| Message ID | 20210902070025.197072-1-philmd@redhat.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | block/nvme: Rework error reporting | expand |
On Thu, Sep 02, 2021 at 09:00:14AM +0200, Philippe Mathieu-Daudé wrote: > (Series fully reviewed) > > Hi, > > This series contains various patches sent last year with > review comments addressed, few more cleanups, and a new > patch which remove the spurious "VFIO_MAP_DMA failed: No > space left on device" now poping up since commit 15a730e7a. > (it is the expected behavior, which is why we retry the > same call after flushing the DMA mappings). > > Since v2: > - qemu_vfio_find_[fixed/temp]_iova retun bool (Klaus) > - Add Klaus's R-b > > Since v1: > - Addressed Klaus review comments (cleaner Error* handling) > - Add Klaus's R-b > > Regards, > > Phil. > > Philippe Mathieu-Daudé (11): > block/nvme: Use safer trace format string > util/vfio-helpers: Let qemu_vfio_verify_mappings() use error_report() > util/vfio-helpers: Replace qemu_mutex_lock() calls with > QEMU_LOCK_GUARD > util/vfio-helpers: Remove unreachable code in qemu_vfio_dma_map() > block/nvme: Have nvme_create_queue_pair() report errors consistently > util/vfio-helpers: Pass Error handle to qemu_vfio_dma_map() > util/vfio-helpers: Extract qemu_vfio_water_mark_reached() > util/vfio-helpers: Use error_setg in qemu_vfio_find_[fixed/temp]_iova > util/vfio-helpers: Simplify qemu_vfio_dma_map() returning directly > util/vfio-helpers: Let qemu_vfio_do_mapping() propagate Error > block/nvme: Only report VFIO error on failed retry > > include/qemu/vfio-helpers.h | 2 +- > block/nvme.c | 29 +++++++---- > util/vfio-helpers.c | 99 ++++++++++++++++++++----------------- > block/trace-events | 2 +- > 4 files changed, 76 insertions(+), 56 deletions(-) > > -- > 2.31.1 > > Thanks, applied to my block tree: https://gitlab.com/stefanha/qemu/commits/block Stefan
(Series fully reviewed) Hi, This series contains various patches sent last year with review comments addressed, few more cleanups, and a new patch which remove the spurious "VFIO_MAP_DMA failed: No space left on device" now poping up since commit 15a730e7a. (it is the expected behavior, which is why we retry the same call after flushing the DMA mappings). Since v2: - qemu_vfio_find_[fixed/temp]_iova retun bool (Klaus) - Add Klaus's R-b Since v1: - Addressed Klaus review comments (cleaner Error* handling) - Add Klaus's R-b Regards, Phil. Philippe Mathieu-Daudé (11): block/nvme: Use safer trace format string util/vfio-helpers: Let qemu_vfio_verify_mappings() use error_report() util/vfio-helpers: Replace qemu_mutex_lock() calls with QEMU_LOCK_GUARD util/vfio-helpers: Remove unreachable code in qemu_vfio_dma_map() block/nvme: Have nvme_create_queue_pair() report errors consistently util/vfio-helpers: Pass Error handle to qemu_vfio_dma_map() util/vfio-helpers: Extract qemu_vfio_water_mark_reached() util/vfio-helpers: Use error_setg in qemu_vfio_find_[fixed/temp]_iova util/vfio-helpers: Simplify qemu_vfio_dma_map() returning directly util/vfio-helpers: Let qemu_vfio_do_mapping() propagate Error block/nvme: Only report VFIO error on failed retry include/qemu/vfio-helpers.h | 2 +- block/nvme.c | 29 +++++++---- util/vfio-helpers.c | 99 ++++++++++++++++++++----------------- block/trace-events | 2 +- 4 files changed, 76 insertions(+), 56 deletions(-)