From: Matthew Wilcox <willy@xxxxxxxxxxxxx> Sent: Wednesday, January 24, 2024 1:21 PM To: Mo, Yuezhang <Yuezhang.Mo@xxxxxxxx> Subject: Re: [PATCH] exfat: fix file not locking when writing zeros in exfat_file_mmap() > On Wed, Jan 24, 2024 at 05:00:37AM +0000, mailto:Yuezhang.Mo@xxxxxxxx wrote: > > inode->i_rwsem should be locked when writing file. But the lock > > is missing when writing zeros to the file in exfat_file_mmap(). > > This is actually very weird behaviour in exfat. This kind of "I must > manipulate the on-disc layout" is not generally done in mmap(), it's > done in ->page_mkwrite() or even delayed until we actually do writeback. > Why does exfat do this? In exfat, "valid_size" describes how far into the data stream user data has been written and "size" describes the file size. Return zeros if read "valid_size"~"size". For example, (1) xfs_io -t -f -c "pwrite -S 0x59 0 1024" $filename - Write 0x59 to 0~1023 - both "size" and "valid_size" are 1024 (2) xfs_io -t -f -c "truncate 4K" $filename - "valid_size" is still 1024 - "size" is changed to 4096 - 1024~4095 is not zeroed - return zeros if read 1024~4095 (3) xfs_io -t -f -c "mmap -rw 0 3072" -c "mwrite -S 0x5a 2048 512" $filename (3.1) "mmap -rw 0 3072" - write zeros to 1024~3071 - "valid_size" is changed to 3072 - "size" is still 4096 (3.2) "mwrite -S 0x5a 2048 512" - write 0x5a to 2048~2559 - "valid_size" is still 3072 - "size" is still 4096 To avoid 1024~2047 is not zeroed and no need to update "valid_size" in (3.2), I zero 1024~3071 in (3.1). If you have a better solution, welcome to contribute to exfat or share your solution in detail.
