<feed xmlns='http://www.w3.org/2005/Atom'>
<title>kernel/git/next/linux-next.git/drivers/ntb, branch akpm</title>
<subtitle>The linux-next integration testing tree</subtitle>
<id>https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/atom?h=akpm</id>
<link rel='self' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/atom?h=akpm'/>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/'/>
<updated>2022-06-27T16:30:06+00:00</updated>
<entry>
<title>drivers/ntb/test: avoid 64-bit modulus operation</title>
<updated>2022-06-27T16:30:06+00:00</updated>
<author>
<name>Max Filippov</name>
<email>jcmvbkbc@gmail.com</email>
</author>
<published>2022-06-27T15:57:10+00:00</published>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=9297e01b038b5932590f4bebda0361d210432ee2'/>
<id>urn:sha1:9297e01b038b5932590f4bebda0361d210432ee2</id>
<content type='text'>
Redefine RESCHEDULE_RATIO to a closest power of 2 so that the following
code in the perf_run_latency

	/* Avoid processor soft lock-ups */
	if (!(pthr-&gt;tries % RESCHEDULE_RATIO))
		schedule();

doesn't do 64-bit modulus operation.
This fixes the following build failures on 32-bit architectures visible
in linux-next:

  ERROR: modpost: "__umoddi3" [drivers/ntb/test/ntb_perf.ko] undefined!

Fixes: dc150dfb081f ("ntb_perf: extend with burst latency measurement")
Signed-off-by: Max Filippov &lt;jcmvbkbc@gmail.com&gt;
Signed-off-by: Jon Mason &lt;jdmason@kudzu.us&gt;
</content>
</entry>
<entry>
<title>NTB: epf: Added more flexible memory map method</title>
<updated>2022-06-20T13:21:45+00:00</updated>
<author>
<name>Frank Li</name>
<email>Frank.Li@nxp.com</email>
</author>
<published>2022-02-15T05:38:42+00:00</published>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=f1ac4e991dc996629eba3a93e54599fd39f3cb47'/>
<id>urn:sha1:f1ac4e991dc996629eba3a93e54599fd39f3cb47</id>
<content type='text'>
Supported below memory map method

bar 0: config and spad data
bar 2: door bell
bar 4: memory map windows

Signed-off-by: Frank Li &lt;Frank.Li@nxp.com&gt;
Signed-off-by: Jon Mason &lt;jdmason@kudzu.us&gt;
</content>
</entry>
<entry>
<title>ntb_perf: extend with doorbell latency measurement</title>
<updated>2022-06-20T13:16:40+00:00</updated>
<author>
<name>Alexander Fomichev</name>
<email>a.fomichev@yadro.com</email>
</author>
<published>2022-05-13T19:37:04+00:00</published>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=d3b4b9c2142ca288f81a373841989f592f138934'/>
<id>urn:sha1:d3b4b9c2142ca288f81a373841989f592f138934</id>
<content type='text'>
Doorbell latency is a delay between start to ring an NTB doorbell
and receiving the confirmation. The remote system needs to be run in server
mode beforehand. Then the server waits for a doorbell event and immediately
rings self doorbell to confirm.

Thanks-to: Guo Zhengkui &lt;guozhengkui@vivo.com&gt;
Signed-off-by: Alexander Fomichev &lt;a.fomichev@yadro.com&gt;
Reviewed-by: Dave Jiang &lt;dave.jiang@intel.com&gt;
Signed-off-by: Jon Mason &lt;jdmason@kudzu.us&gt;
</content>
</entry>
<entry>
<title>ntb_perf: extend with poll latency measurement</title>
<updated>2022-06-20T13:16:36+00:00</updated>
<author>
<name>Alexander Fomichev</name>
<email>a.fomichev@yadro.com</email>
</author>
<published>2022-05-13T19:37:03+00:00</published>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=9ace38ad4e8adffeb5ac292b8ceb74fb4f564452'/>
<id>urn:sha1:9ace38ad4e8adffeb5ac292b8ceb74fb4f564452</id>
<content type='text'>
Poll latency is a delay between start to send 1 byte to the
remote system and receiving the confirmation. The remote system needs to
be run in server mode beforehand. Then the server polls the input buffer
and on receiving data immediately sends the confirmation back.

