Messages in this thread Patch in this message | | | Date | Fri, 10 Jan 2014 17:25:45 -0700 | From | Shuah Khan <> | Subject | Re: [PATCH v2] dma-debug: enhance dma_debug_device_change() to check for mapping errors |
| |
On 01/07/2014 10:22 AM, Shuah Khan wrote: > On 01/07/2014 08:12 AM, Joerg Roedel wrote: >> On Tue, Jan 07, 2014 at 08:00:33AM -0700, Shuah Khan wrote: >>> This patch and a follow-on cocinelli warning fix patch are in >>> linux-next. Would you like me to send a patch relative to the change >>> in linux-next or cut a new patch against the latest Linus's git. I >>> can go either way. We just have to remember to drop those two >>> patches from linux-next. >> >> Please do a patch against linus.git and drop the two patches from >> linux-next. Over which tree did they go into linux-next anyway? >> >> > > ok. It went through mm tree. >
I have a separate routine now for scanning for entries with map error flag set and I also have a new err_printk() call to print the entry from dma_debug_device_change() right after device_dma_allocations() check and corresponding err_printk(). Snippet of the diff below. Only one warning gets printed. Looking at err_printk(): if (!show_all_errors && show_num_errors > 0) conditional, it appears that is the way it is supposed to work, unless I am missing something
One way to not miss the second warning is to just generate just a pr_err() for the mapping error instead of err_printk(). Leaked entries is a bigger problem that should really be shown and the same entry might not have the map error flag set. Thoughts?
static int dma_debug_device_change(struct notifier_block *nb, unsigned long action, void *data) { struct device *dev = data; @@ -758,6 +784,18 @@ static int dma_debug_device_change(struct notifier_block *nb, unsigned long acti "[mapped with %s] [mapped as %s]\n", count, entry->dev_addr, entry->size, dir2name[entry->direction], type2name[entry->type]); + + count = device_dma_map_errors(dev, &entry); + if (count == 0) + break; + err_printk(dev, entry, + "DMA-API: device driver failed to check map err" + ": [count=%d]\n" + "Details of an entry with map error flag set: " + "[device address=0x%016llx] [size=%llu bytes] " + "[mapped with %s] [mapped as %s]\n", + count, entry->dev_addr, entry->size, + dir2name[entry->direction], type2name[entry->type]); break;
-- Shuah Khan Senior Linux Kernel Developer - Open Source Group Samsung Research America(Silicon Valley) shuah.kh@samsung.com | (970) 672-0658
| |