Skip to contentSkip to conversation

smb: some potential bugfixes

PATCH In review Greg Kroah-Hartman 4 participants 16 messages Apr 6 – Apr 9 1 message not in the archive

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover smb: some potential bugfixes

I spent the time exercising some new fuzzing tools on the ksmbd and smb code purely because it's something that is simple to set up and test locally with virtual machines, and in doing so, potentially found some minor problems for when you have an "untrusted" client.

Here's some fixes for what I happened to notice. They pass my very limited testing here, but please don't trust them at all and verify that I'm not just making this all up before accepting them.

thanks!

greg k-h

Greg Kroah-Hartman (2):
smb: client: fix off-by-8 bounds check in check_wsl_eas()
smb: client: fix OOB reads parsing symlink error response

 fs/smb/client/smb2file.c  | 20 ++++++++++++--------
 fs/smb/client/smb2inode.c |  2 +-
 2 files changed, 13 insertions(+), 9 deletions(-)
Signature · 1 line
2.53.0
#1 lore raw

PATCH 1/2 smb: client: fix off-by-8 bounds check in check_wsl_eas()

The bounds check uses (u8 *)ea + nlen + 1 + vlen as the end of the EA name and value, but ea_data sits at offset sizeof(struct smb2_file_full_ea_info) = 8 from ea, not at offset 0. The strncmp() later reads ea->ea_data[0..nlen-1] and the value bytes follow at ea_data[nlen+1..nlen+vlen], so the actual end is ea->ea_data + nlen + 1
+ vlen. Isn't pointer math fun?

The earlier check (u8 *)ea > end - sizeof(*ea) only guarantees the 8-byte header is in bounds, but since the last EA is placed within 8 bytes of the end of the response, the name and value bytes are read past the end of iov.

Fix this mess all up by using ea->ea_data as the base for the bounds check.

An "untrusted" server can use this to leak up to 8 bytes of kernel heap into the EA name comparison and influence which WSL xattr the data is interpreted as.

Cc Steve French <[email protected]>
Cc Paulo Alcantara <[email protected]>
Cc Ronnie Sahlberg <[email protected]>
Cc Shyam Prasad N <[email protected]>
Cc Tom Talpey <[email protected]>
Cc Bharath SM <[email protected]>
Cc stable <[email protected]>
Assisted-by gregkh_clanker_t1000
---
 fs/smb/client/smb2inode.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/smb/client/smb2inode.c +1 -1
index 364bdcff9c9d..fe1c9d776580 100644
@@ -128,7 +128,7 @@ static int check_wsl_eas(struct kvec *rsp_iov)
 		nlen = ea->ea_name_length;
 		vlen = le16_to_cpu(ea->ea_value_length);
 		if (nlen != SMB2_WSL_XATTR_NAME_LEN ||
-		    (u8 *)ea + nlen + 1 + vlen > end)
+		    (u8 *)ea->ea_data + nlen + 1 + vlen > end)
 			return -EINVAL;
 
 		switch (vlen) {
Signature · 1 line
2.53.0
reviewer
#2 lore raw

Re: [PATCH 1/2] smb: client: fix off-by-8 bounds check in check_wsl_eas()

Sashiko reported another out-of-bounds issue: https://sashiko.dev/#/patchset/2026040635-banking-unsoiled-3250@gregkh

Should we add the following checks in check_wsl_eas()?

```
--- a/fs/smb/client/smb2inode.c
+++ b/fs/smb/client/smb2inode.c
@@ -121,6 +121,9 @@ static int check_wsl_eas(struct kvec *rsp_iov)
        ea = (void *)((u8 *)rsp_iov->iov_base +
                      le16_to_cpu(rsp->OutputBufferOffset));
        end = (u8 *)rsp_iov->iov_base + rsp_iov->iov_len;
+       if (ea + outlen > end)
+               return -EINVAL;
+
        for (;;) {
                if ((u8 *)ea > end - sizeof(*ea))
                        return -EINVAL;
```
Quoted 47 lines · Greg Kroah-Hartman

The bounds check uses (u8 *)ea + nlen + 1 + vlen as the end of the EA name and value, but ea_data sits at offset sizeof(struct smb2_file_full_ea_info) = 8 from ea, not at offset 0. The strncmp() later reads ea->ea_data[0..nlen-1] and the value bytes follow at ea_data[nlen+1..nlen+vlen], so the actual end is ea->ea_data + nlen + 1
+ vlen. Isn't pointer math fun?