Signed-off-by: Alexander Fomichev &lt;a.fomichev@yadro.com&gt;
Reviewed-by: Dave Jiang &lt;dave.jiang@intel.com&gt;
Signed-off-by: Jon Mason &lt;jdmason@kudzu.us&gt;
</content>
</entry>
<entry>
<title>ntb_perf: extend with burst latency measurement</title>
<updated>2022-06-20T13:16:26+00:00</updated>
<author>
<name>Alexander Fomichev</name>
<email>a.fomichev@yadro.com</email>
</author>
<published>2022-05-13T19:37:02+00:00</published>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=dc150dfb081fcd17f8ecfa07d94813e4c445d808'/>
<id>urn:sha1:dc150dfb081fcd17f8ecfa07d94813e4c445d808</id>
<content type='text'>
Burst latency is a delay between start to send 1 byte to the
remote system and hardware readiness to send another byte. The
measurement performed within bandwidth test procedure. The DMA Engine is
off. Data integrity is not checked. This mode can be disabled by
'perf_latency=N' module parameter.

Signed-off-by: Alexander Fomichev &lt;a.fomichev@yadro.com&gt;
Reviewed-by: Dave Jiang &lt;dave.jiang@intel.com&gt;
Signed-off-by: Jon Mason &lt;jdmason@kudzu.us&gt;
</content>
</entry>
<entry>
<title>ntb: intel: fix port config status offset for SPR</title>
<updated>2022-01-28T15:19:16+00:00</updated>
<author>
<name>Dave Jiang</name>
<email>dave.jiang@intel.com</email>
</author>
<published>2022-01-27T20:31:12+00:00</published>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=d5081bf5dcfb1cb83fb538708b0ac07a10a79cc4'/>
<id>urn:sha1:d5081bf5dcfb1cb83fb538708b0ac07a10a79cc4</id>
<content type='text'>
The field offset for port configuration status on SPR has been changed to
bit 14 from ICX where it resides at bit 12. By chance link status detection
continued to work on SPR. This is due to bit 12 being a configuration bit
which is in sync with the status bit. Fix this by checking for a SPR device
and checking correct status bit.

Fixes: 26bfe3d0b227 ("ntb: intel: Add Icelake (gen4) support for Intel NTB")
Tested-by: Jerry Dai &lt;jerry.dai@intel.com&gt;
Signed-off-by: Dave Jiang &lt;dave.jiang@intel.com&gt;
Signed-off-by: Jon Mason &lt;jdmason@kudzu.us&gt;
</content>
</entry>
<entry>
<title>NTB/msi: Use struct_size() helper in devm_kzalloc()</title>
<updated>2022-01-23T21:15:15+00:00</updated>
<author>
<name>Gustavo A. R. Silva</name>
<email>gustavoars@kernel.org</email>
</author>
<published>2022-01-20T23:02:47+00:00</published>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=305325688ff924c52a6e12f92235a89536e022cf'/>
<id>urn:sha1:305325688ff924c52a6e12f92235a89536e022cf</id>
<content type='text'>
Make use of the struct_size() helper instead of an open-coded version,
in order to avoid any potential type mistakes or integer overflows that,
in the worst scenario, could lead to heap overflows.

