Move resources allocated using unmanaged interface to managed devm interface
So today let’s talk about devm functions as that is what I have been upto the past couple of weeks. Yes, should have finished the task by now but due to some reasons was not active a couple of days
:(.
Now what are devm functions. There are some common memory resources used by drivers which are allocated using functions and stored as a linked list of arbitrarily sized memory areas called devres. Some of this memory is allocated in the probe function. But
all of this memory should be freed on a failure path or a driver detach. Else, the driver will be leaking resources and wasting precious main memory!
Detection of such memory leaks is difficult as they not clearly noticeable. So, managed interfaces have been created for common resources used by drivers. They have a name similar to the corresponding unmanaged function making it easy to identify e.g.: dma_alloc_coherent()
and dmam_alloc_coherent(), kzalloc() and devm_kzalloc(). If a managed function is used for memory allocation the resources are guaranteed to be freed both on initialization failure and driver detachment. Many of these managed functions are devm functions.
eg: devm_kzalloc, devm_iio_device_alloc().
My task has been mainly to identify the cases where kzalloc() can be changed to devm_kzalloc() and the corresponding kfree()s can be done away with. Then the changes were applied and minor code changes like removal of labels , removing of brackets in single
statement ifs were done. I have been using a Coccinelle semantic patch to do this task and have made tweaks to handle some specific cases. Here is a major part of the Coccinelle semantic patch making the change:
@platform@
identifier p, probefn, removefn;
@@
struct platform_driver p = {
.probe = probefn,
.remove = removefn,
};
@prb@
identifier platform.probefn, pdev;
expression e, e1, e2;
@@
probefn(struct platform_device *pdev, …) {
<+…
- e = kzalloc(e1, e2)
+ e = devm_kzalloc(&pdev->dev, e1, e2)
…
?-kfree(e);
…+>
}
@rem depends on prb@
identifier platform.removefn;
expression e;
@@
removefn(…) {
<…
- kfree(e);
…>
}
There are many cases in which a lot of code goes away and the probe and remove functions appear really simple and easier to understand.
At the same time, there are issues such as cases to which the patch applies but does not benefit much as there are major resources that are still unmanaged. Also, some more advanced functions that can be managed like IRQs have not been handled at this stage.
The managed resource API
The core idea behind the resource management interface is that remembering to freeallocated resources is hard. It appears to be especially hard for driverwriters who, justly or not, have a reputation for adding more than theirfair share of bugs to the kernel. And even the best driver writers can runinto trouble in situations where device probing fails halfway through; therecovery paths may be there in the code, but they tend not to be welltested. The result of all this is a fair number of resource leaks indriver code.
To address this problem, Tejun Heo created a new set of resource allocationfunctions which track allocations made by the driver. These allocationsare associated with the device structure; when the driver detachesfrom the device, any left-over allocations are cleaned up. The resourcemanagement interface is thus similar to the talloc()API used by the Samba hackers, but it is adapted to the kernelenvironment and covers more than just memory allocations.
Starting with memory allocations, though, the new API is:
void *devm_kzalloc(struct device *dev, size_t size, gfp_t gfp);
void devm_kfree(struct device *dev, void *p);
In a pattern we'll see repeated below, the new functions are similar tokzalloc() and kfree() except for the new names and theaddition of the dev argument. That argument is necessary for theresource management code to know when the memory can be freed. If anymemory allocations are still outstanding when the associated device isremoved, they will all be freed at that time.
Note that there is no managed equivalent to kalloc(); if driverwriters cannot be trusted to free memory, it seems, they cannot be trustedto initialize it either. There are also no managed versions of thepage-level or slab allocation functions.
