Thread (22 messages) 22 messages, 5 authors, 2021-11-10

Re: [PATCH] tty: vt: keyboard: do not copy an extra-byte in copy_to_user

From: Jiri Slaby <jirislaby@kernel.org>
Date: 2021-11-10 05:22:27
Also in: lkml

On 06. 11. 21, 21:48, Pavel Skripkin wrote:
On 11/6/21 23:44, Ajay Garg wrote:
quoted
quoted
quoted
That's the whole point of the discussion :)

The method "vt_do_kdgkb_ioctl" does not manage "func_table[kb_func]".
Thus, the method does not know whether or not
strlen(func_table[kb_func]) < sizeof(user_kdgkb->kb_string).
It manages. The code under `case KDSKBSENT:` sets func_table[] entries
via vt_kdskbsent().

kbs = strndup_user(..., sizeof(user_kdgkb->kb_string));

is used to allocate buffer for the func_table[] entry. That's my main
point :)
func_table is set in vt_kdskbent, which itself is external.

More importantly, vt_kdskbent is handled in case KDSKBSENT:, while the
strlcpy issue we are dealing with is in case KDGKBSENT:
In case KDGKBSENT, following are managed :

                 ssize_t len = sizeof(user_kdgkb->kb_string);
                 kbs = kmalloc(len, GFP_KERNEL);

while func_table[kb_func] is external entity here, so no assumption
ought to be made for it, just my 2 cents though :)

Anyhow, really, it is the maintainers' choice now :), since there
isn't a burning (compilation/runtime) issue.
I fully agree here, it's maintainer's choice. Let's sit down and wait 
what experienced people thing about this :)
I don't quite understand what the problem is. Provided I wrote the code, 
is there something wrong with this commit (and its explanation), in 
particular?

commit 6ca03f90527e499dd5e32d6522909e2ad390896b
Author: Jiri Slaby [off-list ref]
Date:   Mon Oct 19 10:55:16 2020 +0200

     vt: keyboard, simplify vt_kdgkbsent

     Use 'strlen' of the string, add one for NUL terminator and simply do
     'copy_to_user' instead of the explicit 'for' loop. This makes the
     KDGKBSENT case more compact.

     The only thing we need to take care about is NULL 'func_table[i]'. Use
     an empty string in that case.

     The original check for overflow could never trigger as the func_buf
     strings are always shorter or equal to 'struct kbsentry's.

thanks,
-- 
js
suse labs
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help