lkrg-users - pCFI false positive

Follow @Openwall on Twitter for new release announcements and other news

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <9959223f-494e-48be-95db-d9a3a02e048c.lc85446@alibaba-inc.com>
Date: Thu, 14 Apr 2022 18:08:53 +0800
From: "李诚(允诚)" <lc85446@...baba-inc.com>
To: "lkrg-users" <lkrg-users@...ts.openwall.com>
Subject: pCFI  false positive

Hi，all
     I tested lkrg-0.9.2 on an ARM64 board, and met a pCFI false positive issue about 
"Stack pointer corruption (ROP?)" for  the security_capable() function.

   I found that sometimes the security_capable() may be called in the interrupt context , 
then the p_ed_enforce_pcfi() would warn about the stack pointer mismatch and kill the 
victim process.

  Maybe the p_ed_enforce_pcfi() should use (unsigned long)p_task->stack if it finds that
p_regs_get_sp(p_regs) belongs to the irq_stack ?

thanks and best regards,

Li Cheng

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.