The earlier check (u8 *)ea > end - sizeof(*ea) only guarantees the 8-byte header is in bounds, but since the last EA is placed within 8 bytes of the end of the response, the name and value bytes are read past the end of iov.

Fix this mess all up by using ea->ea_data as the base for the bounds check.

An "untrusted" server can use this to leak up to 8 bytes of kernel heap into the EA name comparison and influence which WSL xattr the data is interpreted as.

Cc Steve French <[email protected]>
Cc Paulo Alcantara <[email protected]>
Cc Ronnie Sahlberg <[email protected]>
Cc Shyam Prasad N <[email protected]>
Cc Tom Talpey <[email protected]>
Cc Bharath SM <[email protected]>
Cc stable <[email protected]>
Assisted-by gregkh_clanker_t1000
---
  fs/smb/client/smb2inode.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
fs/smb/client/smb2inode.c +1 -1
index 364bdcff9c9d..fe1c9d776580 100644
@@ -128,7 +128,7 @@ static int check_wsl_eas(struct kvec *rsp_iov)
  		nlen = ea->ea_name_length;
  		vlen = le16_to_cpu(ea->ea_value_length);
  		if (nlen != SMB2_WSL_XATTR_NAME_LEN ||
-		    (u8 *)ea + nlen + 1 + vlen > end)
+		    (u8 *)ea->ea_data + nlen + 1 + vlen > end)
  			return -EINVAL;
 
  		switch (vlen) {
Signature · 3 lines
ChenXiaoSong <[email protected]>
Chinese Homepage: chenxiaosong.com
English Homepage: chenxiaosong.com/en
#3 lore raw

Re: [PATCH 1/2] smb: client: fix off-by-8 bounds check in check_wsl_eas()

Quoted 9 lines · ChenXiaoSong

Sashiko reported another out-of-bounds issue:
https://sashiko.dev/#/patchset/2026040635-banking-unsoiled-3250@gregkh

Should we add the following checks in check_wsl_eas()?

```

fs/smb/client/smb2inode.c static int check_wsl_eas(struct kvec *rsp_iov) · line 125

Quoted 5 lines
fs/smb/client/smb2inode.c +2 -0
@@ -121,6 +121,9 @@ static int check_wsl_eas(struct kvec *rsp_iov)
        ea = (void *)((u8 *)rsp_iov->iov_base +
                      le16_to_cpu(rsp->OutputBufferOffset));
        end = (u8 *)rsp_iov->iov_base + rsp_iov->iov_len;
+       if (ea + outlen > end)
+               return -EINVAL;

Then you would miss any "first" structures here, as I think the for loop catches this later on with the line:

fs/smb/client/smb2inode.c static int check_wsl_eas(struct kvec *rsp_iov) · line 129

Quoted 4 lines
fs/smb/client/smb2inode.c +1 -0
+
        for (;;) {
                if ((u8 *)ea > end - sizeof(*ea))
                        return -EINVAL;

That one, right?

Or am I misreading this?

Pointer math is "fun" :(

thanks,

greg k-h

reviewer
#4 lore raw

Re: [PATCH 1/2] smb: client: fix off-by-8 bounds check in check_wsl_eas()

The for loop does not seem to catch cases where outlen is excessively large. In such cases, smb2_compound_op() would use this large outlen to memcpy(), which could lead to OOB.

