linux.git/drivers/net/ethernet/amd, branch v6.7

amd-xgbe: propagate the correct speed and duplex status

2023-11-23T12:47:23+00:00

xgbe_get_link_ksettings() does not propagate correct speed and duplex
information to ethtool during cable unplug. Due to which ethtool reports
incorrect values for speed and duplex.

Address this by propagating correct information.

Fixes: 7c12aa08779c ("amd-xgbe: Move the PHY support into amd-xgbe")
Acked-by: Shyam Sundar S K 
Signed-off-by: Raju Rangoju 
Reviewed-by: Wojciech Drewek 
Signed-off-by: Paolo Abeni

amd-xgbe: handle the corner-case during tx completion

2023-11-23T12:47:23+00:00

The existing implementation uses software logic to accumulate tx
completions until the specified time (1ms) is met and then poll them.
However, there exists a tiny gap which leads to a race between
resetting and checking the tx_activate flag. Due to this the tx
completions are not reported to upper layer and tx queue timeout
kicks-in restarting the device.

To address this, introduce a tx cleanup mechanism as part of the
periodic maintenance process.

Fixes: c5aa9e3b8156 ("amd-xgbe: Initial AMD 10GbE platform driver")
Acked-by: Shyam Sundar S K 
Signed-off-by: Raju Rangoju 
Reviewed-by: Wojciech Drewek 
Signed-off-by: Paolo Abeni

amd-xgbe: handle corner-case during sfp hotplug

2023-11-23T12:47:23+00:00

Force the mode change for SFI in Fixed PHY configurations. Fixed PHY
configurations needs PLL to be enabled while doing mode set. When the
SFP module isn't connected during boot, driver assumes AN is ON and
attempts auto-negotiation. However, if the connected SFP comes up in
Fixed PHY configuration the link will not come up as PLL isn't enabled
while the initial mode set command is issued. So, force the mode change
for SFI in Fixed PHY configuration to fix link issues.

Fixes: e57f7a3feaef ("amd-xgbe: Prepare for working with more than one type of phy")
Acked-by: Shyam Sundar S K 
Signed-off-by: Raju Rangoju 
Reviewed-by: Wojciech Drewek 
Signed-off-by: Paolo Abeni

pds_core: fix up some format-truncation complaints

2023-11-15T03:52:09+00:00

Our friendly kernel test robot pointed out a couple of potential
string truncation issues.  None of which were we worried about,
but can be relatively easily fixed to quiet the complaints.

Reported-by: kernel test robot 
Closes: https://lore.kernel.org/oe-kbuild-all/202310211736.66syyDpp-lkp@intel.com/
Fixes: 45d76f492938 ("pds_core: set up device and adminq")
Signed-off-by: Shannon Nelson 
Link: https://lore.kernel.org/r/20231113183257.71110-3-shannon.nelson@amd.com
Signed-off-by: Jakub Kicinski

pds_core: use correct index to mask irq

2023-11-15T03:52:09+00:00

Use the qcq's interrupt index, not the irq number, to mask
the interrupt.  Since the irq number can be out of range from
the number of possible interrupts, we can end up accessing
and potentially scribbling on out-of-range and/or unmapped
memory, making the kernel angry.

    [ 3116.039364] BUG: unable to handle page fault for address: ffffbeea1c3edf84
    [ 3116.047059] #PF: supervisor write access in kernel mode
    [ 3116.052895] #PF: error_code(0x0002) - not-present page
    [ 3116.058636] PGD 100000067 P4D 100000067 PUD 1001f2067 PMD 10f82e067 PTE 0
    [ 3116.066221] Oops: 0002 [#1] SMP NOPTI
    [ 3116.092948] RIP: 0010:iowrite32+0x9/0x76
    [ 3116.190452] Call Trace:
    [ 3116.193185]  
    [ 3116.195430]  ? show_trace_log_lvl+0x1d6/0x2f9
    [ 3116.200298]  ? show_trace_log_lvl+0x1d6/0x2f9
    [ 3116.205166]  ? pdsc_adminq_isr+0x43/0x55 [pds_core]
    [ 3116.210618]  ? __die_body.cold+0x8/0xa
    [ 3116.214806]  ? page_fault_oops+0x16d/0x1ac
    [ 3116.219382]  ? exc_page_fault+0xbe/0x13b
    [ 3116.223764]  ? asm_exc_page_fault+0x22/0x27
    [ 3116.228440]  ? iowrite32+0x9/0x76
    [ 3116.232143]  pdsc_adminq_isr+0x43/0x55 [pds_core]
    [ 3116.237627]  __handle_irq_event_percpu+0x3a/0x184
    [ 3116.243088]  handle_irq_event+0x57/0xb0
    [ 3116.247575]  handle_edge_irq+0x87/0x225
    [ 3116.252062]  __common_interrupt+0x3e/0xbc
    [ 3116.256740]  common_interrupt+0x7b/0x98
    [ 3116.261216]  
    [ 3116.263745]  
    [ 3116.266268]  asm_common_interrupt+0x22/0x27

Reported-by: Joao Martins 
Fixes: 01ba61b55b20 ("pds_core: Add adminq processing and commands")
Signed-off-by: Shannon Nelson 
Link: https://lore.kernel.org/r/20231113183257.71110-2-shannon.nelson@amd.com
Signed-off-by: Jakub Kicinski

amd/pds_core: core: No need for Null pointer check before kfree

2023-10-25T09:34:28+00:00

kfree()/vfree() internally perform NULL check on the
pointer handed to it and take no action if it indeed is
NULL. Hence there is no need for a pre-check of the memory
pointer before handing it to kfree()/vfree().

Issue reported by ifnullfree.cocci Coccinelle semantic
patch script.

Signed-off-by: Bragatheswaran Manickavel 
Reviewed-by: Shannon Nelson 
Reviewed-by: Wojciech Drewek 
Signed-off-by: David S. Miller

pds_core: add an error code check in pdsc_dl_info_get

2023-10-21T00:39:32+00:00

check the value of 'ret' after call 'devlink_info_version_stored_put'.

Signed-off-by: Su Hui 
Reviewed-by: Shannon Nelson 
Link: https://lore.kernel.org/r/20231019083351.1526484-1-suhui@nfschina.com
Signed-off-by: Jakub Kicinski

pds_core: devlink health: use retained error fmsg API

2023-10-20T10:34:49+00:00

Drop unneeded error checking.

devlink_fmsg_*() family of functions is now retaining errors,
so there is no need to check for them after each call.

Reviewed-by: Jesse Brandeburg 
Reviewed-by: Jiri Pirko 
Reviewed-by: Shannon Nelson 
Signed-off-by: Przemek Kitszel 
Reviewed-by: Simon Horman 
Signed-off-by: David S. Miller

net: ethernet: Use device_get_match_data()

2023-10-13T09:04:53+00:00

Use preferred device_get_match_data() instead of of_match_device() to
get the driver match data. With this, adjust the includes to explicitly
include the correct headers.

Signed-off-by: Rob Herring 
Signed-off-by: David S. Miller

net: ethernet: amd: Convert to platform remove callback returning void

2023-09-20T08:06:37+00:00

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new() which already returns void. Eventually after all drivers
are converted, .remove_new() is renamed to .remove().

Trivially convert these drivers from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König 
Signed-off-by: David S. Miller