Merge tag 'i2c-host-6.13-p1' of git://git.kernel.org/pub/scm/linux/kernel/git/andi.shyti/linux into i2c/for-mergewindow

i2c-host updates for v6.13, part 1

Major Improvements and Refactoring:

 - All controllers using the 'remove_new' callback have been
   reverted to use the 'remove' callback.

 - Intel SCH controller underwent significant refactoring,
   this brings love and a modern look to the driver.

 - PIIX4 driver refactored to enable usage by other drivers
   (e.g., AMD ASF).

 - iMX/MXC improved message handling to reduce protocol overhead:
     Refactored DMA/non-DMA read/write and bus polling mechanisms
     to achieve this.

 - ACPI documentation for PIIX4.

New Features:

 - i2c-cadence added support for atomic transfers.
 - Qualcomm CII added support for a 32MHz serial engine clock.

Deprecated Features:

 - Dropped outdated support for AMD756 S4882 and NFORCE2 S4985. If
   somebody misses this, Jean will rewrite support using the proper
   i2c mux framework.

New Hardware Support:

 - Added support for:
   - Intel Panther Lake (new ID)
   - AMD ASF (new driver)
   - S32G2/S32G3 SoCs (new ID)
   - Realtek RTL I2C Controller (new driver)
   - HJMC01 DesignWare ACPI HID (new ID)
   - PIC64GX to Microchip Core (new ID)
   - Qualcomm SDM670 to Qualcomm CCI (new ID)

1 files changed, 32 insertions, 0 deletions

diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c
index d1915427b4ffc..47a01d4028f60 100644
--- a/arch/x86/kernel/cpu/bugs.c
+++ b/arch/x86/kernel/cpu/bugs.c
@@ -1115,8 +1115,25 @@ do_cmd_auto:
 
 	case RETBLEED_MITIGATION_IBPB:
 		setup_force_cpu_cap(X86_FEATURE_ENTRY_IBPB);
+
+		/*
+		 * IBPB on entry already obviates the need for
+		 * software-based untraining so clear those in case some
+		 * other mitigation like SRSO has selected them.
+		 */
+		setup_clear_cpu_cap(X86_FEATURE_UNRET);
+		setup_clear_cpu_cap(X86_FEATURE_RETHUNK);
+
 		setup_force_cpu_cap(X86_FEATURE_IBPB_ON_VMEXIT);
 		mitigate_smt = true;
+
+		/*
+		 * There is no need for RSB filling: entry_ibpb() ensures
+		 * all predictions, including the RSB, are invalidated,
+		 * regardless of IBPB implementation.
+		 */
+		setup_clear_cpu_cap(X86_FEATURE_RSB_VMEXIT);
+
 		break;
 
 	case RETBLEED_MITIGATION_STUFF:
@@ -2627,6 +2644,14 @@ static void __init srso_select_mitigation(void)
 			if (has_microcode) {
 				setup_force_cpu_cap(X86_FEATURE_ENTRY_IBPB);
 				srso_mitigation = SRSO_MITIGATION_IBPB;
+
+				/*
+				 * IBPB on entry already obviates the need for
+				 * software-based untraining so clear those in case some
+				 * other mitigation like Retbleed has selected them.
+				 */
+				setup_clear_cpu_cap(X86_FEATURE_UNRET);
+				setup_clear_cpu_cap(X86_FEATURE_RETHUNK);
 			}
 		} else {
 			pr_err("WARNING: kernel not compiled with MITIGATION_IBPB_ENTRY.\n");
@@ -2638,6 +2663,13 @@ static void __init srso_select_mitigation(void)
 			if (!boot_cpu_has(X86_FEATURE_ENTRY_IBPB) && has_microcode) {
 				setup_force_cpu_cap(X86_FEATURE_IBPB_ON_VMEXIT);
 				srso_mitigation = SRSO_MITIGATION_IBPB_ON_VMEXIT;
+
+				/*
+				 * There is no need for RSB filling: entry_ibpb() ensures
+				 * all predictions, including the RSB, are invalidated,
+				 * regardless of IBPB implementation.
+				 */
+				setup_clear_cpu_cap(X86_FEATURE_RSB_VMEXIT);
 			}
 		} else {
 			pr_err("WARNING: kernel not compiled with MITIGATION_SRSO.\n");