smb2_query_directory_req )); req->FileNameLength =cpu_to_le16 (len);/* * BB could be 30 bytes or so longer if we used SMB2 specific
@@ -4873,7 +4872,7 @@ SMB2_set_info_init(struct cifs_tcon *tcon, struct TCP_Server_Info *server, req->AdditionalInformation =cpu_to_le32 (additional_info); req->BufferOffset = -cpu_to_le16 (sizeof (struct smb2_set_info_req ) -1 ); +cpu_to_le16 (sizeof (struct smb2_set_info_req )); req->BufferLength =cpu_to_le32 (*size);memcpy (req->Buffer, *data, *size);
@@ -5105,9 +5104,9 @@ build_qfs_info_req(struct kvec *iov, struct cifs_tcon *tcon, req->VolatileFileId = volatile_fid;/* 1 for pad */ req->InputBufferOffset = -cpu_to_le16 (sizeof (struct smb2_query_info_req ) -1 ); +cpu_to_le16 (sizeof (struct smb2_query_info_req )); req->OutputBufferLength =cpu_to_le32 ( - outbuf_len +sizeof (struct smb2_query_info_rsp ) -1 ); + outbuf_len +sizeof (struct smb2_query_info_rsp )); iov->iov_base = (char *)req; iov->iov_len = total_len;
fs/cifs/smb2pdu.h +24 -18
150 lines not on this page · raw message
Signature · 1 line
2.33.8
Quoted 15 lines · On Fri, Jan 26, 2024 at 10:31:41PM +0300, [email protected] wrote:
After mounting a remote cifs resource, it becomes unavailable:
df:/mnt/sambashare: Resource temporarily unavailableIt was tested on the following Linux kernels:
Linux altlinux 5.10.208-std-def-alt1
Linux fedora 5.10.208-200.el8.x86_64The error appeared starting from kernel 5.10.206 after adding the commit [1] "smb: client: fix OOB in
SMB2_query_info_init()", in which the buffer length increases by 1 as a result of changes:
...(quoted diff) +1
-1- iov[0].iov_len = total_len - 1 + input_len; + iov[0].iov_len = len;
Quoted 8 lines
[1] https://patchwork.kernel.org/project/cifs-client/patch/[email protected]/
Error fixed by backported commits in next two patches adapted for the 5.10 kernel:
[PATCH 1/2] stddef: Introduce
DECLARE_FLEX_ARRAY()helper [PATCH 2/2] smb3: Replace smb2pdu 1-element arrays with flex-arrays
Now queued up, thanks.
greg k-h
Hi,
Adding more people to CC.(who have looked at this issue)
Quoted 8 lines · Greg KH
Quoted 7 lines · On Fri, Jan 26, 2024 at 10:31:41PM +0300, [email protected] wrote:
After mounting a remote cifs resource, it becomes unavailable:
df:/mnt/sambashare: Resource temporarily unavailableIt was tested on the following Linux kernels:
Linux altlinux 5.10.208-std-def-alt1
Linux fedora 5.10.208-200.el8.x86_64
Quoted 8 lines
Review
Quoted 8 lines
The error appeared starting from kernel 5.10.206 after adding the commit [1] "smb: client: fix OOB in
SMB2_query_info_init()", in which the buffer length increases by 1 as a result of changes:
...(quoted diff) +1
-1- iov[0].iov_len = total_len - 1 + input_len; + iov[0].iov_len = len;
We can reproduce this on 5.15.148(latest 5.15.y) and Mohamed reported this on 6.1.y, so we need backports there as well.
https://lore.kernel.org/all/[email protected]/
[root@vm1 xfstests-dev]# ./check -g quick -s smb3 TEST_DEV=//<SERVER_IP>/TEST is mounted but not a type cifs filesystem [root@vm1 xfstests-dev]# df
df: /mnt/test: Resource temporarily unavailable
This two patch series doesn't cleanly apply to 5.15.y.
Also I am unsure, which is the better approach to go with
Approach 1 - suggested by Paulo:
https://lore.kernel.org/all/[email protected]/
Approach 2 - this series
Pulling in [PATCH 2/2] smb3: Replace smb2pdu 1-element arrays with flex-arrays like this series did.
I think approach 1 is better as the changes are minimal, but please correct me if that seems wrong.
