powernow: add fixups for AMD P-state figures

In the Linux kernel, these two git commits:

- f594065faf4f9067c2283a34619fc0714e79a98d
  ACPI: Add fixups for AMD P-state figures
- 9855d8ce41a7801548a05d844db2f46c3e810166
  ACPI: Check MSR valid bit before using P-state frequencies

Try to fix the the issue that "some AMD systems may round the
frequencies in ACPI tables to 100MHz boundaries. We can obtain the real
frequencies from MSRs, so add a quirk to fix these frequencies up
on AMD systems." (from f594065..)

In discussion (around 9855d8..) "it turned out that indeed real
HW/BIOSes may choose to not set the valid bit and thus mark the
P-state as invalid. So this could be considered a fix for broken
BIOSes." (from 9855d8..)

which is great for Linux. Unfortunatly the Linux kernel, when
it tries to do the RDMSR under Xen it fails to get the right
value (it gets zero) as Xen traps it and returns zero. Hence
when dom0 uploads the P-states they will be unmodified and
we should take care of updating the frequencies with the right
values.

I've tested it under Dell Inc. PowerEdge T105 /0RR825, BIOS 1.3.2
08/20/2008 where this quirk can be observed (x86 == 0x10, model == 2).
Also on other AMD (x86 == 0x12, A8-3850; x86 = 0x14, AMD E-350) to
make sure the quirk is not applied there.

Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Acked-by: stefan.bader@canonical.com

Do the MSR access here (and while at it, also the one reading
MSR_PSTATE_CUR_LIMIT) on the target CPU, and bound the loop over
amd_fixup_frequency() by max_hw_pstate (matching the one in
powernow_cpufreq_cpu_init()).

Signed-off-by: Jan Beulich <jbeulich@suse.com>
master changeset: 1d80765b504b34b63a42a63aff4291e07e29f0c5
master date: 2013-03-12 15:34:22 +0100

1 files changed, 50 insertions, 6 deletions

diff --git a/xen/arch/x86/acpi/cpufreq/powernow.c b/xen/arch/x86/acpi/cpufreq/powernow.c
index db1806f8f0..398c1c6618 100644
--- a/xen/arch/x86/acpi/cpufreq/powernow.c
+++ b/xen/arch/x86/acpi/cpufreq/powernow.c
@@ -146,6 +146,51 @@ static int powernow_cpufreq_target(struct cpufreq_policy *policy,
     return result;
 }
 
+static void amd_fixup_frequency(struct xen_processor_px *px)
+{
+    u32 hi, lo, fid, did;
+    int index = px->control & 0x00000007;
+    const struct cpuinfo_x86 *c = &current_cpu_data;
+
+    if ((c->x86 != 0x10 || c->x86_model >= 10) && c->x86 != 0x11)
+        return;
+
+    rdmsr(MSR_PSTATE_DEF_BASE + index, lo, hi);
+    /*
+     * MSR C001_0064+:
+     * Bit 63: PstateEn. Read-write. If set, the P-state is valid.
+     */
+    if (!(hi & (1U << 31)))
+        return;
+
+    fid = lo & 0x3f;
+    did = (lo >> 6) & 7;
+    if (c->x86 == 0x10)
+        px->core_frequency = (100 * (fid + 16)) >> did;
+    else
+        px->core_frequency = (100 * (fid + 8)) >> did;
+}
+
+struct amd_cpu_data {
+    struct processor_performance *perf;
+    u32 max_hw_pstate;
+};
+
+static void get_cpu_data(void *arg)
+{
+    struct amd_cpu_data *data = arg;
+    struct processor_performance *perf = data->perf;
+    uint64_t msr_content;
+    unsigned int i;
+
+    rdmsrl(MSR_PSTATE_CUR_LIMIT, msr_content);
+    data->max_hw_pstate = (msr_content & HW_PSTATE_MAX_MASK) >>
+                          HW_PSTATE_MAX_SHIFT;
+
+    for (i = 0; i < perf->state_count && i <= data->max_hw_pstate; i++)
+        amd_fixup_frequency(&perf->states[i]);
+}
+
 static int powernow_cpufreq_verify(struct cpufreq_policy *policy)
 {
     struct powernow_cpufreq_data *data;
@@ -192,8 +237,7 @@ static int powernow_cpufreq_cpu_init(struct cpufreq_policy *policy)
     struct powernow_cpufreq_data *data;
     unsigned int result = 0;
     struct processor_performance *perf;
-    u32 max_hw_pstate;
-    uint64_t msr_content;
+    struct amd_cpu_data info;
     struct cpuinfo_x86 *c = &cpu_data[policy->cpu];
 
     data = xmalloc(struct powernow_cpufreq_data);
@@ -205,7 +249,7 @@ static int powernow_cpufreq_cpu_init(struct cpufreq_policy *policy)
 
     data->acpi_data = &processor_pminfo[cpu]->perf;
 
-    perf = data->acpi_data;
+    info.perf = perf = data->acpi_data;
     policy->shared_type = perf->shared_type;
 
     /*
@@ -225,8 +269,6 @@ static int powernow_cpufreq_cpu_init(struct cpufreq_policy *policy)
         result = -ENODEV;
         goto err_unreg;
     }
-    rdmsrl(MSR_PSTATE_CUR_LIMIT, msr_content);
-    max_hw_pstate = (msr_content & HW_PSTATE_MAX_MASK) >> HW_PSTATE_MAX_SHIFT;
 
     if (perf->control_register.space_id != perf->status_register.space_id) {
         result = -ENODEV;
@@ -251,9 +293,11 @@ static int powernow_cpufreq_cpu_init(struct cpufreq_policy *policy)
 
     policy->governor = cpufreq_opt_governor ? : CPUFREQ_DEFAULT_GOVERNOR;
 
+    on_selected_cpus(cpumask_of(cpu), get_cpu_data, &info, 1);
+
     data->max_freq = perf->states[0].core_frequency * 1000;
     /* table init */
-    for (i = 0; i < perf->state_count && i <= max_hw_pstate; i++) {
+    for (i = 0; i < perf->state_count && i <= info.max_hw_pstate; i++) {
         if (i > 0 && perf->states[i].core_frequency >=
             data->freq_table[valid_states-1].frequency / 1000)
             continue;