| Message ID | 512B339A.7010606@ce.jp.nec.com (mailing list archive) |
|---|---|
| State | Deferred, archived |
| Headers | show |
On 02/25/13 10:49, Jun'ichi Nomura wrote: > diff --git a/drivers/md/dm.c b/drivers/md/dm.c > index 314a0e2..51fefb5 100644 > --- a/drivers/md/dm.c > +++ b/drivers/md/dm.c > @@ -1973,15 +1973,27 @@ static void __bind_mempools(struct mapped_device *md, struct dm_table *t) > { > struct dm_md_mempools *p = dm_table_get_md_mempools(t); > > - if (md->io_pool && (md->tio_pool || dm_table_get_type(t) == DM_TYPE_BIO_BASED) && md->bs) { > - /* > - * The md already has necessary mempools. Reload just the > - * bioset because front_pad may have changed because > - * a different table was loaded. > - */ > - bioset_free(md->bs); > - md->bs = p->bs; > - p->bs = NULL; > + if (md->io_pool && md->bs) { > + /* The md already has necessary mempools. */ > + if (dm_table_get_type(t) == DM_TYPE_BIO_BASED) { > + /* > + * Reload bioset because front_pad may have changed > + * because a different table was loaded. > + */ > + bioset_free(md->bs); > + md->bs = p->bs; > + p->bs = NULL; > + } else if (dm_table_get_type(t) == DM_TYPE_REQUEST_BASED) { > + BUG_ON(!md->tio_pool); > + /* > + * No need to reload in case of request-based dm > + * because of fixed size front_pad. > + * Note for future: if you are to reload bioset, > + * prep-ed requests in queue may have reference > + * to bio from the old bioset. > + * So you must walk through the queue to unprep. > + */ > + } > goto out; > } Without your patch my test failed after two or three iterations. With your patch my test is still running after 53 iterations. So if you want you can add Tested-by: Bart Van Assche <bvanassche@acm.org>. Your e-mail and the above patch are also interesting because these explain why reverting to the v3.7 of drivers/md made my test succeed. Note: even if this patch gets accepted I think it's still useful to modify blk_run_queue() such that it converts recursion into iteration. Bart. -- dm-devel mailing list dm-devel@redhat.com https://www.redhat.com/mailman/listinfo/dm-devel
On 02/26/13 00:09, Bart Van Assche wrote: > Without your patch my test failed after two or three iterations. With your patch my test is still running after 53 iterations. So if you want you can add Tested-by: Bart Van Assche <bvanassche@acm.org>. Great. Thanks for testing. I'll submit a patch with your Reported-by and Tested-by. > Your e-mail and the above patch are also interesting because these explain why reverting to the v3.7 of drivers/md made my test succeed. > > Note: even if this patch gets accepted I think it's still useful to modify blk_run_queue() such that it converts recursion into iteration. Yes. That's a separate discussion. Though I'm not sure if it's ok in general to implicitly convert sync run-queue to async one.
diff --git a/drivers/md/dm.c b/drivers/md/dm.c index 314a0e2..51fefb5 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -1973,15 +1973,27 @@ static void __bind_mempools(struct mapped_device *md, struct dm_table *t) { struct dm_md_mempools *p = dm_table_get_md_mempools(t); - if (md->io_pool && (md->tio_pool || dm_table_get_type(t) == DM_TYPE_BIO_BASED) && md->bs) { - /* - * The md already has necessary mempools. Reload just the - * bioset because front_pad may have changed because - * a different table was loaded. - */ - bioset_free(md->bs); - md->bs = p->bs; - p->bs = NULL; + if (md->io_pool && md->bs) { + /* The md already has necessary mempools. */ + if (dm_table_get_type(t) == DM_TYPE_BIO_BASED) { + /* + * Reload bioset because front_pad may have changed + * because a different table was loaded. + */ + bioset_free(md->bs); + md->bs = p->bs; + p->bs = NULL; + } else if (dm_table_get_type(t) == DM_TYPE_REQUEST_BASED) { + BUG_ON(!md->tio_pool); + /* + * No need to reload in case of request-based dm + * because of fixed size front_pad. + * Note for future: if you are to reload bioset, + * prep-ed requests in queue may have reference + * to bio from the old bioset. + * So you must walk through the queue to unprep. + */ + } goto out; }