Thanks,
Harshit
Quoted 13 lines
Quoted 8 lines
[1] https://patchwork.kernel.org/project/cifs-client/patch/[email protected]/
Error fixed by backported commits in next two patches adapted for the 5.10 kernel:
[PATCH 1/2] stddef: Introduce
DECLARE_FLEX_ARRAY()helper [PATCH 2/2] smb3: Replace smb2pdu 1-element arrays with flex-arraysNow queued up, thanks.
greg k-h
Hi,
Quoted 1 line · Harshit Mogalapalli
We can reproduce this on 5.15.148(latest 5.15.y) and Mohamed reported this on 6.1.y, so we need backports there as well.
in the 6.1.72 kernel, this problem was fixed by the commit [1] "smb3: Replace smb2pdu 1-element arrays with flex-arrays", which was proposed in this series of patches.
[1] https://lore.kernel.org/all/2024010937-eggplant-bauble-d556@gregkh/T/
Signature · 2 lines
Regards, Vasiliy Kovalev
Hi Kovalev,
Quoted 7 lines · On 27/01/24 1:32 pm, [email protected] wrote:
Hi,
Quoted 1 line · Harshit Mogalapalli
We can reproduce this on 5.15.148(latest 5.15.y) and Mohamed reported this on 6.1.y, so we need backports there as well.
in the 6.1.72 kernel, this problem was fixed by the commit [1] "smb3: Replace smb2pdu 1-element arrays with flex-arrays", which was proposed in this series of patches.
Thanks for sharing this, I didnot notice that the above commit was backported to 6.1.72.
I think we still need fixing in 5.15.y as the commit eb3e28c1e89b ("smb3: Replace smb2pdu 1-element arrays with flex-arrays") is not in 5.15.148
Thanks,
Harshit
Quoted 3 lines
[1] https://lore.kernel.org/all/2024010937-eggplant-bauble-d556@gregkh/T/
Quoted 13 lines · Harshit Mogalapalli
Hi Kovalev,
Quoted 10 lines · On 27/01/24 1:32 pm, [email protected] wrote:
Hi,
Quoted 2 lines · Harshit Mogalapalli
We can reproduce this on 5.15.148(latest 5.15.y) and Mohamed reported this on 6.1.y, so we need backports there as well.
in the 6.1.72 kernel, this problem was fixed by the commit [1] "smb3:
Replace smb2pdu 1-element arrays with flex-arrays", which was proposed in this series of patches.
Quoted 5 lines
Thanks for sharing this, I didnot notice that the above commit was backported to 6.1.72.
I think we still need fixing in 5.15.y as the commit
eb3e28c1e89b("smb3:
Replace smb2pdu 1-element arrays with flex-arrays") is not in 5.15.148
Patches gladly accepted :)
Quoted 42 lines · Harshit Mogalapalli
Hi,
Adding more people to CC.(who have looked at this issue)
Quoted 8 lines · Greg KH
Quoted 7 lines · On Fri, Jan 26, 2024 at 10:31:41PM +0300, [email protected] wrote:
After mounting a remote cifs resource, it becomes unavailable:
df:/mnt/sambashare: Resource temporarily unavailableIt was tested on the following Linux kernels:
Linux altlinux 5.10.208-std-def-alt1
Linux fedora 5.10.208-200.el8.x86_64Quoted 8 lines
Review
Quoted 8 lines
The error appeared starting from kernel 5.10.206 after adding the commit [1] "smb: client: fix OOB in
SMB2_query_info_init()", in which the buffer length increases by 1 as a result of changes:
...(quoted diff) +1
-1- iov[0].iov_len = total_len - 1 + input_len; + iov[0].iov_len = len;We can reproduce this on 5.15.148(latest 5.15.y) and Mohamed reported this on 6.1.y, so we need backports there as well.
https://lore.kernel.org/all/[email protected]/
[root@vm1 xfstests-dev]# ./check -g quick -s smb3 TEST_DEV=//<SERVER_IP>/TEST is mounted but not a type cifs filesystem [root@vm1 xfstests-dev]# df
df:/mnt/test: Resource temporarily unavailableThis two patch series doesn't cleanly apply to 5.15.y.
Also I am unsure, which is the better approach to go with
Approach 1 - suggested by Paulo:
https://lore.kernel.org/all/[email protected]/
Quoted 6 lines
Approach 2 - this series
Pulling in [PATCH 2/2] smb3: Replace smb2pdu 1-element arrays with flex-arrays like this series did.I think approach 1 is better as the changes are minimal, but please correct me if that seems wrong.
Yes - Paulo's fix looks simple
Signature · 3 lines
Thanks, Steve