nbd/client: Use smarter assert
PATCH v2 1/1 nbd/client: Use smarter assert
Assigning strlen() to a uint32_t and then asserting that it isn't too large doesn't catch the case of an input string 4G in length. Thankfully, the incoming strings can never be that large: if the export name or query is reflecting a string the client got from the server, we already guarantee that we dropped the NBD connection if the server sent more than 32M in a single reply to our NBD_OPT_* request; if the export name is coming from qemu, nbd_receive_negotiate() asserted that strlen(info->name) <= NBD_MAX_STRING_SIZE; and similarly, a query string via x->dirty_bitmap coming from the user was bounds-checked in either qemu-nbd or by the limitations of QMP. Still, it doesn't hurt to be more explicit in how we write our assertions to not have to analyze whether inadvertent wraparound is possible.
---
v2: update subject line and commit message to reflect file being touched; adjust a second nearby assertion with the same issue
nbd/client.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
nbd/client.c +2 -2
@@ -658,11 +658,11 @@ static int nbd_send_meta_query(QIOChannel *ioc, uint32_t opt,char *p; data_len =sizeof (export_len) + export_len +sizeof (queries); -assert (export_len <=NBD_MAX_STRING_SIZE ); +assert (strlen (export) <=NBD_MAX_STRING_SIZE );if (query) { query_len =strlen (query); data_len +=sizeof (query_len) + query_len; -assert (query_len <=NBD_MAX_STRING_SIZE ); +assert (strlen (query) <=NBD_MAX_STRING_SIZE ); }else {assert (opt ==NBD_OPT_LIST_META_CONTEXT ); }
Signature · 1 line
2.37.3
Quoted 14 lines · * Eric Blake ([email protected]) wrote:
Assigning
strlen()to auint32_tand then asserting that it isn't too large doesn't catch the case of an input string 4G in length. Thankfully, the incoming strings can never be that large: if the export name or query is reflecting a string the client got from the server, we already guarantee that we dropped the NBD connection if the server sent more than 32M in a single reply to our NBD_OPT_* request; if the export name is coming from qemu,nbd_receive_negotiate()asserted thatstrlen(info->name)<=NBD_MAX_STRING_SIZE; and similarly, a query string viax->dirty_bitmapcoming from the user was bounds-checked in either qemu-nbd or by the limitations of QMP. Still, it doesn't hurt to be more explicit in how we write our assertions to not have to analyze whether inadvertent wraparound is possible.
Quoted 3 lines
Fixes 93676c88 ("nbd: Don't send oversize strings", v4.2.0)Reported-by Dr. David Alan Gilbert <[email protected]>Signed-off-by Eric Blake <[email protected]>
Quoted 29 lines
---
v2: update subject line and commit message to reflect file being touched; adjust a second nearby assertion with the same issue
nbd/client.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)nbd/client.c +2
-2@@ -658,11 +658,11 @@ static int nbd_send_meta_query(QIOChannel *ioc, uint32_t opt,char *p; data_len =sizeof (export_len) + export_len +sizeof (queries); -assert (export_len <=NBD_MAX_STRING_SIZE ); +assert (strlen (export) <=NBD_MAX_STRING_SIZE );if (query) { query_len =strlen (query); data_len +=sizeof (query_len) + query_len; -assert (query_len <=NBD_MAX_STRING_SIZE ); +assert (strlen (query) <=NBD_MAX_STRING_SIZE ); }else {assert (opt ==NBD_OPT_LIST_META_CONTEXT ); }Signature · 1 line
2.37.3
Signature · 1 line
Dr. David Alan Gilbert / [email protected] / Manchester, UK
Quoted 33 lines · Eric Blake
Assigning
strlen()to auint32_tand then asserting that it isn't too large doesn't catch the case of an input string 4G in length. Thankfully, the incoming strings can never be that large: if the export name or query is reflecting a string the client got from the server, we already guarantee that we dropped the NBD connection if the server sent more than 32M in a single reply to our NBD_OPT_* request; if the export name is coming from qemu,nbd_receive_negotiate()asserted thatstrlen(info->name)<=NBD_MAX_STRING_SIZE; and similarly, a query string viax->dirty_bitmapcoming from the user was bounds-checked in either qemu-nbd or by the limitations of QMP. Still, it doesn't hurt to be more explicit in how we write our assertions to not have to analyze whether inadvertent wraparound is possible.Fixes 93676c88 ("nbd: Don't send oversize strings", v4.2.0)Reported-by Dr. David Alan Gilbert <[email protected]>Signed-off-by Eric Blake <[email protected]>---
v2: update subject line and commit message to reflect file being touched; adjust a second nearby assertion with the same issue
nbd/client.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)nbd/client.c +0
-0@@ -658,11 +658,11 @@ static int nbd_send_meta_query(QIOChannel *ioc, uint32_t opt,char *p; data_len =sizeof (export_len) + export_len +sizeof (queries);
nbd/client.c static int nbd_send_meta_query(QIOChannel *ioc, uint32_t opt, · line 668
Quoted 10 lines
nbd/client.c +2
-2-assert (export_len <=NBD_MAX_STRING_SIZE ); +assert (strlen (export) <=NBD_MAX_STRING_SIZE );if (query) { query_len =strlen (query); data_len +=sizeof (query_len) + query_len; -assert (query_len <=NBD_MAX_STRING_SIZE ); +assert (strlen (query) <=NBD_MAX_STRING_SIZE ); }else {assert (opt ==NBD_OPT_LIST_META_CONTEXT ); }
Nitpicking (pre-existing) the assertions could be moved before the assignations.
