"value":"In the Linux kernel, the following vulnerability has been resolved:\n\ncifs: Fix writeback data corruption\n\ncifs writeback doesn't correctly handle the case where\ncifs_extend_writeback() hits a point where it is considering an additional\nfolio, but this would overrun the wsize - at which point it drops out of\nthe xarray scanning loop and calls xas_pause(). The problem is that\nxas_pause() advances the loop counter - thereby skipping that page.\n\nWhat needs to happen is for xas_reset() to be called any time we decide we\ndon't want to process the page we're looking at, but rather send the\nrequest we are building and start a new one.\n\nFix this by copying and adapting the netfslib writepages code as a\ntemporary measure, with cifs writeback intending to be offloaded to\nnetfslib in the near future.\n\nThis also fixes the issue with the use of filemap_get_folios_tag() causing\nretry of a bunch of pages which the extender already dealt with.\n\nThis can be tested by creating, say, a 64K file somewhere not on cifs\n(otherwise copy-offload may get underfoot), mounting a cifs share with a\nwsize of 64000, copying the file to it and then comparing the original file\nand the copy:\n\n dd if=/dev/urandom of=/tmp/64K bs=64k count=1\n mount //192.168.6.1/test /mnt -o user=...,pass=...,wsize=64000\n cp /tmp/64K /mnt/64K\n cmp /tmp/64K /mnt/64K\n\nWithout the fix, the cmp fails at position 64000 (or shortly thereafter)."
