mbox series

[v22,0/5] Microchip Polarfire FPGA manager

Message ID 20220623163248.3672-1-i.bornyakov@metrotek.ru (mailing list archive)
Headers show
Series Microchip Polarfire FPGA manager | expand

Message

Ivan Bornyakov June 23, 2022, 4:32 p.m. UTC 
Add support to the FPGA manager for programming Microchip Polarfire
FPGAs over slave SPI interface with .dat formatted bitsream image.

Changelog:
  v1 -> v2: fix printk formating
  v2 -> v3:
   * replace "microsemi" with "microchip"
   * replace prefix "microsemi_fpga_" with "mpf_"
   * more sensible .compatible and .name strings
   * remove unused defines STATUS_SPI_VIOLATION and STATUS_SPI_ERROR
  v3 -> v4: fix unused variable warning
    Put 'mpf_of_ids' definition under conditional compilation, so it
    would not hang unused if CONFIG_OF is not enabled.
  v4 -> v5:
   * prefix defines with MPF_
   * mdelay() -> usleep_range()
   * formatting fixes
   * add DT bindings doc
   * rework fpga_manager_ops.write() to fpga_manager_ops.write_sg()
     We can't parse image header in write_init() because image header
     size is not known beforehand. Thus parsing need to be done in
     fpga_manager_ops.write() callback, but fpga_manager_ops.write()
     also need to be reenterable. On the other hand,
     fpga_manager_ops.write_sg() is called once. Thus, rework usage of
     write() callback to write_sg().
  v5 -> v6: fix patch applying
     I forgot to clean up unrelated local changes which lead to error on
     patch 0001-fpga-microchip-spi-add-Microchip-MPF-FPGA-manager.patch
     applying on vanilla kernel.
  v6 -> v7: fix binding doc to pass dt_binding_check
  v7 -> v8: another fix for dt_binding_check warning
  v8 -> v9:
   * add another patch to support bitstream offset in FPGA image buffer
   * rework fpga_manager_ops.write_sg() back to fpga_manager_ops.write()
   * move image header parsing from write() to write_init()
  v9 -> v10:
   * add parse_header() callback to fpga_manager_ops
   * adjust fpga_mgr_write_init[_buf|_sg]() for parse_header() usage
   * implement parse_header() in microchip-spi driver
  v10 -> v11: include missing unaligned.h to microchip-spi
     fix error: implicit declaration of function 'get_unaligned_le[16|32]'
  v11 -> v12:
   * microchip-spi: double read hw status, ignore first read, because it
     can be unreliable.
   * microchip-spi: remove sleep between status readings in
     poll_status_not_busy() to save a few seconds. Status is polled on
     every 16 byte writes - that is quite often, therefore
     usleep_range() accumulate to a considerable number of seconds.
  v12 -> v13:
   * fpga-mgr: separate fpga_mgr_parse_header_buf() from
     fpga_mgr_write_init_buf()
   * fpga-mgr: introduce FPGA_MGR_STATE_PARSE_HEADER and
     FPGA_MGR_STATE_PARSE_HEADER_ERR fpga_mgr_states
   * fpga-mgr: rename fpga_mgr_write_init_sg() to fpga_mgr_prepare_sg()
     and rework with respect to a new fpga_mgr_parse_header_buf()
   * fpga-mgr: rework write accounting in fpga_mgr_buf_load_sg() for
     better clarity
   * microchip-spi: rename MPF_STATUS_POLL_TIMEOUT to
     MPF_STATUS_POLL_RETRIES
   * microchip-spi: add comment about status reading quirk to
     mpf_read_status()
   * microchip-spi: rename poll_status_not_busy() to mpf_poll_status()
     and add comment.
   * microchip-spi: make if statement in mpf_poll_status() easier to
     read.
  v13 -> v14:
   * fpga-mgr: improvements from Xu Yilun in
      - fpga_mgr_parse_header_buf()
      - fpga_mgr_write_init_buf()
