powerpc/perf/hv-gpci: Increase request buffer size
authorSukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Tue, 9 Feb 2016 07:47:45 +0000 (02:47 -0500)
committerMichael Ellerman <mpe@ellerman.id.au>
Wed, 10 Feb 2016 03:30:56 +0000 (14:30 +1100)
The GPCI hcall allows for a 4K buffer but we limit the buffer to 1K.
The problem with a 1K buffer is if a request results in returning
more values than can be accomodated in the 1K buffer the request will
fail.

The buffer we are using is currently allocated on the stack and hence
limited in size. Instead use a per-CPU 4K buffer like we do with 24x7
counters (hv-24x7.c).

While here, rename the macro GPCI_MAX_DATA_BYTES to HGPCI_MAX_DATA_BYTES
for consistency with 24x7 counters.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
arch/powerpc/perf/hv-gpci.c

index 856fe6e03c2a84a122456fd001f30834c4253940..7aa37236bb70bf38ae4d4bb2cc5e9e118d1ad5da 100644 (file)
@@ -127,8 +127,16 @@ static const struct attribute_group *attr_groups[] = {
        NULL,
 };
 
-#define GPCI_MAX_DATA_BYTES \
-       (1024 - sizeof(struct hv_get_perf_counter_info_params))
+#define HGPCI_REQ_BUFFER_SIZE  4096
+#define HGPCI_MAX_DATA_BYTES \
+       (HGPCI_REQ_BUFFER_SIZE - sizeof(struct hv_get_perf_counter_info_params))
+
+DEFINE_PER_CPU(char, hv_gpci_reqb[HGPCI_REQ_BUFFER_SIZE]) __aligned(sizeof(uint64_t));
+
+struct hv_gpci_request_buffer {
+       struct hv_get_perf_counter_info_params params;
+       uint8_t bytes[HGPCI_MAX_DATA_BYTES];
+} __packed;
 
 static unsigned long single_gpci_request(u32 req, u32 starting_index,
                u16 secondary_index, u8 version_in, u32 offset, u8 length,
@@ -137,24 +145,21 @@ static unsigned long single_gpci_request(u32 req, u32 starting_index,
        unsigned long ret;
        size_t i;
        u64 count;
+       struct hv_gpci_request_buffer *arg;
+
+       arg = (void *)get_cpu_var(hv_gpci_reqb);
+       memset(arg, 0, HGPCI_REQ_BUFFER_SIZE);
 
-       struct {
-               struct hv_get_perf_counter_info_params params;
-               uint8_t bytes[GPCI_MAX_DATA_BYTES];
-       } __packed __aligned(sizeof(uint64_t)) arg = {
-               .params = {
-                       .counter_request = cpu_to_be32(req),
-                       .starting_index = cpu_to_be32(starting_index),
-                       .secondary_index = cpu_to_be16(secondary_index),
-                       .counter_info_version_in = version_in,
-               }
-       };
+       arg->params.counter_request = cpu_to_be32(req);
+       arg->params.starting_index = cpu_to_be32(starting_index);
+       arg->params.secondary_index = cpu_to_be16(secondary_index);
+       arg->params.counter_info_version_in = version_in;
 
        ret = plpar_hcall_norets(H_GET_PERF_COUNTER_INFO,
-                       virt_to_phys(&arg), sizeof(arg));
+                       virt_to_phys(arg), HGPCI_REQ_BUFFER_SIZE);
        if (ret) {
                pr_devel("hcall failed: 0x%lx\n", ret);
-               return ret;
+               goto out;
        }
 
        /*
@@ -163,9 +168,11 @@ static unsigned long single_gpci_request(u32 req, u32 starting_index,
         */
        count = 0;
        for (i = offset; i < offset + length; i++)
-               count |= arg.bytes[i] << (i - offset);
+               count |= arg->bytes[i] << (i - offset);
 
        *value = count;
+out:
+       put_cpu_var(hv_gpci_reqb);
        return ret;
 }
 
@@ -245,10 +252,10 @@ static int h_gpci_event_init(struct perf_event *event)
        }
 
        /* last byte within the buffer? */
-       if ((event_get_offset(event) + length) > GPCI_MAX_DATA_BYTES) {
+       if ((event_get_offset(event) + length) > HGPCI_MAX_DATA_BYTES) {
                pr_devel("request outside of buffer: %zu > %zu\n",
                                (size_t)event_get_offset(event) + length,
-                               GPCI_MAX_DATA_BYTES);
+                               HGPCI_MAX_DATA_BYTES);
                return -EINVAL;
        }