| Message ID | 20191022203113.30015-3-snelson@pensando.io |
|---|---|
| State | Changes Requested |
| Delegated to: | David Miller |
| Headers | show |
| Series | ionic updates | expand |
On Tue, Oct 22, 2019 at 01:31:09PM -0700, Shannon Nelson wrote: > Fix the initial interrupt coalesce usec-to-hw setting > to actually be usec-to-hw. > > Fixes: 780eded34ccc ("ionic: report users coalesce request") > Signed-off-by: Shannon Nelson <snelson@pensando.io> Hi Shannon How bad is this? Should it be backported? If so, you should send it as a separate patch for net. Andrew
On 10/22/19 6:22 PM, Andrew Lunn wrote: > On Tue, Oct 22, 2019 at 01:31:09PM -0700, Shannon Nelson wrote: >> Fix the initial interrupt coalesce usec-to-hw setting >> to actually be usec-to-hw. >> >> Fixes: 780eded34ccc ("ionic: report users coalesce request") >> Signed-off-by: Shannon Nelson <snelson@pensando.io> > Hi Shannon > > How bad is this? Should it be backported? If so, you should send it as > a separate patch for net. > > Andrew Well, it doesn't "break" anything, but it does end up initializing the coalesce value to 0, aka turns it off. The easy work-around is for the user to set a new value, then all is well. However, since v5.4 is intended as an LTS, this probably should get promoted into net. If there aren't many other comments, I'll likely turn this around tomorrow afternoon. sln
On Tue, 22 Oct 2019 19:09:40 -0700, Shannon Nelson wrote: > On 10/22/19 6:22 PM, Andrew Lunn wrote: > > On Tue, Oct 22, 2019 at 01:31:09PM -0700, Shannon Nelson wrote: > >> Fix the initial interrupt coalesce usec-to-hw setting > >> to actually be usec-to-hw. > >> > >> Fixes: 780eded34ccc ("ionic: report users coalesce request") > >> Signed-off-by: Shannon Nelson <snelson@pensando.io> > > Hi Shannon > > > > How bad is this? Should it be backported? If so, you should send it as > > a separate patch for net. > > Well, it doesn't "break" anything, but it does end up initializing the > coalesce value to 0, aka turns it off. The easy work-around is for the > user to set a new value, then all is well. However, since v5.4 is > intended as an LTS, this probably should get promoted into net. If > there aren't many other comments, I'll likely turn this around tomorrow > afternoon. I think the commit in question only landed in net-next after the merge window. Perhaps the larger problem with the coalescing could have been fixed in net, but now I think it may be too late, no?
diff --git a/drivers/net/ethernet/pensando/ionic/ionic_lif.c b/drivers/net/ethernet/pensando/ionic/ionic_lif.c index 559b96ae48f5..cf64dea53f82 100644 --- a/drivers/net/ethernet/pensando/ionic/ionic_lif.c +++ b/drivers/net/ethernet/pensando/ionic/ionic_lif.c @@ -1686,7 +1686,7 @@ static struct ionic_lif *ionic_lif_alloc(struct ionic *ionic, unsigned int index /* Convert the default coalesce value to actual hw resolution */ lif->rx_coalesce_usecs = IONIC_ITR_COAL_USEC_DEFAULT; - lif->rx_coalesce_hw = ionic_coal_hw_to_usec(lif->ionic, + lif->rx_coalesce_hw = ionic_coal_usec_to_hw(lif->ionic, lif->rx_coalesce_usecs); snprintf(lif->name, sizeof(lif->name), "lif%u", index);
Fix the initial interrupt coalesce usec-to-hw setting to actually be usec-to-hw. Fixes: 780eded34ccc ("ionic: report users coalesce request") Signed-off-by: Shannon Nelson <snelson@pensando.io> --- drivers/net/ethernet/pensando/ionic/ionic_lif.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)