      - fpga_mgr_prepare_sg()
      - fpga_mgr_buf_load_sg()
   * fpga-mgr: add check for -EAGAIN from fpga_mgr_parse_header_buf()
     when called from fpga_mgr_buf_load_mapped()
   * microchip-spi: remove excessive cs_change from second spi_transfer
     in mpf_read_status()
   * microchip-spi: change type of components_size_start,
     bitstream_start, i from size_t to u32 in mpf_ops_parse_header()
  v14 -> v15: eliminate memcpy() in mpf_ops_write()
    Eliminate excessive memcpy() in mpf_ops_write() by using
    spi_sync_transfer() instead of spi_write().
  v15 -> v16:
   * microchip-spi: change back components_size_start and
     bitstream_start variables types to size_t, i - to u16 in
     mpf_ops_parse_header()
   * fpga-mgr: rename fpga_parse_header_buf() to
     fpga_parse_header_mapped(). It serves only mapped FPGA image now,
     adjust it accordingly.
   * fpga-mgr: separate fpga_mgr_parse_header_sg_first() and
     fpga_mgr_parse_header_sg() from fpga_mgr_prepare_sg()
  v16 -> v17:
   * fpga-mgr: return size of allocated header from
     fpga_mgr_parse_header_sg(), add `char **ret_buf` to function args
     to save pointer to allocated header. This allow us to call
     fpga_mgr_write_init_buf() with exact size of allocated header.
   * document parse_header() callback in fpga-mgr.rst
  v17 -> v18:
   * fpga-mgr: change back fpga_mgr_parse_header_sg() to return
     allocated buffer but set buffer size into output parameter
   * fpga-mgr: check returned pointer from krealloc for ZERO_OR_NULL_PTR
     in fpga_mgr_paese_header_sg() as krealloc may return ZERO_SIZE_PTR.
   * fpga-mgr: in fpga_mgr_prepare_sg() return fpga_mgr_write_init() on
     fast path only when both initial_header_size and parse_header() are
     not defined.
   * docs: fpga-mgr: a few rewords from Xu Yilun
  v18 -> v19:
   * microchip-spi: split multiple assignments on a single line in
     functions mpf_read_status() and mpf_ops_parse_header()
   * fpga-mgr: add braces {} around "else if" arm in
     fpga_mgr_prepare_sg()
   * fpga-mgr: don't reuse krealloc() arg in fpga_mgr_parse_header_sg().
     If krealloc() returns NULL, it doesn't free the original.
  v19 -> v20:
   * fpga-mgr: initialize info->header_size with
     mops->initial_header_size at fpga_mgr_load().
   * fpga-mgr: add mops->skip_header boolean flag. Adjust skipping
     header before write to check against skip_header flag instead of
     mere presence of info->header_size.
   * fpga-mgr: split check for ZERO_OR_NULL_PTR() after realloc() in
     fpga_mgr_parse_header_sg() function to check against zero header +
     check against NULL returned from realloc().
   * docs: fpga-mgr: adjust for skip_header flag.
   * microchip-spi: add skip_header to mpf_ops.
  v20 -> v21:
   * fpga-mgr: in function fpga_mgr_parse_header_sg(), after
     reallocation of a buffer for a new header, copy only exceeding
     part, since realloc preserve old buffer content.
   * fpga-mgr: in function fpga_mgr_buf_load_sg() instead of manually
     counting fragments with header that should be skipped, use
     sg_miter_skip().
   * fpga-mgr: add tag "Suggested-by: Xu Yilun <yilun.xu@intel.com>"
     since all changes in fpga-mgr was his ideas.
   * MAINTAINERS: add entry for Microchip PolarFire FPGA drivers as
     Conor Dooley suggested.
  v21 -> v22:
   * fpga-mgr: in function fpga_mgr_buf_load_sg() drop excessive
     sg_miter_stop() when sg_miter_skip() reaches end of sg_mapping_iter