```
smb2_compound_op()
{
	...
	size[0] = outlen; // very large
	check_wsl_eas()
	memcpy(..., outlen) // out-of-bounds
	...
}
```
Quoted 34 lines · Greg Kroah-Hartman
Quoted 9 lines · ChenXiaoSong

Sashiko reported another out-of-bounds issue:
https://sashiko.dev/#/patchset/2026040635-banking-unsoiled-3250@gregkh

Should we add the following checks in check_wsl_eas()?

```

fs/smb/client/smb2inode.c static int check_wsl_eas(struct kvec *rsp_iov) · line 125

Quoted 5 lines
fs/smb/client/smb2inode.c +2 -0
@@ -121,6 +121,9 @@ static int check_wsl_eas(struct kvec *rsp_iov)
         ea = (void *)((u8 *)rsp_iov->iov_base +
                       le16_to_cpu(rsp->OutputBufferOffset));
         end = (u8 *)rsp_iov->iov_base + rsp_iov->iov_len;
+       if (ea + outlen > end)
+               return -EINVAL;

Then you would miss any "first" structures here, as I think the for loop catches this later on with the line:

fs/smb/client/smb2inode.c static int check_wsl_eas(struct kvec *rsp_iov) · line 129

Quoted 4 lines
fs/smb/client/smb2inode.c +1 -0
+
         for (;;) {
                 if ((u8 *)ea > end - sizeof(*ea))
                         return -EINVAL;

That one, right?

Or am I misreading this?

Pointer math is "fun" :(

thanks,

greg k-h

Signature · 3 lines
ChenXiaoSong <[email protected]>
Chinese Homepage: chenxiaosong.com
English Homepage: chenxiaosong.com/en
#5 lore raw

Re: [PATCH 1/2] smb: client: fix off-by-8 bounds check in check_wsl_eas()

Quoted 4 lines · ChenXiaoSong

The for loop does not seem to catch cases where outlen is excessively large. In such cases, smb2_compound_op() would use this large outlen to memcpy(), which could lead to OOB.

Quoted 10 lines
```
smb2_compound_op()
{
	...
	size[0] = outlen; // very large
	check_wsl_eas()
	memcpy(..., outlen) // out-of-bounds
	...
}
```

Ah, I missed the caller site. Yeah, probably a good thing to check as well, want to make up a patch?

thanks,
greg k-h

reviewer
#6 lore raw

Re: [PATCH 1/2] smb: client: fix off-by-8 bounds check in check_wsl_eas()

Could you send v2 with this check added?

I am currently reviewing another patch. Perhaps you could wait until I have finished reviewing it.

Quoted 2 lines · Greg Kroah-Hartman

Ah, I missed the caller site. Yeah, probably a good thing to check as well, want to make up a patch?

Signature · 3 lines
ChenXiaoSong <[email protected]>
Chinese Homepage: chenxiaosong.com
English Homepage: chenxiaosong.com/en
#7 lore raw

Re: [PATCH 1/2] smb: client: fix off-by-8 bounds check in check_wsl_eas()

Quoted 1 line · ChenXiaoSong

Could you send v2 with this check added?

Looks like this was already accepted, and the additional check would be just that, an additional check :)

Quoted 2 lines

I am currently reviewing another patch. Perhaps you could wait until I have finished reviewing it.

I have no problem waiting, but as you found this one, I want to give you proper credit for the find/fix.

thanks,

greg k-h

reviewer
#8 lore raw

Re: [PATCH 1/2] smb: client: fix off-by-8 bounds check in check_wsl_eas()

Sashiko found it, I just checked and confirmed it.

Could you create another patch? I enjoy the review process more than creating patches :)

Quoted 15 lines · Greg Kroah-Hartman
Quoted 1 line · ChenXiaoSong

Could you send v2 with this check added?

Looks like this was already accepted, and the additional check would be just that, an additional check :)

Quoted 2 lines

I am currently reviewing another patch. Perhaps you could wait until I have finished reviewing it.

