linux-kernel - [PATCH UP] Make smp_call_function

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <20200205143409.GA7021@paulmck-ThinkPad-P72>
Date:   Wed, 5 Feb 2020 06:34:09 -0800
From:   "Paul E. McKenney" <paulmck@...nel.org>
To:     linux-kernel@...r.kernel.org
Cc:     kernel-team@...com, tglx@...utronix.de, gregkh@...uxfoundation.org,
        rmk+kernel@...linux.org.uk, nico@...xnic.net
Subject: [PATCH UP] Make smp_call_function_single() match SMP semantics

In CONFIG_SMP=y kernels, smp_call_function_single() returns -ENXIO when
invoked for a non-existent CPU.  In contrast, in CONFIG_SMP=n kernels,
a splat is emitted and smp_call_function_single() otherwise silently
ignores its "cpu" argument, instead pretending that the caller intended
to have something happen on CPU 0.  Given that there is now code that
expects smp_call_function_single() to return an error if a bad CPU was
specified, this difference in semantics needs to be addressed.

This commit therefore brings the semantics of the CONFIG_SMP=n version
of smp_call_function_single() into alignment with its CONFIG_SMP=y
counterpart.

Signed-off-by: Paul E. McKenney <paulmck@...nel.org>
---
 up.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/kernel/up.c b/kernel/up.c
index 862b460..a504e81 100644
--- a/kernel/up.c
+++ b/kernel/up.c
@@ -14,7 +14,8 @@ int smp_call_function_single(int cpu, void (*func) (void *info), void *info,
 {
 	unsigned long flags;

-	WARN_ON(cpu != 0);
+	if (cpu != 0)
+		return -ENXIO;

 	local_irq_save(flags);
 	func(info);