On Mon, 2012-04-16 at 17:38 -0300, Mauro Carvalho Chehab wrote:
Kernel kobjects have rigid rules: each container object should be
dynamically allocated, and can't be allocated into a single kmalloc.
EDAC never obeyed this rule: it has a single malloc function that
allocates all needed data into a single kzalloc.
As this is not accepted anymore, change the allocation schema of the
EDAC *_info structs to enforce this Kernel standard.
[]
quoted hunk ↗ jump to hunk
diff --git a/drivers/edac/amd64_edac.c b/drivers/edac/amd64_edac.c
[]
quoted hunk ↗ jump to hunk
@@ -2228,9 +2228,9 @@ static int init_csrows(struct mem_ctl_info *mci)
edac_mode = EDAC_NONE;
for (j = 0; j < pvt->channel_count; j++) {
- csrow->channels[j].dimm->mtype = mtype;
- csrow->channels[j].dimm->edac_mode = edac_mode;
- csrow->channels[j].dimm->nr_pages = nr_pages;
+ csrow->channels[j]->dimm->mtype = mtype;
+ csrow->channels[j]->dimm->edac_mode = edac_mode;
+ csrow->channels[j]->dimm->nr_pages = nr_pages;
It might be better to use an automatic for
typeof foo dimm = csrow->channels[j]->dimm;
dimm->mtype = mtype;
etc...
[]
quoted hunk ↗ jump to hunk
@@ -293,39 +286,56 @@ struct mem_ctl_info *edac_mc_alloc(unsigned edac_index,
mci->mem_is_per_rank = per_rank;
/*
- * Fills the csrow struct
+ * Alocate and fill the csrow/channels structs
*/
+ mci->csrows = kzalloc(sizeof(*mci->csrows) * tot_csrows, GFP_KERNEL);
kcalloc
[]
+ csr->channels = kzalloc(sizeof(*csr->channels) * tot_cschannels,
+ GFP_KERNEL);
here too
quoted hunk ↗ jump to hunk
@@ -391,7 +401,31 @@ struct mem_ctl_info *edac_mc_alloc(unsigned edac_index,
trace_hw_event_init("edac", (unsigned)edac_index);
+ debugf1("EDAC MCI allocated\n");
return mci;
+
+error:
+ debugf1("Failed to allocate one or more EDAC MCI structs\n");
Generally, it's not necessary to have specific "OOM" messages
as allocations without GFP_NOWARN do dump_stack()s