On Tue 24-01-23 05:25:09, Christoph Hellwig wrote: > As a pure mechanical move this looks good: > > Reviewed-by: Christoph Hellwig <hch@xxxxxx> Thanks! > But some comments: > > > + struct inode *inode = page->mapping->host; > > + char *kaddr; > > + struct udf_inode_info *iinfo = UDF_I(inode); > > + > > + BUG_ON(!PageLocked(page)); > > + > > + kaddr = kmap_atomic(page); > > + memcpy(iinfo->i_data + iinfo->i_lenEAttr, kaddr, i_size_read(inode)); > > + SetPageUptodate(page); > > + kunmap_atomic(kaddr); > > + unlock_page(page); > > This really should be using memcpy_to_page. And the SetPageUptodate > here in ->writepages loos a little odd as well. Good points. Added a cleanup patch for this. And then one more to get rid of the last remaining kmap_atomic() use in UDF (BTW that would benefit from memcpy_to_page_page() helper as well). Honza -- Jan Kara <jack@xxxxxxxx> SUSE Labs, CR
