On Wed, Sep 6, 2017 at 10:28 PM, Dan Carpenter <dan.carpenter@xxxxxxxxxx> wrote: > On Sun, Sep 03, 2017 at 11:56:50AM +0300, Gilad Ben-Yossef wrote: >> @@ -1154,7 +1150,8 @@ static inline int ssi_buffer_mgr_aead_chain_data( >> //if have reached the end of the sgl, then this is unexpected >> if (!areq_ctx->src_sgl) { >> SSI_LOG_ERR("reached end of sg list. unexpected\n"); >> - BUG(); >> + return -EINVAL; >> + goto chain_data_exit; > > You've got a direct return followed by a goto. Yes, that is silly. Thank you for spotting that. > > It's a do-nothing goto that just returns rc. I hate those. I've tried > to review locking bugs to see if single returns prevent future > programmers from introducing new error paths which don't unlock. They > don't really help... I've replaced the error handling from goto to return as you suggested for the changes introduced by this patch. Still need to clean up a lot of the other code doing this though. Gilad -- Gilad Ben-Yossef Chief Coffee Drinker "If you take a class in large-scale robotics, can you end up in a situation where the homework eats your dog?" -- Jean-Baptiste Queru _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel