Re: [ 026/171 ] sfc: Only use TX push if a single descriptor is tobe written

From: Ben Hutchings
Date: Fri Apr 12 2013 - 18:05:39 EST

On Thu, 2013-04-11 at 22:15 +0100, Ben Hutchings wrote:
> Aside from #21-26 in this series, and the deadlock fix required on top
> of #24, there are several more fixes for sfc that I think are suitable
> for 3.6.11.y.
> These commits were cherry-picked for 3.4.38 and can also be
> cherry-picked cleanly on top of plus the 7 patches you already
> have:
> d5e8cc6c946e sfc: Really disable flow control while flushing
> bfeed902946a sfc: Convert firmware subtypes to native byte order in efx_mcdi_get_board_cfg()
> 9724a8504c87 sfc: Add parentheses around use of bitfield macro arguments
> 0a6e5008a9df sfc: Fix MCDI structure field lookup
> 450783747f42 sfc: Avoid generating over-length MC_CMD_FLUSH_RX_QUEUES request
> 525d9e824018 sfc: Work-around flush timeout when flushes have completed
> ef492f11efed sfc: Correctly initialise reset_method in siena_test_chip()
> ebf98e797b4e sfc: Fix timekeeping in efx_mcdi_poll()
> Please let me know whether you're prepared to include these in the
> current update. I can then run some automated tests with your selected
> set of patches applied.

The test suite found a regression which I'd forgotten about. It
was introduced in 3.6 by commit b7f514af7d6f 'sfc: Fix interface
statistics running backward' and fixed in 3.8 by commit 876be083b669
'sfc: Reset driver's MAC stats after MC reboot seen'.

That latter fix is, again, a clean cherry-pick onto 3.6.y. I don't
think I'm going to be able to re-test with this but it's sufficiently
low-risk that I'd be happy for you to add it anyway.


Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at