On 2024-02-08 16:32, Dan Williams wrote:
Mathieu Desnoyers wrote:
In preparation for checking whether the architecture has data cache
aliasing within alloc_dax(), modify the error handling of nvdimm/pmem
pmem_attach_disk() to treat alloc_dax() -EOPNOTSUPP failure as non-fatal.
For the transition, consider that alloc_dax() returning NULL is the
same as returning -EOPNOTSUPP.
Fixes: d92576f1167c ("dax: does not work correctly with virtual aliasing caches")
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@xxxxxxxxxxxx>
Cc: Alasdair Kergon <agk@xxxxxxxxxx>
Cc: Mike Snitzer <snitzer@xxxxxxxxxx>
Cc: Mikulas Patocka <mpatocka@xxxxxxxxxx>
Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Cc: Dan Williams <dan.j.williams@xxxxxxxxx>
Cc: Vishal Verma <vishal.l.verma@xxxxxxxxx>
Cc: Dave Jiang <dave.jiang@xxxxxxxxx>
Cc: Matthew Wilcox <willy@xxxxxxxxxxxxx>
Cc: Arnd Bergmann <arnd@xxxxxxxx>
Cc: Russell King <linux@xxxxxxxxxxxxxxx>
Cc: linux-arch@xxxxxxxxxxxxxxx
Cc: linux-cxl@xxxxxxxxxxxxxxx
Cc: linux-fsdevel@xxxxxxxxxxxxxxx
Cc: linux-mm@xxxxxxxxx
Cc: linux-xfs@xxxxxxxxxxxxxxx
Cc: dm-devel@xxxxxxxxxxxxxxx
Cc: nvdimm@xxxxxxxxxxxxxxx
---
drivers/nvdimm/pmem.c | 26 ++++++++++++++------------
1 file changed, 14 insertions(+), 12 deletions(-)
diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c
index 9fe358090720..f1d9f5c6dbac 100644
--- a/drivers/nvdimm/pmem.c
+++ b/drivers/nvdimm/pmem.c
@@ -558,19 +558,21 @@ static int pmem_attach_disk(struct device *dev,
disk->bb = &pmem->bb;
dax_dev = alloc_dax(pmem, &pmem_dax_ops);
- if (IS_ERR(dax_dev)) {
- rc = PTR_ERR(dax_dev);
- goto out;
+ if (IS_ERR_OR_NULL(dax_dev)) {
alloc_dax() should never return NULL. I.e. the lead in before this patch
should fix this misunderstanding:
diff --git a/include/linux/dax.h b/include/linux/dax.h
index b463502b16e1..df2d52b8a245 100644
--- a/include/linux/dax.h
+++ b/include/linux/dax.h
@@ -86,11 +86,7 @@ static inline void *dax_holder(struct dax_device *dax_dev)
static inline struct dax_device *alloc_dax(void *private,
const struct dax_operations *ops)
{
- /*
- * Callers should check IS_ENABLED(CONFIG_DAX) to know if this
- * NULL is an error or expected.
- */
- return NULL;
+ return ERR_PTR(-EOPNOTSUPP);
}
static inline void put_dax(struct dax_device *dax_dev)
{
+ rc = IS_ERR(dax_dev) ? PTR_ERR(dax_dev) : -EOPNOTSUPP;
Then this ternary can be replaced with just a check of which PTR_ERR()
value is being returned.
As you noted, I've introduced this as cleanups in later patches. I don't
mind folding these into their respective per-driver commits and moving
the alloc_dax() hunk earlier in the series.
Thanks,
Mathieu
--
Mathieu Desnoyers
EfficiOS Inc.
https://www.efficios.com
