> Begin forwarded message: > > From: Greg KH <greg@xxxxxxxxx> > Subject: Re: Linux kernel development reports for the 4.7 release > Date: July 28, 2016 at 6:48:24 PM ADT > To: Aaron Campbell <aaron@xxxxxxxxxx> […] >> Thank you for your work in maintaining -stable! We upgraded from 3.2 to >> 4.1-stable about 6 months ago and had only one regression, which we >> fixed by pulling in the following commits: >> >> aa65fa3 may_follow_link() should use nd->inode >> 97242f9 link_path_walk(): be careful when failing with ENOTDIR >> >> Actually, it looks like those still haven’t hit the 4.1 branch? If you are >> interested I could dig up the shell script I used to trigger these bugs. > > Yes, please email stable@xxxxxxxxxxxxxxx and let the developers know > that those patches need to be applied. Note the following regression observed in 4.1-stable (specifically we were using 4.1.15 when this was tested): $ while [ 1 ]; do > cat /proc/net/dev 1>/dev/null > done cat: /proc/net/dev: Not a directory cat: /proc/net/dev: Not a directory … YMMV on how many loop iterations it will take to reproduce the ENOTDIR errors. It seems to be some sort of race, and may or may not be easier to trigger on a machine with more cores. We fixed the regression pulling in aa65fa3 and 97242f9, neither of which are currently on the 4.1-stable branch: $ git branch —contains aa65fa3 linux-4.3.y linux-4.6.y master $ git branch —contains 97242f9 linux-4.3.y linux-4.6.y master $ We have been running with these patches in our 4.1 tree for 6+ months with no issues. Thanks, -Aaron-- To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html