On Tue, Mar 03, 2020 at 05:18:40PM -0800, Darrick J. Wong wrote: > On Tue, Mar 03, 2020 at 10:51:22PM +0000, Domenico Andreoli wrote: > > > > I don't see the need of reverting anything, I can deal with these > > issues if you are busy on something else. > > If you want to work on the patch, please do! Starting from the revert > patch I sent earlier, I /think/ only the first chunk (the one that > touches blkdev_write_iter) of that patch actually has to be applied to > re-enable uswsusp. That could probably be turned into: > > if (IS_SWAPFILE(...) && !IS_ENABLED(HIBERNATION)) > return -ETXTBSY; I've just sent such patch, I don't know how it will play with the whole revert of yesterday and that akpm has already taken in his tree. Ideally this should go in 5.6-rc and also in stable kernels > 5.2. > > Though perhaps a better thing to check here rather than the Kconfig > option is whether or not the system is locked out against hibernation? > e.g., > > if (IS_SWAPFILE(...) && !hibernation_available()) > return -EXTBSY; This is the kind of improved fix I'm going to prepare for a coming merge window. Regards, Domenico -- rsa4096: 3B10 0CA1 8674 ACBA B4FE FCD2 CE5B CF17 9960 DE13 ed25519: FFB4 0CC3 7F2E 091D F7DA 356E CC79 2832 ED38 CB05