Message ID | 20211210184636.7273-2-jim2101024@gmail.com |
---|---|
State | Accepted, archived |
Headers | show |
Series | of: unittest: fix warning on PowerPC frame size warning | expand |
Context | Check | Description |
---|---|---|
robh/checkpatch | warning | total: 0 errors, 1 warnings, 29 lines checked |
robh/dtbs-check | success | |
robh/dt-meta-schema | success |
moved the file's maintainers from the "cc:" list to the "to:" list review comments below On 12/10/21 1:46 PM, Jim Quinlan wrote: > The struct device variable "dev_bogus" was triggering this warning > on a PowerPC build: > > drivers/of/unittest.c: In function 'of_unittest_dma_ranges_one.constprop': > [...] >> The frame size of 1424 bytes is larger than 1024 bytes > [-Wframe-larger-than=] > > This variable is now dynamically allocated. A side effect of the change is that dev_bogus is initialized to all zeros instead of containing random data from the stack. > > Fixes: e0d072782c734 ("dma-mapping: introduce DMA range map, supplanting dma_pfn_offset") > Reported-by: kernel test robot <lkp@intel.com> > Signed-off-by: Jim Quinlan <jim2101024@gmail.com> > --- > drivers/of/unittest.c | 16 ++++++++++++---- > 1 file changed, 12 insertions(+), 4 deletions(-) > > diff --git a/drivers/of/unittest.c b/drivers/of/unittest.c > index 481ba8682ebf..02c5cd06ad19 100644 > --- a/drivers/of/unittest.c > +++ b/drivers/of/unittest.c > @@ -911,11 +911,18 @@ static void __init of_unittest_dma_ranges_one(const char *path, > if (!rc) { > phys_addr_t paddr; > dma_addr_t dma_addr; > - struct device dev_bogus; > + struct device *dev_bogus; > > - dev_bogus.dma_range_map = map; > - paddr = dma_to_phys(&dev_bogus, expect_dma_addr); > - dma_addr = phys_to_dma(&dev_bogus, expect_paddr); > + dev_bogus = kzalloc(sizeof(struct device), GFP_KERNEL); > + if (!dev_bogus) { > + unittest(0, "kzalloc() failed\n"); > + kfree(map); > + return; > + } > + > + dev_bogus->dma_range_map = map; > + paddr = dma_to_phys(dev_bogus, expect_dma_addr); > + dma_addr = phys_to_dma(dev_bogus, expect_paddr); > > unittest(paddr == expect_paddr, > "of_dma_get_range: wrong phys addr %pap (expecting %llx) on node %pOF\n", > @@ -925,6 +932,7 @@ static void __init of_unittest_dma_ranges_one(const char *path, > &dma_addr, expect_dma_addr, np); > > kfree(map); > + kfree(dev_bogus); > } > of_node_put(np); > #endif > Reviewed-by: Frank Rowand <frank.rowand@sony.com>
Looks good:
Reviewed-by: Christoph Hellwig <hch@lst.de>
On 12/10/21 10:46 AM, Jim Quinlan wrote: > The struct device variable "dev_bogus" was triggering this warning > on a PowerPC build: > > drivers/of/unittest.c: In function 'of_unittest_dma_ranges_one.constprop': > [...] >> The frame size of 1424 bytes is larger than 1024 bytes > [-Wframe-larger-than=] > > This variable is now dynamically allocated. > > Fixes: e0d072782c734 ("dma-mapping: introduce DMA range map, supplanting dma_pfn_offset") > Reported-by: kernel test robot <lkp@intel.com> > Signed-off-by: Jim Quinlan <jim2101024@gmail.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
On Fri, 10 Dec 2021 13:46:35 -0500, Jim Quinlan wrote: > The struct device variable "dev_bogus" was triggering this warning > on a PowerPC build: > > drivers/of/unittest.c: In function 'of_unittest_dma_ranges_one.constprop': > [...] >> The frame size of 1424 bytes is larger than 1024 bytes > [-Wframe-larger-than=] > > This variable is now dynamically allocated. > > Fixes: e0d072782c734 ("dma-mapping: introduce DMA range map, supplanting dma_pfn_offset") > Reported-by: kernel test robot <lkp@intel.com> > Signed-off-by: Jim Quinlan <jim2101024@gmail.com> > --- > drivers/of/unittest.c | 16 ++++++++++++---- > 1 file changed, 12 insertions(+), 4 deletions(-) > Applied, thanks!
diff --git a/drivers/of/unittest.c b/drivers/of/unittest.c index 481ba8682ebf..02c5cd06ad19 100644 --- a/drivers/of/unittest.c +++ b/drivers/of/unittest.c @@ -911,11 +911,18 @@ static void __init of_unittest_dma_ranges_one(const char *path, if (!rc) { phys_addr_t paddr; dma_addr_t dma_addr; - struct device dev_bogus; + struct device *dev_bogus; - dev_bogus.dma_range_map = map; - paddr = dma_to_phys(&dev_bogus, expect_dma_addr); - dma_addr = phys_to_dma(&dev_bogus, expect_paddr); + dev_bogus = kzalloc(sizeof(struct device), GFP_KERNEL); + if (!dev_bogus) { + unittest(0, "kzalloc() failed\n"); + kfree(map); + return; + } + + dev_bogus->dma_range_map = map; + paddr = dma_to_phys(dev_bogus, expect_dma_addr); + dma_addr = phys_to_dma(dev_bogus, expect_paddr); unittest(paddr == expect_paddr, "of_dma_get_range: wrong phys addr %pap (expecting %llx) on node %pOF\n", @@ -925,6 +932,7 @@ static void __init of_unittest_dma_ranges_one(const char *path, &dma_addr, expect_dma_addr, np); kfree(map); + kfree(dev_bogus); } of_node_put(np); #endif
The struct device variable "dev_bogus" was triggering this warning on a PowerPC build: drivers/of/unittest.c: In function 'of_unittest_dma_ranges_one.constprop': [...] >> The frame size of 1424 bytes is larger than 1024 bytes [-Wframe-larger-than=] This variable is now dynamically allocated. Fixes: e0d072782c734 ("dma-mapping: introduce DMA range map, supplanting dma_pfn_offset") Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: Jim Quinlan <jim2101024@gmail.com> --- drivers/of/unittest.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-)