powerpc/perf/hv-24x7: use kmem_cache instead of aligned stack allocations
authorCody P Schafer <dev@codyps.com>
Wed, 1 Oct 2014 06:03:17 +0000 (23:03 -0700)
committerMichael Ellerman <mpe@ellerman.id.au>
Tue, 7 Oct 2014 05:52:58 +0000 (16:52 +1100)
Ian pointed out the use of __aligned(4096) caused rather large stack
consumption in single_24x7_request(), so use the kmem_cache
hv_page_cache (which we've already got set up for other allocations)
insead of allocating locally.

CC: Haren Myneni <hbabu@us.ibm.com>
Reported-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
arch/powerpc/perf/hv-24x7.c

index 70d4f748b54bead98a8889ea3de6d344dbb3ca31..697759d8eb1677ebf0a3c621dd82379a42f1e53a 100644 (file)
@@ -294,7 +294,7 @@ static unsigned long single_24x7_request(u8 domain, u32 offset, u16 ix,
                                         u16 lpar, u64 *res,
                                         bool success_expected)
 {
-       unsigned long ret;
+       unsigned long ret = -ENOMEM;
 
        /*
         * request_buffer and result_buffer are not required to be 4k aligned,
@@ -304,7 +304,27 @@ static unsigned long single_24x7_request(u8 domain, u32 offset, u16 ix,
        struct reqb {
                struct hv_24x7_request_buffer buf;
                struct hv_24x7_request req;
-       } __packed __aligned(4096) request_buffer = {
+       } __packed *request_buffer;
+
+       struct {
+               struct hv_24x7_data_result_buffer buf;
+               struct hv_24x7_result res;
+               struct hv_24x7_result_element elem;
+               __be64 result;
+       } __packed *result_buffer;
+
+       BUILD_BUG_ON(sizeof(*request_buffer) > 4096);
+       BUILD_BUG_ON(sizeof(*result_buffer) > 4096);
+
+       request_buffer = kmem_cache_zalloc(hv_page_cache, GFP_USER);
+       if (!request_buffer)
+               goto out;
+
+       result_buffer = kmem_cache_zalloc(hv_page_cache, GFP_USER);
+       if (!result_buffer)
+               goto out_free_request_buffer;
+
+       *request_buffer = (struct reqb) {
                .buf = {
                        .interface_version = HV_24X7_IF_VERSION_CURRENT,
                        .num_requests = 1,
@@ -320,28 +340,27 @@ static unsigned long single_24x7_request(u8 domain, u32 offset, u16 ix,
                }
        };
 
-       struct resb {
-               struct hv_24x7_data_result_buffer buf;
-               struct hv_24x7_result res;
-               struct hv_24x7_result_element elem;
-               __be64 result;
-       } __packed __aligned(4096) result_buffer = {};
-
        ret = plpar_hcall_norets(H_GET_24X7_DATA,
-                       virt_to_phys(&request_buffer), sizeof(request_buffer),
-                       virt_to_phys(&result_buffer),  sizeof(result_buffer));
+                       virt_to_phys(request_buffer), sizeof(*request_buffer),
+                       virt_to_phys(result_buffer),  sizeof(*result_buffer));
 
        if (ret) {
                if (success_expected)
-                       pr_err_ratelimited("hcall failed: %d %#x %#x %d => 0x%lx (%ld) detail=0x%x failing ix=%x\n",
-                                       domain, offset, ix, lpar,
-                                       ret, ret,
-                                       result_buffer.buf.detailed_rc,
-                                       result_buffer.buf.failing_request_ix);
-               return ret;
+                       pr_err_ratelimited("hcall failed: %d %#x %#x %d => "
+                               "0x%lx (%ld) detail=0x%x failing ix=%x\n",
+                               domain, offset, ix, lpar, ret, ret,
+                               result_buffer->buf.detailed_rc,
+                               result_buffer->buf.failing_request_ix);
+               goto out_free_result_buffer;
        }
 
-       *res = be64_to_cpu(result_buffer.result);
+       *res = be64_to_cpu(result_buffer->result);
+
+out_free_result_buffer:
+       kfree(result_buffer);
+out_free_request_buffer:
+       kfree(request_buffer);
+out:
        return ret;
 }