Skip to content

Commit fdf82a7

Browse files
Jiri KosinaKAGA-KOKO
Jiri Kosina
authored andcommitted
x86/speculation: Protect against userspace-userspace spectreRSB
The article "Spectre Returns! Speculation Attacks using the Return Stack Buffer" [1] describes two new (sub-)variants of spectrev2-like attacks, making use solely of the RSB contents even on CPUs that don't fallback to BTB on RSB underflow (Skylake+). Mitigate userspace-userspace attacks by always unconditionally filling RSB on context switch when the generic spectrev2 mitigation has been enabled. [1] https://arxiv.org/pdf/1807.07940.pdf Signed-off-by: Jiri Kosina <jkosina@suse.cz> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Josh Poimboeuf <jpoimboe@redhat.com> Acked-by: Tim Chen <tim.c.chen@linux.intel.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Borislav Petkov <bp@suse.de> Cc: David Woodhouse <dwmw@amazon.co.uk> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/nycvar.YFH.7.76.1807261308190.997@cbobk.fhfr.pm
1 parent ca38dc8 commit fdf82a7

File tree

1 file changed

+7
-31
lines changed

1 file changed

+7
-31
lines changed

Diff for: arch/x86/kernel/cpu/bugs.c

+7-31
Original file line numberDiff line numberDiff line change
@@ -313,23 +313,6 @@ static enum spectre_v2_mitigation_cmd __init spectre_v2_parse_cmdline(void)
313313
return cmd;
314314
}
315315

316-
/* Check for Skylake-like CPUs (for RSB handling) */
317-
static bool __init is_skylake_era(void)
318-
{
319-
if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL &&
320-
boot_cpu_data.x86 == 6) {
321-
switch (boot_cpu_data.x86_model) {
322-
case INTEL_FAM6_SKYLAKE_MOBILE:
323-
case INTEL_FAM6_SKYLAKE_DESKTOP:
324-
case INTEL_FAM6_SKYLAKE_X:
325-
case INTEL_FAM6_KABYLAKE_MOBILE:
326-
case INTEL_FAM6_KABYLAKE_DESKTOP:
327-
return true;
328-
}
329-
}
330-
return false;
331-
}
332-
333316
static void __init spectre_v2_select_mitigation(void)
334317
{
335318
enum spectre_v2_mitigation_cmd cmd = spectre_v2_parse_cmdline();
@@ -390,22 +373,15 @@ static void __init spectre_v2_select_mitigation(void)
390373
pr_info("%s\n", spectre_v2_strings[mode]);
391374

392375
/*
393-
* If neither SMEP nor PTI are available, there is a risk of
394-
* hitting userspace addresses in the RSB after a context switch
395-
* from a shallow call stack to a deeper one. To prevent this fill
396-
* the entire RSB, even when using IBRS.
376+
* If spectre v2 protection has been enabled, unconditionally fill
377+
* RSB during a context switch; this protects against two independent
378+
* issues:
397379
*
398-
* Skylake era CPUs have a separate issue with *underflow* of the
399-
* RSB, when they will predict 'ret' targets from the generic BTB.
400-
* The proper mitigation for this is IBRS. If IBRS is not supported
401-
* or deactivated in favour of retpolines the RSB fill on context
402-
* switch is required.
380+
* - RSB underflow (and switch to BTB) on Skylake+
381+
* - SpectreRSB variant of spectre v2 on X86_BUG_SPECTRE_V2 CPUs
403382
*/
404-
if ((!boot_cpu_has(X86_FEATURE_PTI) &&
405-
!boot_cpu_has(X86_FEATURE_SMEP)) || is_skylake_era()) {
406-
setup_force_cpu_cap(X86_FEATURE_RSB_CTXSW);
407-
pr_info("Spectre v2 mitigation: Filling RSB on context switch\n");
408-
}
383+
setup_force_cpu_cap(X86_FEATURE_RSB_CTXSW);
384+
pr_info("Spectre v2 / SpectreRSB mitigation: Filling RSB on context switch\n");
409385

410386
/* Initialize Indirect Branch Prediction Barrier if supported */
411387
if (boot_cpu_has(X86_FEATURE_IBPB)) {

0 commit comments

Comments
 (0)