On Mon, Jun 8, 2015 at 9:29 AM, Elliott, Robert (Server Storage) <Elliott@xxxxxx> wrote: >> -----Original Message----- >> From: Linux-nvdimm [mailto:linux-nvdimm-bounces@xxxxxxxxxxxx] On Behalf >> Of Dan Williams >> Sent: Friday, June 05, 2015 3:19 PM >> Subject: [PATCH v4 4/9] dax: fix mapping lifetime handling, convert to >> __pfn_t + kmap_atomic_pfn_t() > ... >> diff --git a/arch/powerpc/sysdev/axonram.c >> b/arch/powerpc/sysdev/axonram.c >> index e8657d3bc588..20725006592e 100644 >> --- a/arch/powerpc/sysdev/axonram.c >> +++ b/arch/powerpc/sysdev/axonram.c > ... >> @@ -165,9 +166,13 @@ static int axon_ram_probe(struct platform_device >> *device) >> { >> static int axon_ram_bank_id = -1; >> struct axon_ram_bank *bank; >> - struct resource resource; >> + struct resource *resource; >> int rc = 0; >> >> + resource = devm_kzalloc(&device->dev, sizeof(*resource), >> GFP_KERNEL); >> + if (!resource) >> + return -ENOMEM; >> + > > Since resource is now a pointer... > > ... >> @@ -184,13 +189,13 @@ static int axon_ram_probe(struct platform_device >> *device) >> >> bank->device = device; >> >> - if (of_address_to_resource(device->dev.of_node, 0, &resource) != 0) >> { >> + if (of_address_to_resource(device->dev.of_node, 0, resource) != 0) { >> dev_err(&device->dev, "Cannot access device tree\n"); >> rc = -EFAULT; >> goto failed; >> } > ... > > ... I'd expect to see a devm_kfree call added after the failed label, > like was done here: No, because unlike dccsblk the axon_ram driver actually adheres to the device model, so it will be auto-freed on failed probe. The dcssblk implementation is not a proper "driver" so we don't get the devres_release_all() until the device is unregistered. That causes the "resource" object to stick around longer than necessary unless we force devm_kfree() it. -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
