在 2026/9/22 12:19, Oscar Salvador (SUSE) 写道: > On Thu, Sep 03, 2026 at 01:35:35PM +0800, Kaitao Cheng wrote: >> From: Kaitao Cheng <[email protected]> >> >> TestClearPageHWPoison() is now defined regardless of whether >> CONFIG_MEMORY_FAILURE is enabled, returning false when memory failure >> handling is unavailable. >> >> The test_and_clear_pmem_poison() wrapper duplicates this configuration >> handling and has no other pmem-specific behavior. >> >> Call TestClearPageHWPoison() directly and remove the redundant wrapper. >> This also removes the need to include page-flags.h from pmem.h. >> >> No functional change is intended. >> >> Signed-off-by: Kaitao Cheng <[email protected]> > > Nice cleanup, but as David noted, please squash patch#1 into this one.
Although NVDIMM is currently the only external user of TestClearPageHWPoison, I still consider it a generic declaration. Combining the two patches might give the impression that TestClearPageHWPoison exists solely to serve a specific driver. That said, I’m not opposed to merging the two patches, as doing so would make the submission more concise. I think both approaches have their own advantages and disadvantages, which is why I haven’t sent a v2 yet. > Reviewed-by: Oscar Salvador <[email protected]> > > -- Thanks Kaitao Cheng
