x86/intel_rdt: Fix size reporting of MBA resource
authorReinette Chatre <reinette.chatre@intel.com>
Sat, 15 Sep 2018 21:58:20 +0000 (14:58 -0700)
committerThomas Gleixner <tglx@linutronix.de>
Tue, 18 Sep 2018 21:38:05 +0000 (23:38 +0200)
Chen Yu reported a divide-by-zero error when accessing the 'size'
resctrl file when a MBA resource is enabled.

divide error: 0000 [#1] SMP PTI
CPU: 93 PID: 1929 Comm: cat Not tainted 4.19.0-rc2-debug-rdt+ #25
RIP: 0010:rdtgroup_cbm_to_size+0x7e/0xa0
Call Trace:
rdtgroup_size_show+0x11a/0x1d0
seq_read+0xd8/0x3b0

Quoting Chen Yu's report: This is because for MB resource, the
r->cache.cbm_len is zero, thus calculating size in rdtgroup_cbm_to_size()
will trigger the exception.

Fix this issue in the 'size' file by getting correct memory bandwidth value
which is in MBps when MBA software controller is enabled or in percentage
when MBA software controller is disabled.

Fixes: d9b48c86eb38 ("x86/intel_rdt: Display resource groups' allocations in bytes")
Reported-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Fenghua Yu <fenghua.yu@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Chen Yu <yu.c.chen@intel.com>
Cc: "H Peter Anvin" <hpa@zytor.com>
Cc: "Tony Luck" <tony.luck@intel.com>
Cc: "Xiaochen Shen" <xiaochen.shen@intel.com>
Link: https://lkml.kernel.org/r/20180904174614.26682-1-yu.c.chen@intel.com
Link: https://lkml.kernel.org/r/1537048707-76280-3-git-send-email-fenghua.yu@intel.com
arch/x86/kernel/cpu/intel_rdt_rdtgroup.c

index b799c00..32e8bbd 100644 (file)
@@ -1155,8 +1155,8 @@ static int rdtgroup_size_show(struct kernfs_open_file *of,
        struct rdt_resource *r;
        struct rdt_domain *d;
        unsigned int size;
-       bool sep = false;
-       u32 cbm;
+       bool sep;
+       u32 ctrl;
 
        rdtgrp = rdtgroup_kn_lock_live(of->kn);
        if (!rdtgrp) {
@@ -1174,6 +1174,7 @@ static int rdtgroup_size_show(struct kernfs_open_file *of,
        }
 
        for_each_alloc_enabled_rdt_resource(r) {
+               sep = false;
                seq_printf(s, "%*s:", max_name_width, r->name);
                list_for_each_entry(d, &r->domains, list) {
                        if (sep)
@@ -1181,8 +1182,13 @@ static int rdtgroup_size_show(struct kernfs_open_file *of,
                        if (rdtgrp->mode == RDT_MODE_PSEUDO_LOCKSETUP) {
                                size = 0;
                        } else {
-                               cbm = d->ctrl_val[rdtgrp->closid];
-                               size = rdtgroup_cbm_to_size(r, d, cbm);
+                               ctrl = (!is_mba_sc(r) ?
+                                               d->ctrl_val[rdtgrp->closid] :
+                                               d->mbps_val[rdtgrp->closid]);
+                               if (r->rid == RDT_RESOURCE_MBA)
+                                       size = ctrl;
+                               else
+                                       size = rdtgroup_cbm_to_size(r, d, ctrl);
                        }
                        seq_printf(s, "%d=%u", d->id, size);
                        sep = true;