Re: [dpdk-dev] [PATCH v8 0/9] make rte_intr_handle internal

[David Marchand <david.marchand@redhat.com>]

On Mon, Oct 25, 2021 at 4:27 PM David Marchand
<david.marchand@redhat.com> wrote:
>
> Moving struct rte_intr_handle as an internal structure to
> avoid any ABI breakages in future. Since this structure defines
> some static arrays and changing respective macros breaks the ABI.
> Eg:
> Currently RTE_MAX_RXTX_INTR_VEC_ID imposes a limit of maximum 512
> MSI-X interrupts that can be defined for a PCI device, while PCI
> specification allows maximum 2048 MSI-X interrupts that can be used.
> If some PCI device requires more than 512 vectors, either change the
> RTE_MAX_RXTX_INTR_VEC_ID limit or dynamically allocate based on
> PCI device MSI-X size on probe time. Either way its an ABI breakage.
>
> Change already included in 21.11 ABI improvement spreadsheet (item 42):
> https://urldefense.proofpoint.com/v2/url?u=https-3A__docs.google.com_s
> preadsheets_d_1betlC000ua5SsSiJIcC54mCCCJnW6voH5Dqv9UxeyfE_edit-23gid-
> 3D0&d=DwICaQ&c=nKjWec2b6R0mOyPaz7xtfQ&r=5ESHPj7V-7JdkxT_Z_SU6RrS37ys4U
> XudBQ_rrS5LRo&m=7dl3OmXU7QHMmWYB6V1hYJtq1cUkjfhXUwze2Si_48c&s=lh6DEGhR
> Bg1shODpAy3RQk-H-0uQx5icRfUBf9dtCp4&e=
>
> This series makes struct rte_intr_handle totally opaque to the outside
> world by wrapping it inside a .c file and providing get set wrapper APIs
> to read or manipulate its fields.. Any changes to be made to any of the
> fields should be done via these get set APIs.
> Introduced a new eal_common_interrupts.c where all these APIs are defined
> and also hides struct rte_intr_handle definition.
>
> v1:
> * Fixed freebsd compilation failure
> * Fixed seg fault in case of memif
>
> v2:
> * Merged the prototype and implementation patch to 1.
> * Restricting allocation of single interrupt instance.
> * Removed base APIs, as they were exposing internally
> allocated memory information.
> * Fixed some memory leak issues.
> * Marked some library specific APIs as internal.
>
> v3:
> * Removed flag from instance alloc API, rather auto detect
> if memory should be allocated using glibc malloc APIs or
> rte_malloc*
> * Added APIs for get/set windows handle.
> * Defined macros for repeated checks.
>
> v4:
> * Rectified some typo in the APIs documentation.
> * Better names for some internal variables.
>
> v5:
> * Reverted back to passing flag to instance alloc API, as
> with auto detect some multiprocess issues existing in the
> library were causing tests failure.
> * Rebased to top of tree.
>
> v6:
> * renamed RTE_INTR_INSTANCE_F_UNSHARED as RTE_INTR_INSTANCE_F_PRIVATE,
> * changed API and removed need for alloc_flag content exposure
>   (see rte_intr_instance_dup() in patch 1 and 2),
> * exported all symbols for Windows,
> * fixed leak in unit tests in case of alloc failure,
> * split (previously) patch 4 into three patches
>   * (now) patch 4 only concerns alarm and (previously) patch 6 cleanup bits
>     are squashed in it,
>   * (now) patch 5 concerns other libraries updates,
>   * (now) patch 6 concerns drivers updates:
>     * instance allocation is moved to probing for auxiliary,
>     * there might be a bug for PCI drivers non requesting
>       RTE_PCI_DRV_NEED_MAPPING, but code is left as v5,
> * split (previously) patch 5 into three patches
>   * (now) patch 7 only hides structure, but keep it in a EAL private
>     header, this makes it possible to keep info in tracepoints,
>   * (now) patch 8 deals with VFIO/UIO internal fds merge,
>   * (now) patch 9 extends event list,
>
> v7:
> * fixed compilation on FreeBSD,
> * removed unused interrupt handle in FreeBSD alarm code,
> * fixed interrupt handle allocation for PCI drivers without
>   RTE_PCI_DRV_NEED_MAPPING,
>
> v8:
> * lowered logs level to DEBUG in sanity checks,
> * fixed corner case with vector list access,
>
> --
> David Marchand
>
> Harman Kalra (9):
>   interrupts: add allocator and accessors
>   interrupts: remove direct access to interrupt handle
>   test/interrupts: remove direct access to interrupt handle
>   alarm: remove direct access to interrupt handle
>   lib: remove direct access to interrupt handle
>   drivers: remove direct access to interrupt handle
>   interrupts: make interrupt handle structure opaque
>   interrupts: rename device specific file descriptor
>   interrupts: extend event list

Series applied, thanks.


-- 
David Marchand