I have no problem waiting, but as you found this one, I want to give you proper credit for the find/fix.

thanks,

greg k-h

Signature · 3 lines
ChenXiaoSong <[email protected]>
Chinese Homepage: chenxiaosong.com
English Homepage: chenxiaosong.com/en
#9 lore raw

Re: [PATCH 1/2] smb: client: fix off-by-8 bounds check in check_wsl_eas()

Quoted 4 lines · ChenXiaoSong

Sashiko found it, I just checked and confirmed it.

Could you create another patch? I enjoy the review process more than creating patches :)

Will do, I'll add that to my queue, but might take a few days if someone wants to "beat" me to it, I'll have no objection at all :)

thanks,

greg k-h

reviewer
#10 lore raw

Re: [PATCH 1/2] smb: client: fix off-by-8 bounds check in check_wsl_eas()

In reply to Greg Kroah-Hartman

Hi Steve,

Should we use get_unaligned_le32() to get unaligned data?

Quoted 20 lines

static inline kuid_t wsl_make_kuid(struct cifs_sb_info *cifs_sb,
void *ptr)

{                                                               
        u32 uid = le32_to_cpu(*(__le32 *)ptr);
        ...
}                                                               

Additionally, does parsing these extended attributes cause unaligned memory accesses?
When parsing WSL extended attributes, the code derives the value pointer at an offset of ea_name_length (which is 6) plus 1 from ea_data. Since ea is a 4-byte aligned structure, the value sits at an unaligned offset. Helper functions like wsl_make_kuid() explicitly cast this unaligned pointer to a 32-bit type and dereference it:

    le32_to_cpu(*(__le32 *)v);

Could this trigger an unaligned access exception and crash the kernel on architectures with strict alignment requirements?

Signature · 3 lines
ChenXiaoSong <[email protected]>
Chinese Homepage: chenxiaosong.com
English Homepage: chenxiaosong.com/en
#11 lore raw

Re: [PATCH 1/2] smb: client: fix off-by-8 bounds check in check_wsl_eas()

In reply to Greg Kroah-Hartman

[PATCH] smb: client: add bounds checks when iterating SMB1 WSL EAs

Add boundary and overflow checks when traversing extended attribute (EA) entries for SMB1 WSL reparse points in cifs_query_path_info(), to prevent out-of-bounds memory accesses from malformed next_entry_offset values.

This is necessary because SMB1 WSL EA iteration does not currently have any explicit bounds checking, unlike SMB2 path which has check_wsl_eas(). The new checks ensure the current EA pointer plus its size does not exceed the end of the EA buffer during iteration and before parsing.

---
 fs/smb/client/smb1ops.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)
fs/smb/client/smb1ops.c +10 -2
index 9694117050a6..940bae49b2cc 100644
@@ -645,13 +645,17 @@ static int cifs_query_path_info(const unsigned int xid,
 	if (!rc && data->reparse_point) {
 		struct smb2_file_full_ea_info *ea;
 		u32 next = 0;
+		u8 *eas_end = data->wsl.eas + data->wsl.eas_len;
 
 		ea = (struct smb2_file_full_ea_info *)data->wsl.eas;
 		do {
 			ea = (void *)((u8 *)ea + next);
+			if ((u8 *)ea + sizeof(*ea) > eas_end)
+				break;
 			next = le32_to_cpu(ea->next_entry_offset);
 		} while (next);
-		if (le16_to_cpu(ea->ea_value_length)) {
+		if ((u8 *)ea + sizeof(*ea) <= eas_end &&
+		    le16_to_cpu(ea->ea_value_length)) {
 			ea->next_entry_offset = cpu_to_le32(ALIGN(sizeof(*ea) +
 						ea->ea_name_length + 1 +
 						le16_to_cpu(ea->ea_value_length), 4));
@@ -691,13 +695,17 @@ static int cifs_query_path_info(const unsigned int xid,
 	if (!rc && data->reparse_point) {
 		struct smb2_file_full_ea_info *ea;
 		u32 next = 0;
+		u8 *eas_end = data->wsl.eas + data->wsl.eas_len;
 
