bootconfig: Fix off-by-one in xbc_node_compose_key_after()
Subsystems (2)
PATCH 1/1 bootconfig: Fix off-by-one in xbc_node_compose_key_after()
From: Steven Rostedt (VMware) <[email protected]>
While reviewing some patches for bootconfig, I noticed the following code in xbc_node_compose_key_after():
ret = snprintf(buf, size, "%s%s", xbc_node_get_data(node),
depth ? "." : "");
if (ret < 0)
return ret;
if (ret > size) {
size = 0;
} else {
size -= ret;
buf += ret;
}But snprintf() returns the number of bytes that would be written, not the number of bytes that are written (ignoring the nul terminator). This means that if the number of non null bytes written were to equal size, then the nul byte, which snprintf() always adds, will overwrite that last byte.
ret = snprintf(buf, 5, "hello");
printf("buf = '%s'\n", buf);
printf("ret = %d\n", ret);produces:
buf = 'hell' ret = 5
The string was truncated without ret being greater than 5. Test (ret >= size) for overwrite.
---
lib/bootconfig.c +1 -1
@@ -248,7 +248,7 @@ int __init xbc_node_compose_key_after(struct xbc_node *root, depth ? "." : "");if (ret <0 )return ret; -if (ret > size) { +if (ret >= size) { size =0 ; }else { size -= ret;
Andrew Morton
maintainerQuoted 26 lines · On Thu, 13 Aug 2020 18:30:50 -0400 Steven Rostedt <[email protected]> wrote:
From: Steven Rostedt (VMware) <[email protected]>
While reviewing some patches for bootconfig, I noticed the following code in
xbc_node_compose_key_after():ret = snprintf(buf, size, "%s%s", xbc_node_get_data(node), depth ? "." : ""); if (ret < 0) return ret; if (ret > size) { size = 0; } else { size -= ret; buf += ret; }But
snprintf()returns the number of bytes that would be written, not the number of bytes that are written (ignoring the nul terminator). This means that if the number of non null bytes written were to equal size, then the nul byte, whichsnprintf()always adds, will overwrite that last byte.ret = snprintf(buf, 5, "hello"); printf("buf = '%s'\n", buf); printf("ret = %d\n", ret);
Quoted 7 lines
produces:
buf = 'hell' ret = 5The string was truncated without ret being greater than 5. Test (ret >= size) for overwrite.
What are the end-user visible effects of the bug? IOW, why cc:stable?
Quoted 30 lines · On Thu, 13 Aug 2020 19:38:18 -0700 Andrew Morton <[email protected]>…
Quoted 28 lines · On Thu, 13 Aug 2020 18:30:50 -0400 Steven Rostedt <[email protected]> wrote:
From: Steven Rostedt (VMware) <[email protected]>
While reviewing some patches for bootconfig, I noticed the following code in
xbc_node_compose_key_after():ret = snprintf(buf, size, "%s%s", xbc_node_get_data(node), depth ? "." : ""); if (ret < 0) return ret; if (ret > size) { size = 0; } else { size -= ret; buf += ret; }But
snprintf()returns the number of bytes that would be written, not the number of bytes that are written (ignoring the nul terminator). This means that if the number of non null bytes written were to equal size, then the nul byte, whichsnprintf()always adds, will overwrite that last byte.ret = snprintf(buf, 5, "hello"); printf("buf = '%s'\n", buf); printf("ret = %d\n", ret);produces:
Quoted 8 lines
Quoted 5 lines
buf = 'hell' ret = 5The string was truncated without ret being greater than 5. Test (ret >= size) for overwrite.
What are the end-user visible effects of the bug? IOW, why cc:stable?
Hmm, looking at it at a wider view, it may not be an issue. The tools code calls this code, and I looked to see if it was possible to corrupt the buffer by an incorrect size. But now that I'm looking at the else part of the section, it may not be a problem as it may act the same.
That is, ret == size will make size = 0 with the size -= ret, and we get the same result.
OK, you can drop the patch. Thanks for the review!
Although, there's no error message if the buffer is not big enough to hold the fields.
Masami?
-- Steve
Masami Hiramatsu
maintainerQuoted 47 lines · On Thu, 13 Aug 2020 23:04:06 -0400 Steven Rostedt <[email protected]> wrote:
Quoted 30 lines · On Thu, 13 Aug 2020 19:38:18 -0700 Andrew Morton <[email protected]>…
Quoted 28 lines · On Thu, 13 Aug 2020 18:30:50 -0400 Steven Rostedt <[email protected]> wrote:
From: Steven Rostedt (VMware) <[email protected]>
While reviewing some patches for bootconfig, I noticed the following code in
xbc_node_compose_key_after():ret = snprintf(buf, size, "%s%s", xbc_node_get_data(node), depth ? "." : ""); if (ret < 0) return ret; if (ret > size) { size = 0; } else { size -= ret; buf += ret; }But
snprintf()returns the number of bytes that would be written, not the number of bytes that are written (ignoring the nul terminator). This means that if the number of non null bytes written were to equal size, then the nul byte, whichsnprintf()always adds, will overwrite that last byte.ret = snprintf(buf, 5, "hello"); printf("buf = '%s'\n", buf); printf("ret = %d\n", ret);produces:
Quoted 8 lines
Quoted 5 lines
buf = 'hell' ret = 5The string was truncated without ret being greater than 5. Test (ret >= size) for overwrite.
What are the end-user visible effects of the bug? IOW, why cc:stable?
Hmm, looking at it at a wider view, it may not be an issue. The tools code calls this code, and I looked to see if it was possible to corrupt the buffer by an incorrect size. But now that I'm looking at the else part of the section, it may not be a problem as it may act the same.
Quoted 4 lines
That is, ret == size will make size = 0 with the size -= ret, and we get the same result.
OK, you can drop the patch. Thanks for the review!
Thanks Andrew, you're right. If size == ret, then size -= ret makes
size = 0 and after that, the loop doesn't change the size.
Quoted 5 lines
Although, there's no error message if the buffer is not big enough to hold the fields.
Masami?
Ah, I think we need a patch to fix a comment of the function. It says
* Returns the total length of the key stored in @buf.
But this must be
* Returns the total bytes of the key which would be written.
As similar to the snprintf(), so that the user can compare the result with the given buffer size.
Thank you,
Signature · 1 line
Masami Hiramatsu <[email protected]>