[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20220302143331.654426-1-Jason@zx2c4.com>
Date: Wed, 2 Mar 2022 15:33:31 +0100
From: "Jason A. Donenfeld" <Jason@...c4.com>
To: linux-kernel@...r.kernel.org
Cc: "Jason A. Donenfeld" <Jason@...c4.com>,
Dominik Brodowski <linux@...inikbrodowski.net>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Theodore Ts'o <tytso@....edu>,
Alexander Graf <graf@...zon.com>,
Adrian Catangiu <adrian@...ity.io>,
Jann Horn <jannh@...gle.com>
Subject: [PATCH RFC v1] random: add vm_id sysctl with polling
This is an RFC, because I'm in no rush to add new userspace APIs without
being quite sure that this is the way forward.
In order to inform userspace of virtual machine forks, this commit adds
a "vm_id" sysctl, with the same semantics as the existing "boot_id" one,
except it changes after a VM fork and allows userspace processes to
poll() on it for changes.
It avoids exposing the actual vmgenid from the hypervisor to userspace,
in case there is any randomness value in keeping it secret.
The main purpose of posting this patch is just to have something
concrete to demonstrate the idea, as well as to show how simple this is
to implement. The discussion of races and such is for the thread in the
link below.
Link: https://lore.kernel.org/lkml/Yh4+9+UpanJWAIyZ@zx2c4.com/
Cc: Dominik Brodowski <linux@...inikbrodowski.net>
Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc: Theodore Ts'o <tytso@....edu>
Cc: Alexander Graf <graf@...zon.com>
Cc: Adrian Catangiu <adrian@...ity.io>
Cc: Jann Horn <jannh@...gle.com>
Signed-off-by: Jason A. Donenfeld <Jason@...c4.com>
---
drivers/char/random.c | 15 +++++++++++++++
1 file changed, 15 insertions(+)
diff --git a/drivers/char/random.c b/drivers/char/random.c
index 07ca3164522c..942a6dc4f00b 100644
--- a/drivers/char/random.c
+++ b/drivers/char/random.c
@@ -1142,6 +1142,8 @@ EXPORT_SYMBOL_GPL(add_bootloader_randomness);
#if IS_ENABLED(CONFIG_VMGENID)
static BLOCKING_NOTIFIER_HEAD(vmfork_chain);
+static DEFINE_CTL_TABLE_POLL(sysctl_vmid_poll);
+static u8 sysctl_vmid[UUID_SIZE];
/*
* Handle a new unique VM ID, which is unique, not secret, so we
@@ -1156,6 +1158,10 @@ void add_vmfork_randomness(const void *unique_vm_id, size_t size)
pr_notice("crng reseeded due to virtual machine fork\n");
}
blocking_notifier_call_chain(&vmfork_chain, 0, NULL);
+ if (IS_ENABLED(CONFIG_SYSCTL)) {
+ WRITE_ONCE(sysctl_vmid[8], 0);
+ proc_sys_poll_notify(&sysctl_vmid_poll);
+ }
}
#if IS_MODULE(CONFIG_VMGENID)
EXPORT_SYMBOL_GPL(add_vmfork_randomness);
@@ -1684,6 +1690,15 @@ static struct ctl_table random_table[] = {
.mode = 0444,
.proc_handler = proc_do_uuid,
},
+#if IS_ENABLED(CONFIG_VMGENID)
+ {
+ .procname = "vm_id",
+ .data = &sysctl_vmid,
+ .mode = 0444,
+ .proc_handler = proc_do_uuid,
+ .poll = &sysctl_vmid_poll,
+ },
+#endif
{
.procname = "uuid",
.mode = 0444,
--
2.35.1
Powered by blists - more mailing lists