 		ea = (struct smb2_file_full_ea_info *)data->wsl.eas;
 		do {
 			ea = (void *)((u8 *)ea + next);
+			if ((u8 *)ea + sizeof(*ea) > eas_end)
+				break;
 			next = le32_to_cpu(ea->next_entry_offset);
 		} while (next);
-		if (le16_to_cpu(ea->ea_value_length)) {
+		if ((u8 *)ea + sizeof(*ea) <= eas_end &&
+		    le16_to_cpu(ea->ea_value_length)) {
 			ea->next_entry_offset = cpu_to_le32(ALIGN(sizeof(*ea) +
 						ea->ea_name_length + 1 +
 						le16_to_cpu(ea->ea_value_length), 4));
Signature · 1 line
2.43.0
Quoted 55 lines · Greg Kroah-Hartman

The bounds check uses (u8 *)ea + nlen + 1 + vlen as the end of the EA name and value, but ea_data sits at offset sizeof(struct smb2_file_full_ea_info) = 8 from ea, not at offset 0. The strncmp() later reads ea->ea_data[0..nlen-1] and the value bytes follow at ea_data[nlen+1..nlen+vlen], so the actual end is ea->ea_data + nlen + 1
+ vlen. Isn't pointer math fun?

The earlier check (u8 *)ea > end - sizeof(*ea) only guarantees the 8-byte header is in bounds, but since the last EA is placed within 8 bytes of the end of the response, the name and value bytes are read past the end of iov.

Fix this mess all up by using ea->ea_data as the base for the bounds check.

An "untrusted" server can use this to leak up to 8 bytes of kernel heap into the EA name comparison and influence which WSL xattr the data is interpreted as.

Cc Steve French <[email protected]>
Cc Paulo Alcantara <[email protected]>
Cc Ronnie Sahlberg <[email protected]>
Cc Shyam Prasad N <[email protected]>
Cc Tom Talpey <[email protected]>
Cc Bharath SM <[email protected]>
Cc stable <[email protected]>
Assisted-by gregkh_clanker_t1000

---
fs/smb/client/smb2inode.c | 2 +-

1 file changed, 1 insertion(+), 1 deletion(-)
fs/smb/client/smb2inode.c +1 -1
index 364bdcff9c9d..fe1c9d776580 100644
@@ -128,7 +128,7 @@ static int check_wsl_eas(struct kvec *rsp_iov)
 		nlen = ea->ea_name_length;
 		vlen = le16_to_cpu(ea->ea_value_length);
 		if (nlen != SMB2_WSL_XATTR_NAME_LEN ||
-		    (u8 *)ea + nlen + 1 + vlen > end)
+		    (u8 *)ea->ea_data + nlen + 1 + vlen > end)
 			return -EINVAL;
 
 		switch (vlen) {
Signature · 1 line
2.53.0
#12 lore raw

PATCH 2/2 smb: client: fix OOB reads parsing symlink error response

In reply to Greg Kroah-Hartman

When a CREATE returns STATUS_STOPPED_ON_SYMLINK, smb2_check_message() returns success without any length validation, leaving the symlink parsers as the only defense against an untrusted server.

symlink_data() walks SMB 3.1.1 error contexts with the loop test "p < end", but reads p->ErrorId at offset 4 and p->ErrorDataLength at offset
0. When the server-controlled ErrorDataLength advances p to within 1-7 bytes of end, the next iteration will read past it. When the matching context is found, sym->SymLinkErrorTag is read at offset 4 from p->ErrorContextData with no check that the symlink header itself fits.

smb2_parse_symlink_response() then bounds-checks the substitute name using SMB2_SYMLINK_STRUCT_SIZE as the offset of PathBuffer from iov_base. That value is computed as sizeof(smb2_err_rsp) + sizeof(smb2_symlink_err_rsp), which is correct only when ErrorContextCount == 0.

With at least one error context the symlink data sits 8 bytes deeper, and each skipped non-matching context shifts it further by 8 + ALIGN(ErrorDataLength, 8). The check is too short, allowing the substitute name read to run past iov_len. The out-of-bound heap bytes are UTF-16-decoded into the symlink target and returned to userspace via readlink(2).