Also, address the following sparse warnings:
drivers/ntb/msi.c:46:23: warning: using sizeof on a flexible structure

Link: https://github.com/KSPP/linux/issues/174
Signed-off-by: Gustavo A. R. Silva &lt;gustavoars@kernel.org&gt;
Reviewed-by: Kees Cook &lt;keescook@chromium.org&gt;
Reviewed-by: Logan Gunthorpe &lt;logang@deltatee.com&gt;
Signed-off-by: Jon Mason &lt;jdmason@kudzu.us&gt;
</content>
</entry>
<entry>
<title>Merge tag 'ntb-5.17' of git://github.com/jonmason/ntb</title>
<updated>2022-01-17T06:14:18+00:00</updated>
<author>
<name>Linus Torvalds</name>
<email>torvalds@linux-foundation.org</email>
</author>
<published>2022-01-17T06:14:18+00:00</published>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=96000bc956628dc0cc608696d5385ffc19e65a76'/>
<id>urn:sha1:96000bc956628dc0cc608696d5385ffc19e65a76</id>
<content type='text'>
Pull NTB updates from Jon Mason:
 "New AMD PCI ID for NTB, and a number of bug fixes for ntb_hw_switchtec
  for Linux v5.17"

* tag 'ntb-5.17' of git://github.com/jonmason/ntb:
  ntb_hw_switchtec: Fix a minor issue in config_req_id_table()
  ntb_hw_switchtec: Remove code for disabling ID protection
  ntb_hw_switchtec: Update the way of getting VEP instance ID
  ntb_hw_switchtec: AND with the part_map for a valid tpart_vec
  ntb_hw_switchtec: Fix bug with more than 32 partitions
  ntb_hw_switchtec: Fix pff ioread to read into mmio_part_cfg_all
  ntb_hw_switchtec: fix the spelling of "its"
  NTB/msi: Fix ntbm_msi_request_threaded_irq() kernel-doc comment
  ntb_hw_amd: Add NTB PCI ID for new gen CPU
</content>
</entry>
<entry>
<title>ntb_hw_switchtec: Fix a minor issue in config_req_id_table()</title>
<updated>2022-01-11T20:38:59+00:00</updated>
<author>
<name>Kelvin Cao</name>
<email>kelvin.cao@microchip.com</email>
</author>
<published>2021-12-24T01:23:34+00:00</published>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=8cd778650ae223cd306588042b55d0290ef81037'/>
<id>urn:sha1:8cd778650ae223cd306588042b55d0290ef81037</id>
<content type='text'>
The req_id_table_size field is 16-bit wide, use ioread16() to read the
value.

Signed-off-by: Kelvin Cao &lt;kelvin.cao@microchip.com&gt;
Signed-off-by: Jon Mason &lt;jdmason@kudzu.us&gt;
</content>
</entry>
<entry>
<title>ntb_hw_switchtec: Remove code for disabling ID protection</title>
<updated>2022-01-11T20:38:59+00:00</updated>
<author>
<name>Kelvin Cao</name>
<email>kelvin.cao@microchip.com</email>
</author>
<published>2021-12-24T01:23:33+00:00</published>
<link rel='alternate' type='text/html' href='https://git.rulkc.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=1d3cfc2835c1754d19a743dc346a9e58cf0c07c0'/>
<id>urn:sha1:1d3cfc2835c1754d19a743dc346a9e58cf0c07c0</id>
<content type='text'>
ID protection is a firmware setting for NT window access control. With
it enabled, only the posted requests with requester IDs in the requester
ID table will be allowed to access the NT windows. Otherwise all posted
requests are allowed. Normally user will configure it statically via the
Switchtec config file, and it will take effect when the firmware boots
up. The driver can also toggle the ID protection setting dynamically,
which will overwrite the static setting in the Switchtec config file as
a side effect.

Currently, the driver disables the ID protection. However, it's not
necessary to disable the ID protection at the driver level as the driver
has already configured the proper requester IDs in the requester ID
table to allow the corresponding posted requests to hit the NT windows.
Remove the code that disables the ID protection to make the static
setting prevail.

Note: ID protection is not applicable to non-posted requests.

Signed-off-by: Kelvin Cao &lt;kelvin.cao@microchip.com&gt;
Signed-off-by: Jon Mason &lt;jdmason@kudzu.us&gt;
</content>
</entry>
</feed>
