On 01/18/2019 09:46 AM, Greg KH wrote:
On Mon, Jan 14, 2019 at 10:02:21AM +0000, Suzuki K Poulose wrote:
On 09/01/2019 23:55, Jeremy Linton wrote:
As suggested on the list, https://lkml.org/lkml/2019/1/4/282, there are
a number of cases where its useful for a system to avoid exporting a
sysfs entry for a given vulnerability. This set adds an architecture
specific callback which returns the bitmap of vulnerabilities the
architecture would like to advertise.
Signed-off-by: Jeremy Linton <[email protected]>
Cc: Greg Kroah-Hartman <[email protected]>
Cc: Rafael J. Wysocki <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Josh Poimboeuf <[email protected]>
Cc: Konrad Rzeszutek Wilk <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Waiman Long <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Jiri Kosina <[email protected]>
---
drivers/base/cpu.c | 19 +++++++++++++++++++
include/linux/cpu.h | 7 +++++++
2 files changed, 26 insertions(+)
diff --git a/drivers/base/cpu.c b/drivers/base/cpu.c
index eb9443d5bae1..35f6dfb24cd6 100644
--- a/drivers/base/cpu.c
+++ b/drivers/base/cpu.c
@@ -561,6 +561,11 @@ static struct attribute *cpu_root_vulnerabilities_attrs[]
= {
NULL
};
+uint __weak arch_supported_vuln_attr_fields(void)
+{
+ return VULN_MELTDOWN|VULN_SPECTREV1|VULN_SPECTREV2|VULN_SSB|VULN_L1TF;
+}
+
static const struct attribute_group cpu_root_vulnerabilities_group = {
.name = "vulnerabilities",
.attrs = cpu_root_vulnerabilities_attrs,
@@ -568,6 +573,20 @@ static const struct attribute_group
cpu_root_vulnerabilities_group = {
static void __init cpu_register_vulnerabilities(void)
{
+ int fld;
+ int max_fields = ARRAY_SIZE(cpu_root_vulnerabilities_attrs) - 1;
+ struct attribute **hd = cpu_root_vulnerabilities_attrs;
+ uint enabled_fields = arch_supported_vuln_attr_fields();
+
+ /* only enable entries requested by the arch code */
+ for (fld = 0; fld < max_fields; fld++) {
+ if (enabled_fields & 1 << fld) {
+ *hd = cpu_root_vulnerabilities_attrs[fld];
+ hd++;
+ }
+ }
+ *hd = NULL;
+
nit: Could we use "is_visible" callback in the attribute group to check this
dynamically ?
You should, that is what it is there for.
Yes, its a good suggestion. OTOH, I think the plan is to drop this
functionality all together by removing the ability to build kernels
without the vulnerability checking/processor white lists. That will
simplify some of the #ifdef'ing going on as well.