Fix this all up by making the loops test require the full context header to fit, rejecting sym if its header runs past end, and bound the substitute name against the actual position of sym->PathBuffer rather than a fixed offset.

Because sub_offs and sub_len are 16bits, the pointer math will not overflow here with the new greater-than.

Cc Steve French <[email protected]>
Cc Paulo Alcantara <[email protected]>
Cc Ronnie Sahlberg <[email protected]>
Cc Shyam Prasad N <[email protected]>
Cc Tom Talpey <[email protected]>
Cc Bharath SM <[email protected]>
Cc stable <[email protected]>
Assisted-by gregkh_clanker_t1000
---
 fs/smb/client/smb2file.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)
fs/smb/client/smb2file.c +12 -8
index ed651c946251..b292aa94a593 100644
@@ -27,10 +27,11 @@ static struct smb2_symlink_err_rsp *symlink_data(const struct kvec *iov)
 {
 	struct smb2_err_rsp *err = iov->iov_base;
 	struct smb2_symlink_err_rsp *sym = ERR_PTR(-EINVAL);
+	u8 *end = (u8 *)err + iov->iov_len;
 	u32 len;
 
 	if (err->ErrorContextCount) {
-		struct smb2_error_context_rsp *p, *end;
+		struct smb2_error_context_rsp *p;
 
 		len = (u32)err->ErrorContextCount * (offsetof(struct smb2_error_context_rsp,
 							      ErrorContextData) +
@@ -39,8 +40,7 @@ static struct smb2_symlink_err_rsp *symlink_data(const struct kvec *iov)
 			return ERR_PTR(-EINVAL);
 
 		p = (struct smb2_error_context_rsp *)err->ErrorData;
-		end = (struct smb2_error_context_rsp *)((u8 *)err + iov->iov_len);
-		do {
+		while ((u8 *)p + sizeof(*p) <= end) {
 			if (le32_to_cpu(p->ErrorId) == SMB2_ERROR_ID_DEFAULT) {
 				sym = (struct smb2_symlink_err_rsp *)p->ErrorContextData;
 				break;
@@ -50,14 +50,16 @@ static struct smb2_symlink_err_rsp *symlink_data(const struct kvec *iov)
 
 			len = ALIGN(le32_to_cpu(p->ErrorDataLength), 8);
 			p = (struct smb2_error_context_rsp *)(p->ErrorContextData + len);
-		} while (p < end);
+		}
 	} else if (le32_to_cpu(err->ByteCount) >= sizeof(*sym) &&
 		   iov->iov_len >= SMB2_SYMLINK_STRUCT_SIZE) {
 		sym = (struct smb2_symlink_err_rsp *)err->ErrorData;
 	}
 
-	if (!IS_ERR(sym) && (le32_to_cpu(sym->SymLinkErrorTag) != SYMLINK_ERROR_TAG ||
-			     le32_to_cpu(sym->ReparseTag) != IO_REPARSE_TAG_SYMLINK))
+	if (!IS_ERR(sym) &&
+	    ((u8 *)sym + sizeof(*sym) > end ||
+	     le32_to_cpu(sym->SymLinkErrorTag) != SYMLINK_ERROR_TAG ||
+	     le32_to_cpu(sym->ReparseTag) != IO_REPARSE_TAG_SYMLINK))
 		sym = ERR_PTR(-EINVAL);
 
 	return sym;
@@ -128,8 +130,10 @@ int smb2_parse_symlink_response(struct cifs_sb_info *cifs_sb, const struct kvec
 	print_len = le16_to_cpu(sym->PrintNameLength);
 	print_offs = le16_to_cpu(sym->PrintNameOffset);
 
-	if (iov->iov_len < SMB2_SYMLINK_STRUCT_SIZE + sub_offs + sub_len ||
-	    iov->iov_len < SMB2_SYMLINK_STRUCT_SIZE + print_offs + print_len)
+	if ((char *)sym->PathBuffer + sub_offs + sub_len >
+		(char *)iov->iov_base + iov->iov_len ||
+	    (char *)sym->PathBuffer + print_offs + print_len >
+		(char *)iov->iov_base + iov->iov_len)
 		return -EINVAL;
 
 	return smb2_parse_native_symlink(path,
Signature · 1 line
2.53.0
reviewer
#13 lore raw

Re: [PATCH 2/2] smb: client: fix OOB reads parsing symlink error response

Sashiko reported the following out-of-bounds issue. I have checked and confirmed that this indeed causes an OOB access.

