On Fri, Dec 18, 2020 at 12:29:10AM -0700, Allison Henderson wrote: > Because xattrs can be over a page in size, we need to handle possible > krealloc errors to avoid warnings Which warnings? > Signed-off-by: Allison Henderson <allison.henderson@xxxxxxxxxx> > --- > fs/xfs/xfs_log_recover.c | 5 ++++- > 1 file changed, 4 insertions(+), 1 deletion(-) > > diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c > index 97f3130..295a5c6 100644 > --- a/fs/xfs/xfs_log_recover.c > +++ b/fs/xfs/xfs_log_recover.c > @@ -2061,7 +2061,10 @@ xlog_recover_add_to_cont_trans( > old_ptr = item->ri_buf[item->ri_cnt-1].i_addr; > old_len = item->ri_buf[item->ri_cnt-1].i_len; > > - ptr = krealloc(old_ptr, len + old_len, GFP_KERNEL | __GFP_NOFAIL); > + ptr = krealloc(old_ptr, len + old_len, GFP_KERNEL); Does the removal of NOFAIL increase the likelihood that log recovery will fail instead of looping around looking for more memory? Hm, what /are/ we doing here, anyway? I guess someone logged a gigantic xattri item, which gets split across multiple log records, and now we're trying to staple all that back together? And perhaps the xattri item is larger than a ... page(?) which causes dmesg warnings when combined with NOFAIL? --D > + if (ptr == NULL) > + return -ENOMEM; > + > memcpy(&ptr[old_len], dp, len); > item->ri_buf[item->ri_cnt-1].i_len += len; > item->ri_buf[item->ri_cnt-1].i_addr = ptr; > -- > 2.7.4 >
