2013-05-31 21:22:08

by Suman Anna

Subject: [PATCH 3/3] remoteproc: free carveout memories only after unmapping them

It is not preferable to have the allocated pages for carveout
memories freed before they are unmapped. The code that deals
with the cleanup of carveout memories is therefore moved after
the corresponding mapping entries were cleaned up.

This is mostly a no-op since the remote processors are already
stopped when the cleanup function is called, but this will make
the cleanup code follow the exact reverse path of allocation.

Signed-off-by: Suman Anna <[email protected]>
---
drivers/remoteproc/remoteproc_core.c | 16 ++++++++--------
1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c
index bc078c0..1ea7460 100644
--- a/drivers/remoteproc/remoteproc_core.c
+++ b/drivers/remoteproc/remoteproc_core.c
@@ -762,14 +762,6 @@ static void rproc_resource_cleanup(struct rproc *rproc)
kfree(entry);
}

- /* clean up carveout allocations */
- list_for_each_entry_safe(entry, tmp, &rproc->carveouts, node) {
- dma_free_coherent(dev->parent, entry->len, entry->va,
- entry->dma);
- list_del(&entry->node);
- kfree(entry);
- }
-
/* clean up iommu mapping entries */
list_for_each_entry_safe(entry, tmp, &rproc->mappings, node) {
size_t unmapped;
@@ -784,6 +776,14 @@ static void rproc_resource_cleanup(struct rproc *rproc)
list_del(&entry->node);
kfree(entry);
}
+
+ /* clean up carveout allocations */
+ list_for_each_entry_safe(entry, tmp, &rproc->carveouts, node) {
+ dma_free_coherent(dev->parent, entry->len, entry->va,
+ entry->dma);
+ list_del(&entry->node);
+ kfree(entry);
+ }
}

/*
--
1.8.2

2013-07-01 14:06:32

by Ohad Ben Cohen

Subject: Re: [PATCH 3/3] remoteproc: free carveout memories only after unmapping them

On Sat, Jun 1, 2013 at 12:16 AM, Suman Anna <[email protected]> wrote:
> It is not preferable to have the allocated pages for carveout
> memories freed before they are unmapped. The code that deals
> with the cleanup of carveout memories is therefore moved after
> the corresponding mapping entries were cleaned up.
>
> This is mostly a no-op since the remote processors are already
> stopped when the cleanup function is called, but this will make
> the cleanup code follow the exact reverse path of allocation.
>
> Signed-off-by: Suman Anna <[email protected]>

Applied, thanks.