When create fails on symlink, len in smb2_check_message() may be smaller than calc_len. The function flow is as follows:

```
smb2_check_message()
  // ensure StructureSize2 is 9
  if (... pdu->StructureSize2 != SMB2_ERROR_STRUCTURE_SIZE2_LE ...) // 
false
  smb2_calc_size()
    len = le16_to_cpu(shdr->StructureSize) == 64
    len += le16_to_cpu(pdu->StructureSize2) == 64 + 9
    smb2_get_data_area_len
      if (shdr->StructureSize == 9) // true, return NULL
  calc_len == 64 + 9
  if (len != calc_len) { // true
  /* create failed on symlink */
  if (command == SMB2_CREATE_HE && shdr->Status == 
STATUS_STOPPED_ON_SYMLINK) // true
```

Should we add the following check? Or check it in symlink_data()?

```
-- a/fs/smb/client/smb2misc.c
+++ b/fs/smb/client/smb2misc.c
@@ -241,7 +241,8 @@ smb2_check_message(char *buf, unsigned int pdu_len, unsigned int len,
        if (len != calc_len) {
                /* create failed on symlink */
                if (command == SMB2_CREATE_HE &&
-                   shdr->Status == STATUS_STOPPED_ON_SYMLINK)
+                   shdr->Status == STATUS_STOPPED_ON_SYMLINK &&
+                   len > calc_len)
                        return 0;
                /* Windows 7 server returns 24 bytes more */
                if (calc_len + 24 == len && command == 
SMB2_OPLOCK_BREAK_HE)

```
Quoted 16 lines

fs/smb/client/smb2file.c static struct smb2_symlink_err_rsp *symlink_data(const struct kvec *iov) · line 33

Quoted 10 lines
fs/smb/client/smb2file.c +1 -0
@@ -27,10 +27,11 @@ static struct smb2_symlink_err_rsp *symlink_data(const struct kvec *iov)
 {
 	struct smb2_err_rsp *err = iov->iov_base;
 	struct smb2_symlink_err_rsp *sym = ERR_PTR(-EINVAL);
+	u8 *end = (u8 *)err + iov->iov_len;
 	u32 len;
 
 	if (err->ErrorContextCount) {

Since smb2_check_message() returns success without length validation for the symlink error response, is it possible for iov->iov_len to be smaller than sizeof(struct smb2_err_rsp)?
If the buffer only contains the base SMB2 header (64 bytes), does accessing err->ErrorContextCount (at offset 66) or err->ByteCount later in this function cause an out-of-bounds read?

Signature · 3 lines
ChenXiaoSong <[email protected]>
Chinese Homepage: chenxiaosong.com
English Homepage: chenxiaosong.com/en
maintainer Reviewed-by on the v1 series
#14 lore raw
In reply to Greg Kroah-Hartman
Quoted 8 lines · Greg Kroah-Hartman

I spent the time exercising some new fuzzing tools on the ksmbd and smb code purely because it's something that is simple to set up and test locally with virtual machines, and in doing so, potentially found some minor problems for when you have an "untrusted" client.

Here's some fixes for what I happened to notice. They pass my very limited testing here, but please don't trust them at all and verify that I'm not just making this all up before accepting them.

A message that is not in the archive (sent off-list or not yet received): CAH2r5mvXFzWxs70Jmg=yWddmcHs+Bpf8eiBUvyc4oOMLzdCY7w@mail.gmail.com