Managed versions of a subset of the DMA allocation functions have beenprovided:
void *dmam_alloc_coherent(struct device *dev, size_t size,
dma_addr_t *dma_handle, gfp_t gfp);
void dmam_free_coherent(struct device *dev, size_t size, void *vaddr,
dma_addr_t dma_handle);
void *dmam_alloc_noncoherent(struct device *dev, size_t size,
dma_addr_t *dma_handle, gfp_t gfp);
void dmam_free_noncoherent(struct device *dev, size_t size, void *vaddr,
dma_addr_t dma_handle);
int dmam_declare_coherent_memory(struct device *dev, dma_addr_t bus_addr,
dma_addr_t device_addr, size_t size,
int flags);
void dmam_release_declared_memory(struct device *dev);
struct dma_pool *dmam_pool_create(const char *name, struct device *dev,
size_t size, size_t align,
size_t allocation);
void dmam_pool_destroy(struct dma_pool *pool);
All of these functions have the same arguments and functionality as theirdma_* equivalents, but they will clean up the DMA areas on deviceshutdown. One still has to hope that the driver has ensuredthat no DMA remains active on those areas, or unpleasant things couldhappen.
There is a managed version of pci_enable_device():
int pcim_enable_device(struct pci_dev *pdev);
There is no pcim_disable_device(), however; code should just usepci_disable_device() as usual. A new function:
void pcim_pin_device(struct pci_dev *pdev);
will cause the given pdev to be left enabled even after the driverdetaches from it.
The patch makes the allocation of I/O memory regions withpci_request_region() managed by default - there is nopcim_ version of that interface. The higher-level allocation andmapping interfaces do have managed versions:
void __iomem *pcim_iomap(struct pci_dev *pdev, int bar,
unsigned long maxlen);
void pcim_iounmap(struct pci_dev *pdev, void __iomem *addr);
For the allocation of interrupts, the managed API is:
int devm_request_irq(struct device *dev, unsigned int irq,
irq_handler_t handler, unsigned long irqflags,
const char *devname, void *dev_id);
void devm_free_irq(struct device *dev, unsigned int irq, void *dev_id);
For these functions, the addition of a struct device argument wasrequired.
There is a new set of functions for the mapping of of I/O ports and memory:
void __iomem *devm_ioport_map(struct device *dev, unsigned long port,
unsigned int nr);
void devm_ioport_unmap(struct device *dev, void __iomem *addr);
void __iomem *devm_ioremap(struct device *dev, unsigned long offset,
unsigned long size);
void __iomem *devm_ioremap_nocache(struct device *dev,
unsigned long offset,
unsigned long size);
void devm_iounmap(struct device *dev, void __iomem *addr);
Once again, these functions required the addition of a structdevice argument for the managed form.
Finally, for those using the low-level resource allocation functions, themanaged versions are:
struct resource *devm_request_region(struct device *dev,
resource_size_t start,
resource_size_t n,
const char *name);
void devm_release_region(resource_size_t start, resource_size_t n);
struct resource *devm_request_mem_region(struct device *dev,
resource_size_t start,
resource_size_t n,
const char *name);
void devm_release_mem_region(resource_size_t start, resource_size_t n);
The resource management layer includes a "group" mechanism, accessed viathese functions:
void *devres_open_group(struct device *dev, void *id, gfp_t gfp);
void devres_close_group(struct device *dev, void *id);
void devres_remove_group(struct device *dev, void *id);
int devres_release_group(struct device *dev, void *id);
A group can be thought of as a marker in the list of allocations associatedwith a given device. Groups are created with devres_open_group(),which can be passed an id value to identify the group orNULL to have the ID generated on the fly; either way, theresulting group ID is returned. A call to devres_close_group()marks the end of a given group. Calling devres_remove_group()causes the system to forget about the given group, but does nothing withthe resources allocated within the group. To remove the group andimmediately free all resources allocated within that group,devres_release_group() should be used.
The group functions seem to be primarily aimed at mid-level code - the buslayers, for example. When bus code tries to attach a driver to a device,for example, it can open a group; should the driver attach fail, the groupcan be used to free up any resources allocated by the driver.
There are not many users of this new API in the kernel now. That maychange over time as driver writers become aware of these functions, and,perhaps, as the list of managed allocation types grows. The reward forswitching over to managed allocations should be more robust and simplercode as current failure and cleanup paths are removed.
(http://lwn.net/Articles/222860/)