On 12/05/2012 08:13 AM, Nicholas A. Bellinger wrote:
diff --git a/drivers/target/target_core_file.c b/drivers/target/target_core_file.c index 0360383..319b87a 100644 --- a/drivers/target/target_core_file.c +++ b/drivers/target/target_core_file.c @@ -260,7 +260,7 @@ static int fd_do_readv(struct se_cmd *cmd, struct scatterlist *sgl, for_each_sg(sgl, sg, sgl_nents, i) { iov[i].iov_len = sg->length; - iov[i].iov_base = sg_virt(sg); + iov[i].iov_base = kmap(sg_page(sg)); } old_fs = get_fs(); @@ -268,6 +268,10 @@ static int fd_do_readv(struct se_cmd *cmd, struct scatterlist *sgl, ret = vfs_readv(fd,&iov[0], sgl_nents,&pos); set_fs(old_fs); + for_each_sg(sgl, sg, sgl_nents, i) { + kunmap(sg_page(sg)); + } + kfree(iov); /* * Return zeros and GOOD status even if the READ did not return Does not work. It bails out even earlier, before the XFS mount completes. That means the complete code is not highmem aware and it simply works because page_address() returns something that does not cause a fault and the memory is not important enough to crash the box.Mmmm.. Christoph, what do you think for the highmem FILEIO case..?
Just sent a fix. I forgot to add the offset… Oh well. Sebastian -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
