On Fri, 2011-01-14 at 18:35 +0100, Moritz Mühlenhoff wrote: > What shall we do with CVE-2010-4075, CVE-2010-4076, CVE-2010-4077 > at this point of the freeze? > > Should be fixed by d281da7ff6f70efca0553c288bb883e8605b3862 > and 0587102cf9f427c185bfdeb2cef41e13ee0264b1 , but would change > the ABI. > > We could postpone it to a later point update, where we change the > ABI along with more serious issues requiring an ABI bump? I think I can see how to do this without an ABI bump: - Add the function pointers at the end of the structures (#ifndef __GENKSYMS__). - Define a new flag in tty_driver::flags indicating whether ops->get_icount is valid. - Define a new 1-bit bitfield in struct usb_serial (#ifndef __GENKSYMS__) indicating whether type->get_icount is valid. I'll try to implement this now. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse.
Attachment:
signature.asc
Description: This is a digitally signed message part