bug#80516: df --total returns wrong values
Hi Collin. Thanks for the reply. My report wasn't about the display of the duplicate entries, though. I understand that's how it's supposed to work. My issue was with the total storage display. It simply adds up everything that's shown, which leads to an incorrect result. I do not have 2.6T of disk storage and I do not have 1.5T of stored data. The btrfs-disk has a size of 222G in total and the ntfs-volume has a size of 932G in total. Therefore, the entire row containing the totals (except for the percentage) displays incorrect values. The explanation written by Pádraig Brady shows that the problem is already known. My problem is that I want to use the output of 'df' for a zsh plugin that monitors disk usage during cp/mv operations, and the incorrect values then corrupt the return values of my plugin. I will probably have to come up with a different strategy. Best regards, Tom Am So., 1. März 2026 um 10:27 Uhr schrieb Collin Funk < [email protected]>: > Hi Thomas, > > Rorschach writes: > > > The `df --total` command returns incorrect values for total disk space > > (total, used, available, etc.). It simply adds all values without > > considering whether they are already in use. See the CSV file (output of > > `df --total`) and the LibreOffice Calc spreadsheet, where the values are > > calculated correctly by excluding duplicates. > > I am copying the 'df' output you shared, so that others on the mailing > list can see it more easily: > > Dateisystem Typ Größe Benutzt Verf. Verw% Eingehängt auf > dev devtmpfs32G 0 32G 0% /dev > run tmpfs 32G 2,4M32G 1% /run > efivarfsefivarfs128K47K 77K 38% > /sys/firmware/efi/efivars > /dev/sda2 btrfs 222G126G94G 58% / > tmpfs tmpfs 32G 0 32G 0% /dev/shm > nonetmpfs 1,0M0 1,0M0% > /run/credentials/systemd-journald.service > nonetmpfs 1,0M0 1,0M0% > /run/credentials/systemd-resolved.service > /dev/sda2 btrfs 222G126G94G 58% /srv > /dev/sda2 btrfs 222G126G94G 58% /var/tmp > /dev/sda2 btrfs 222G126G94G 58% /root > /dev/sda2 btrfs 222G126G94G 58% /var/log > /dev/sda2 btrfs 222G126G94G 58% /home > /dev/sda2 btrfs 222G126G94G 58% /var/cache > tmpfs tmpfs 32G 76K 32G 1% /tmp > /dev/sda1 vfat2,0G440M1,6G22% /boot > /dev/sdb1 ntfs3 932G583G349G63% /home/tom/Daten > tmpfs tmpfs 6,3G72K 6,3G1% /run/user/1000 > total - 2,6T1,5T1,2T57% - > > It looks like the duplicate entries you mention are btrfs subvolumes. If > so, that is the expected behavior, albeit a bit awkward. > > I will link a good explanation that Pádraig Brady wrote, instead of > poorly paraphrasing it [1]. > > Collin > > [1] https://debbugs.gnu.org/cgi/bugreport.cgi?bug=17676#8 >
bug#80516: df --total returns wrong values
Hi Thomas, Rorschach writes: > The `df --total` command returns incorrect values for total disk space > (total, used, available, etc.). It simply adds all values without > considering whether they are already in use. See the CSV file (output of > `df --total`) and the LibreOffice Calc spreadsheet, where the values are > calculated correctly by excluding duplicates. I am copying the 'df' output you shared, so that others on the mailing list can see it more easily: Dateisystem Typ Größe Benutzt Verf. Verw% Eingehängt auf dev devtmpfs32G 0 32G 0% /dev run tmpfs 32G 2,4M32G 1% /run efivarfsefivarfs128K47K 77K 38% /sys/firmware/efi/efivars /dev/sda2 btrfs 222G126G94G 58% / tmpfs tmpfs 32G 0 32G 0% /dev/shm nonetmpfs 1,0M0 1,0M0% /run/credentials/systemd-journald.service nonetmpfs 1,0M0 1,0M0% /run/credentials/systemd-resolved.service /dev/sda2 btrfs 222G126G94G 58% /srv /dev/sda2 btrfs 222G126G94G 58% /var/tmp /dev/sda2 btrfs 222G126G94G 58% /root /dev/sda2 btrfs 222G126G94G 58% /var/log /dev/sda2 btrfs 222G126G94G 58% /home /dev/sda2 btrfs 222G126G94G 58% /var/cache tmpfs tmpfs 32G 76K 32G 1% /tmp /dev/sda1 vfat2,0G440M1,6G22% /boot /dev/sdb1 ntfs3 932G583G349G63% /home/tom/Daten tmpfs tmpfs 6,3G72K 6,3G1% /run/user/1000 total - 2,6T1,5T1,2T57% - It looks like the duplicate entries you mention are btrfs subvolumes. If so, that is the expected behavior, albeit a bit awkward. I will link a good explanation that Pádraig Brady wrote, instead of poorly paraphrasing it [1]. Collin [1] https://debbugs.gnu.org/cgi/bugreport.cgi?bug=17676#8
bug#80516: df --total returns wrong values
The `df --total` command returns incorrect values for total disk space (total, used, available, etc.). It simply adds all values without considering whether they are already in use. See the CSV file (output of `df --total`) and the LibreOffice Calc spreadsheet, where the values are calculated correctly by excluding duplicates. Best regard Thomas Bernard Berlin (Germany) Dateisystem Typ Größe Benutzt Verf. Verw% Eingehängt auf dev devtmpfs 32G 0 32G 0% /dev run tmpfs 32G 2,4M 32G 1% /run efivarfs efivarfs 128K 47K 77K 38% /sys/firmware/efi/efivars /dev/sda2 btrfs 222G 126G 94G 58% / tmpfs tmpfs 32G 0 32G 0% /dev/shm none tmpfs 1,0M 0 1,0M 0% /run/credentials/systemd-journald.service none tmpfs 1,0M 0 1,0M 0% /run/credentials/systemd-resolved.service /dev/sda2 btrfs 222G 126G 94G 58% /srv /dev/sda2 btrfs 222G 126G 94G 58% /var/tmp /dev/sda2 btrfs 222G 126G 94G 58% /root /dev/sda2 btrfs 222G 126G 94G 58% /var/log /dev/sda2 btrfs 222G 126G 94G 58% /home /dev/sda2 btrfs 222G 126G 94G 58% /var/cache tmpfs tmpfs 32G 76K 32G 1% /tmp /dev/sda1 vfat 2,0G 440M 1,6G 22% /boot /dev/sdb1 ntfs3 932G 583G 349G 63% /home/tom/Daten tmpfs tmpfs 6,3G 72K 6,3G 1% /run/user/1000 total - 2,6T 1,5T 1,2T 57% - wrong-df-total Description: Binary data