Quoted 33 lines · Eric Blake
Assigning
strlen()to auint32_tand then asserting that it isn't too large doesn't catch the case of an input string 4G in length. Thankfully, the incoming strings can never be that large: if the export name or query is reflecting a string the client got from the server, we already guarantee that we dropped the NBD connection if the server sent more than 32M in a single reply to our NBD_OPT_* request; if the export name is coming from qemu,nbd_receive_negotiate()asserted thatstrlen(info->name)<=NBD_MAX_STRING_SIZE; and similarly, a query string viax->dirty_bitmapcoming from the user was bounds-checked in either qemu-nbd or by the limitations of QMP. Still, it doesn't hurt to be more explicit in how we write our assertions to not have to analyze whether inadvertent wraparound is possible.Fixes 93676c88 ("nbd: Don't send oversize strings", v4.2.0)Reported-by Dr. David Alan Gilbert <[email protected]>Signed-off-by Eric Blake <[email protected]>---
v2: update subject line and commit message to reflect file being touched; adjust a second nearby assertion with the same issue
nbd/client.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)nbd/client.c +0
-0@@ -658,11 +658,11 @@ static int nbd_send_meta_query(QIOChannel *ioc, uint32_t opt,char *p; data_len =sizeof (export_len) + export_len +sizeof (queries);
nbd/client.c static int nbd_send_meta_query(QIOChannel *ioc, uint32_t opt, · line 668
Quoted 10 lines
nbd/client.c +2
-2-assert (export_len <=NBD_MAX_STRING_SIZE ); +assert (strlen (export) <=NBD_MAX_STRING_SIZE );if (query) { query_len =strlen (query); data_len +=sizeof (query_len) + query_len; -assert (query_len <=NBD_MAX_STRING_SIZE ); +assert (strlen (query) <=NBD_MAX_STRING_SIZE ); }else {assert (opt ==NBD_OPT_LIST_META_CONTEXT ); }
I'm a bit late, and this should work as is.
Still, for me it's a bit strange: you point to the fact that we probably overflow uint32_t variable. But we keep this fact hidden in the code. So, everyone who read should guess "aha, this extra strlen here is because the previous one may overflow the variable)".
Could we use strnlen() instead of strlen()? That would be also more effective.
Signature · 2 lines
Best regards, Vladimir
Quoted 47 lines · Vladimir Sementsov-Ogievskiy
Quoted 33 lines · Eric Blake
Assigning
strlen()to auint32_tand then asserting that it isn't too large doesn't catch the case of an input string 4G in length. Thankfully, the incoming strings can never be that large: if the export name or query is reflecting a string the client got from the server, we already guarantee that we dropped the NBD connection if the server sent more than 32M in a single reply to our NBD_OPT_* request; if the export name is coming from qemu,nbd_receive_negotiate()asserted thatstrlen(info->name)<=NBD_MAX_STRING_SIZE; and similarly, a query string viax->dirty_bitmapcoming from the user was bounds-checked in either qemu-nbd or by the limitations of QMP. Still, it doesn't hurt to be more explicit in how we write our assertions to not have to analyze whether inadvertent wraparound is possible.Fixes 93676c88 ("nbd: Don't send oversize strings", v4.2.0)Reported-by Dr. David Alan Gilbert <[email protected]>Signed-off-by Eric Blake <[email protected]>---
v2: update subject line and commit message to reflect file being touched; adjust a second nearby assertion with the same issue
nbd/client.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)nbd/client.c +0
-0@@ -658,11 +658,11 @@ static int nbd_send_meta_query(QIOChannel *ioc, uint32_t opt,char *p; data_len =sizeof (export_len) + export_len +sizeof (queries);nbd/client.c static int nbd_send_meta_query(QIOChannel *ioc, uint32_t opt, · line 668
Quoted 10 lines
nbd/client.c +2
-2-assert (export_len <=NBD_MAX_STRING_SIZE ); +assert (strlen (export) <=NBD_MAX_STRING_SIZE );if (query) { query_len =strlen (query); data_len +=sizeof (query_len) + query_len; -assert (query_len <=NBD_MAX_STRING_SIZE ); +assert (strlen (query) <=NBD_MAX_STRING_SIZE ); }else {assert (opt ==NBD_OPT_LIST_META_CONTEXT ); }I'm a bit late, and this should work as is.
Quoted 3 lines
Still, for me it's a bit strange: you point to the fact that we probably overflow
uint32_tvariable. But we keep this fact hidden in the code. So, everyone who read should guess "aha, this extra strlen here is because the previous one may overflow the variable)".Could we use
strnlen()instead ofstrlen()? That would be also more effective.
Good idea. As in:
assert(strnlen(query, NBD_MAX_STRING_SIZE + 1) <= NBD_MAX_STRING_SIZE);
Signature · 3 lines
Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3266 Virtualization: qemu.org | libvirt.org