x86/fault: Don't send SIGSEGV twice on SEGV_PKUERR

__bad_area_nosemaphore() calls both force_sig_pkuerr() and
force_sig_fault() when handling SEGV_PKUERR. This does not cause
problems because the second signal is filtered by the legacy_queue()
check in __send_signal() because in both cases, the signal is SIGSEGV,
the second one seeing that the first one is already pending.

This causes the kernel to do unnecessary work so send the signal only
once for SEGV_PKUERR.

 [ bp: Massage commit message. ]

Fixes: 9db812dbb29d ("signal/x86: Call force_sig_pkuerr from __bad_area_nosemaphore")
Suggested-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Jiashuo Liang <liangjs@pku.edu.cn>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Link: https://lkml.kernel.org/r/20210601085203.40214-1-liangjs@pku.edu.cn

diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 1c548ad0075204e3d129913be2767b64d1edb2fc..6bda7f67d737e9bb88102b295ff013d210a4a377 100644 (file)
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -836,8 +836,8 @@ __bad_area_nosemaphore(struct pt_regs *regs, unsigned long error_code,
 
        if (si_code == SEGV_PKUERR)
                force_sig_pkuerr((void __user *)address, pkey);
-
-       force_sig_fault(SIGSEGV, si_code, (void __user *)address);
+       else
+               force_sig_fault(SIGSEGV, si_code, (void __user *)address);
 
        local_irq_disable();
 }