Ivan Bornyakov (5):
  fpga: fpga-mgr: support bitstream offset in image buffer
  docs: fpga: mgr: document parse_header() callback
  fpga: microchip-spi: add Microchip MPF FPGA manager
  dt-bindings: fpga: add binding doc for microchip-spi fpga mgr
  MAINTAINERS: add Microchip PolarFire FPGA drivers entry

 .../fpga/microchip,mpf-spi-fpga-mgr.yaml      |  44 ++
 Documentation/driver-api/fpga/fpga-mgr.rst    |  27 +-
 MAINTAINERS                                   |   8 +
 drivers/fpga/Kconfig                          |   8 +
 drivers/fpga/Makefile                         |   1 +
 drivers/fpga/fpga-mgr.c                       | 223 ++++++++--
 drivers/fpga/microchip-spi.c                  | 398 ++++++++++++++++++
 include/linux/fpga/fpga-mgr.h                 |  24 +-
 8 files changed, 700 insertions(+), 33 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/fpga/microchip,mpf-spi-fpga-mgr.yaml
 create mode 100644 drivers/fpga/microchip-spi.c

Comments

Xu Yilun June 24, 2022, 4:16 a.m. UTC | #1 
On Thu, Jun 23, 2022 at 07:32:43PM +0300, Ivan Bornyakov wrote:
> Add support to the FPGA manager for programming Microchip Polarfire
> FPGAs over slave SPI interface with .dat formatted bitsream image.
> 
> Changelog:
>   v1 -> v2: fix printk formating
>   v2 -> v3:
>    * replace "microsemi" with "microchip"
>    * replace prefix "microsemi_fpga_" with "mpf_"
>    * more sensible .compatible and .name strings
>    * remove unused defines STATUS_SPI_VIOLATION and STATUS_SPI_ERROR
>   v3 -> v4: fix unused variable warning
>     Put 'mpf_of_ids' definition under conditional compilation, so it
>     would not hang unused if CONFIG_OF is not enabled.
>   v4 -> v5:
>    * prefix defines with MPF_
>    * mdelay() -> usleep_range()
>    * formatting fixes
>    * add DT bindings doc
>    * rework fpga_manager_ops.write() to fpga_manager_ops.write_sg()
>      We can't parse image header in write_init() because image header
>      size is not known beforehand. Thus parsing need to be done in
>      fpga_manager_ops.write() callback, but fpga_manager_ops.write()
>      also need to be reenterable. On the other hand,
>      fpga_manager_ops.write_sg() is called once. Thus, rework usage of
>      write() callback to write_sg().
>   v5 -> v6: fix patch applying
>      I forgot to clean up unrelated local changes which lead to error on
>      patch 0001-fpga-microchip-spi-add-Microchip-MPF-FPGA-manager.patch
>      applying on vanilla kernel.
>   v6 -> v7: fix binding doc to pass dt_binding_check
>   v7 -> v8: another fix for dt_binding_check warning
>   v8 -> v9:
>    * add another patch to support bitstream offset in FPGA image buffer
>    * rework fpga_manager_ops.write_sg() back to fpga_manager_ops.write()
>    * move image header parsing from write() to write_init()
>   v9 -> v10:
>    * add parse_header() callback to fpga_manager_ops
>    * adjust fpga_mgr_write_init[_buf|_sg]() for parse_header() usage
>    * implement parse_header() in microchip-spi driver
>   v10 -> v11: include missing unaligned.h to microchip-spi
>      fix error: implicit declaration of function 'get_unaligned_le[16|32]'
>   v11 -> v12:
>    * microchip-spi: double read hw status, ignore first read, because it
>      can be unreliable.
>    * microchip-spi: remove sleep between status readings in
>      poll_status_not_busy() to save a few seconds. Status is polled on
>      every 16 byte writes - that is quite often, therefore
>      usleep_range() accumulate to a considerable number of seconds.
>   v12 -> v13:
>    * fpga-mgr: separate fpga_mgr_parse_header_buf() from
>      fpga_mgr_write_init_buf()
>    * fpga-mgr: introduce FPGA_MGR_STATE_PARSE_HEADER and
>      FPGA_MGR_STATE_PARSE_HEADER_ERR fpga_mgr_states
>    * fpga-mgr: rename fpga_mgr_write_init_sg() to fpga_mgr_prepare_sg()
>      and rework with respect to a new fpga_mgr_parse_header_buf()
>    * fpga-mgr: rework write accounting in fpga_mgr_buf_load_sg() for
>      better clarity
>    * microchip-spi: rename MPF_STATUS_POLL_TIMEOUT to
>      MPF_STATUS_POLL_RETRIES
>    * microchip-spi: add comment about status reading quirk to
>      mpf_read_status()
>    * microchip-spi: rename poll_status_not_busy() to mpf_poll_status()
>      and add comment.
>    * microchip-spi: make if statement in mpf_poll_status() easier to
>      read.
>   v13 -> v14:
>    * fpga-mgr: improvements from Xu Yilun in
>       - fpga_mgr_parse_header_buf()
>       - fpga_mgr_write_init_buf()
>       - fpga_mgr_prepare_sg()
>       - fpga_mgr_buf_load_sg()
>    * fpga-mgr: add check for -EAGAIN from fpga_mgr_parse_header_buf()
