perf/x86/intel/pt: Clean up the control flow in pt_pmu_hw_init()
authorIngo Molnar <mingo@kernel.org>
Sun, 12 Apr 2015 09:11:21 +0000 (11:11 +0200)
committerIngo Molnar <mingo@kernel.org>
Sun, 12 Apr 2015 09:21:15 +0000 (11:21 +0200)
Dan Carpenter pointed out that the control flow in pt_pmu_hw_init()
is a bit messy: for example the kfree(de_attrs) is entirely
superfluous.

Another problem is the inconsistent mixing of label based and
direct return error handling.

Add modern, label based error handling instead and clean up the code
a bit as well.

Note that we'll still do a kfree(NULL) in the normal case - this does
not matter as this is an init path and kfree() returns early if it
sees a NULL.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20150409090805.GG17605@mwanda
Signed-off-by: Ingo Molnar <mingo@kernel.org>
arch/x86/kernel/cpu/perf_event_intel_pt.c

index f5a3afc653711264b68154a2d7b6314113844b85..f2770641c0fd212c5e585532d51fc1859ab94540 100644 (file)
@@ -119,48 +119,55 @@ static int __init pt_pmu_hw_init(void)
        struct dev_ext_attribute *de_attrs;
        struct attribute **attrs;
        size_t size;
+       int ret;
        long i;
 
-       if (test_cpu_cap(&boot_cpu_data, X86_FEATURE_INTEL_PT)) {
-               for (i = 0; i < PT_CPUID_LEAVES; i++)
-                       cpuid_count(20, i,
-                                   &pt_pmu.caps[CR_EAX + i * 4],
-                                   &pt_pmu.caps[CR_EBX + i * 4],
-                                   &pt_pmu.caps[CR_ECX + i * 4],
-                                   &pt_pmu.caps[CR_EDX + i * 4]);
-       } else {
-               return -ENODEV;
+       attrs = NULL;
+       ret = -ENODEV;
+       if (!test_cpu_cap(&boot_cpu_data, X86_FEATURE_INTEL_PT))
+               goto fail;
+
+       for (i = 0; i < PT_CPUID_LEAVES; i++) {
+               cpuid_count(20, i,
+                           &pt_pmu.caps[CR_EAX + i*4],
+                           &pt_pmu.caps[CR_EBX + i*4],
+                           &pt_pmu.caps[CR_ECX + i*4],
+                           &pt_pmu.caps[CR_EDX + i*4]);
        }
 
-       size = sizeof(struct attribute *) * (ARRAY_SIZE(pt_caps) + 1);
+       ret = -ENOMEM;
+       size = sizeof(struct attribute *) * (ARRAY_SIZE(pt_caps)+1);
        attrs = kzalloc(size, GFP_KERNEL);
        if (!attrs)
-               goto err_attrs;
+               goto fail;
 
-       size = sizeof(struct dev_ext_attribute) * (ARRAY_SIZE(pt_caps) + 1);
+       size = sizeof(struct dev_ext_attribute) * (ARRAY_SIZE(pt_caps)+1);
        de_attrs = kzalloc(size, GFP_KERNEL);
        if (!de_attrs)
-               goto err_de_attrs;
+               goto fail;
 
        for (i = 0; i < ARRAY_SIZE(pt_caps); i++) {
-               de_attrs[i].attr.attr.name = pt_caps[i].name;
+               struct dev_ext_attribute *de_attr = de_attrs + i;
+
+               de_attr->attr.attr.name = pt_caps[i].name;
+
+               sysfs_attr_init(&de_attrs->attr.attr);
 
-               sysfs_attr_init(&de_attrs[i].attr.attr);
-               de_attrs[i].attr.attr.mode = S_IRUGO;
-               de_attrs[i].attr.show = pt_cap_show;
-               de_attrs[i].var = (void *)i;
-               attrs[i] = &de_attrs[i].attr.attr;
+               de_attr->attr.attr.mode         = S_IRUGO;
+               de_attr->attr.show              = pt_cap_show;
+               de_attr->var                    = (void *)i;
+
+               attrs[i] = &de_attr->attr.attr;
        }
 
        pt_cap_group.attrs = attrs;
+
        return 0;
 
-err_de_attrs:
-       kfree(de_attrs);
-err_attrs:
+fail:
        kfree(attrs);
 
-       return -ENOMEM;
+       return ret;
 }
 
 #define PT_CONFIG_MASK (RTIT_CTL_TSC_EN | RTIT_CTL_DISRETC)