[PATCH v6 0/8] gpiolib: fix chip removal with lines still requested from user space
From: Adrian Ambrozewicz
Date: Mon Nov 16 2026 - 08:47:34 EST
Removing a GPIO chip while user space holds a line request, as when a
hot-pluggable GPIO controller goes away, leaks the line labels (patch 2)
and IRQ names (patch 5), frees IRQs twice (patches 3, 7 and 8) and lets
a blocked read() stall the removal (patch 4). Patch 1 makes
gpiodev_release() wait for pending label frees; without it,
cleanup_srcu_struct() can warn, more often with patch 2.
Patch 7 also affects in-kernel users: their remaining IRQs are freed
after the chip pointer is cleared, when gpiod_*() calls fail. Patch 6
drops edge events whose line read fails instead of reporting an edge.
User-visible change: a read() blocked when the chip is removed returns
the queued events, then fails with -ENODEV.
For stable: patch 2 needs 1, patch 7 needs 6 and patch 8 needs 7;
patches 3, 5 and 7 fix a 6.13 commit also in 6.12.y, patch 8 a 6.19
commit; patch 4 applies to 6.12.y.
Tested under QEMU (x86_64, arm64) with gpio-sim, a shared gpio-sim line
and a test chip on the hierarchical irqchip, with lockdep, kmemleak and
KASAN.
Prepared with Cursor using Claude, which helped analyse the bugs, write
the reproducers and draft the patches; I reviewed and tested every
change.
Changes in v6:
- Rebased on gpio/for-current (patch 2 over the flags rework).
- New patch 8: remove the shared GPIO proxies first (Sashiko).
- Patch 7: cite Herve's proposal as [1] (Andy, Herve).
- Dropped the Assisted-by tags (removed upstream).
- Link to v5: https://lore.kernel.org/r/20261103-gpiolib-chip-removal-v5-0-aambrozewicz@xxxxxxxxxx
Changes in v5:
- Rebased on gpio/for-current.
- Patch 6: describe driver ->get() errors; new Fixes tags (Kent,
Bartosz).
- Link to v4: https://lore.kernel.org/r/20261029-gpiolib-chip-removal-v4-0-aambrozewicz@xxxxxxxxxx
Changes in v4:
- Rebased on gpio/for-current.
- New patch 6: drop edge events when a line read fails (Sashiko).
- Patch 2: the NULL label fix was applied separately (Sashiko):
https://lore.kernel.org/r/20261022083343.3905973-3-aambrozewicz@xxxxxxxxxx
- Patch 5: moving irq_label above irq in struct line moves the 4-byte
hole rather than removing it (Andy).
- Link to v3: https://lore.kernel.org/r/20261020-gpiolib-chip-removal-v3-0-aambrozewicz@xxxxxxxxxx
Changes in v3:
- Rebased on gpio/for-current; shorter commit messages (Kent, Bartosz).
- Patch 6 (now 7): return early when there is no ->to_irq(); drop the
lineevent_free() comment (Kent).
- Sashiko's v2 6/6 finding on gpiochip_setup_dev(): separate fix
planned.
- Link to v2: https://lore.kernel.org/r/20261016-gpiolib-chip-removal-v2-0-aambrozewicz@xxxxxxxxxx
- Link to v1: https://lore.kernel.org/r/20261012-gpiolib-chip-removal-v1-0-aambrozewicz@xxxxxxxxxx
Adrian Ambrozewicz (8):
gpiolib: wait for pending label frees before cleaning up desc_srcu
gpiolib: free the label of lines released after their chip was removed
gpiolib: cdev: do not free line IRQs again after the chip is removed
gpiolib: cdev: don't sleep in the SRCU read section in read()
gpiolib: cdev: don't leak the IRQ name when the chip is removed
gpiolib: cdev: don't report a failed line read as an edge
gpiolib: let the character device free its IRQs first on chip removal
gpiolib: remove the shared GPIO proxies before freeing the remaining
IRQs
drivers/gpio/gpiolib-cdev.c | 155 +++++++++++++++++++++++++++---------
drivers/gpio/gpiolib.c | 38 +++++++--
2 files changed, 147 insertions(+), 46 deletions(-)
base-commit: 3e68810e7c109ae4cc23541f36a1a9fc8ed65700
--
2.43.0