>      when called from fpga_mgr_buf_load_mapped()
>    * microchip-spi: remove excessive cs_change from second spi_transfer
>      in mpf_read_status()
>    * microchip-spi: change type of components_size_start,
>      bitstream_start, i from size_t to u32 in mpf_ops_parse_header()
>   v14 -> v15: eliminate memcpy() in mpf_ops_write()
>     Eliminate excessive memcpy() in mpf_ops_write() by using
>     spi_sync_transfer() instead of spi_write().
>   v15 -> v16:
>    * microchip-spi: change back components_size_start and
>      bitstream_start variables types to size_t, i - to u16 in
>      mpf_ops_parse_header()
>    * fpga-mgr: rename fpga_parse_header_buf() to
>      fpga_parse_header_mapped(). It serves only mapped FPGA image now,
>      adjust it accordingly.
>    * fpga-mgr: separate fpga_mgr_parse_header_sg_first() and
>      fpga_mgr_parse_header_sg() from fpga_mgr_prepare_sg()
>   v16 -> v17:
>    * fpga-mgr: return size of allocated header from
>      fpga_mgr_parse_header_sg(), add `char **ret_buf` to function args
>      to save pointer to allocated header. This allow us to call
>      fpga_mgr_write_init_buf() with exact size of allocated header.
>    * document parse_header() callback in fpga-mgr.rst
>   v17 -> v18:
>    * fpga-mgr: change back fpga_mgr_parse_header_sg() to return
>      allocated buffer but set buffer size into output parameter
>    * fpga-mgr: check returned pointer from krealloc for ZERO_OR_NULL_PTR
>      in fpga_mgr_paese_header_sg() as krealloc may return ZERO_SIZE_PTR.
>    * fpga-mgr: in fpga_mgr_prepare_sg() return fpga_mgr_write_init() on
>      fast path only when both initial_header_size and parse_header() are
>      not defined.
>    * docs: fpga-mgr: a few rewords from Xu Yilun
>   v18 -> v19:
>    * microchip-spi: split multiple assignments on a single line in
>      functions mpf_read_status() and mpf_ops_parse_header()
>    * fpga-mgr: add braces {} around "else if" arm in
>      fpga_mgr_prepare_sg()
>    * fpga-mgr: don't reuse krealloc() arg in fpga_mgr_parse_header_sg().
>      If krealloc() returns NULL, it doesn't free the original.
>   v19 -> v20:
>    * fpga-mgr: initialize info->header_size with
>      mops->initial_header_size at fpga_mgr_load().
>    * fpga-mgr: add mops->skip_header boolean flag. Adjust skipping
>      header before write to check against skip_header flag instead of
>      mere presence of info->header_size.
>    * fpga-mgr: split check for ZERO_OR_NULL_PTR() after realloc() in
>      fpga_mgr_parse_header_sg() function to check against zero header +
>      check against NULL returned from realloc().
>    * docs: fpga-mgr: adjust for skip_header flag.
>    * microchip-spi: add skip_header to mpf_ops.
>   v20 -> v21:
>    * fpga-mgr: in function fpga_mgr_parse_header_sg(), after
>      reallocation of a buffer for a new header, copy only exceeding
>      part, since realloc preserve old buffer content.
>    * fpga-mgr: in function fpga_mgr_buf_load_sg() instead of manually
>      counting fragments with header that should be skipped, use
>      sg_miter_skip().
>    * fpga-mgr: add tag "Suggested-by: Xu Yilun <yilun.xu@intel.com>"
>      since all changes in fpga-mgr was his ideas.
>    * MAINTAINERS: add entry for Microchip PolarFire FPGA drivers as
>      Conor Dooley suggested.
>   v21 -> v22:
>    * fpga-mgr: in function fpga_mgr_buf_load_sg() drop excessive
>      sg_miter_stop() when sg_miter_skip() reaches end of sg_mapping_iter
> 
> Ivan Bornyakov (5):
>   fpga: fpga-mgr: support bitstream offset in image buffer
>   docs: fpga: mgr: document parse_header() callback
>   fpga: microchip-spi: add Microchip MPF FPGA manager
>   dt-bindings: fpga: add binding doc for microchip-spi fpga mgr
>   MAINTAINERS: add Microchip PolarFire FPGA drivers entry
> 
>  .../fpga/microchip,mpf-spi-fpga-mgr.yaml      |  44 ++
>  Documentation/driver-api/fpga/fpga-mgr.rst    |  27 +-
>  MAINTAINERS                                   |   8 +
>  drivers/fpga/Kconfig                          |   8 +
>  drivers/fpga/Makefile                         |   1 +
>  drivers/fpga/fpga-mgr.c                       | 223 ++++++++--
>  drivers/fpga/microchip-spi.c                  | 398 ++++++++++++++++++
>  include/linux/fpga/fpga-mgr.h                 |  24 +-
>  8 files changed, 700 insertions(+), 33 deletions(-)
>  create mode 100644 Documentation/devicetree/bindings/fpga/microchip,mpf-spi-fpga-mgr.yaml
>  create mode 100644 drivers/fpga/microchip-spi.c

Applied the series to for-next

Thanks for the improvement to FPGA core.
Yilun

> 
> -- 
> 2.36.1
>