<div dir="ltr">Hi Dan,<div><br></div><div>SD8787 does not support interleaved scanning i.e. scanning one channel and then coming back to operating channel for TX/RX.</div><div><br></div><div>We have recently added channel scan gap support to mwifiex driver; but that is applicable only for newer chipsets.</div><div><br></div><div><a href="http://git.kernel.org/cgit/linux/kernel/git/linville/wireless-testing.git/commit/drivers/net/wireless/mwifiex?id=cb91be87e12c079fcf25069a21123756c0931dd6">http://git.kernel.org/cgit/linux/kernel/git/linville/wireless-testing.git/commit/drivers/net/wireless/mwifiex?id=cb91be87e12c079fcf25069a21123756c0931dd6</a><br></div><div><br></div><div>Thanks,</div><div>Avinash</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Fri, Oct 10, 2014 at 10:59 PM, Dan Williams <span dir="ltr"><<a href="mailto:dcbw@redhat.com" target="_blank">dcbw@redhat.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">On Fri, 2014-10-10 at 11:56 +0200, Víctor Andrés wrote:<br>
> Hi Avinash,<br>
><br>
> I'm using the firmware version 14.66.9.p192<br>
> (<a href="http://git.marvell.com/?p=mwifiex-firmware.git;a=commit;h=633d06edab7a2b82efaa134ec5591cd20f9ac0a3" target="_blank">http://git.marvell.com/?p=mwifiex-firmware.git;a=commit;h=633d06edab7a2b82efaa134ec5591cd20f9ac0a3</a>)<br>
> and testing the recent commit p52 version<br>
> (<a href="http://git.marvell.com/?p=mwifiex-firmware.git;a=commit;h=3f45b8c4cc1eb1d102bc3486b19677332dd215ab" target="_blank">http://git.marvell.com/?p=mwifiex-firmware.git;a=commit;h=3f45b8c4cc1eb1d102bc3486b19677332dd215ab</a>),<br>
</span>> but I'm not sure of this last commit, becauseeven if it was posted more<br>
> recent, it's a previous versionof the FW.<br>
<span class="">><br>
> Yes, the roaming is working now. I'm trying to reduce the scanning<br>
</span>> timein order toscan only the channel in use or only for the known bssid.<br>
<span class="">> When the interface is scanning it isn't transmitting data, and what I<br>
> need is a continuous flow of data.<br>
<br>
</span>Scanning only the channel-in-use doesn't work very well for roaming,<br>
because close-by APs are quite likely to be on a different channel to<br>
ensure they don't interfere with each other. Scanning for a known BSSID<br>
doesn't work very well because you still have to scan a bunch of<br>
channels to find the BSSID you're looking for. The driver/mac80211<br>
should be capable of partial scans (send PS, scan 2 or 3 channels, jump<br>
back to operating channel for a while), is that not happening? How long<br>
is the card off the operating channel for, and what are your latency<br>
requirements? Is 100ms too long to pause data transmission?<br>
<br>
Dan<br>
<span class=""><br>
> Víctor<br>
><br>
> El 10/10/2014 10:55, Avinash Patil escribió:<br>
> > Hi Victor,<br>
> ><br>
> > What is SD8787 FW version you are using? As I understand from your<br>
> > last email, roaming is now working fine, right?<br>
> ><br>
> > Thanks,<br>
> > Avinash<br>
> ><br>
> > On Fri, Oct 10, 2014 at 12:38 PM, Víctor Andrés <<a href="mailto:victor@cymonline.com">victor@cymonline.com</a><br>
</span><div><div class="h5">> > <mailto:<a href="mailto:victor@cymonline.com">victor@cymonline.com</a>>> wrote:<br>
> ><br>
> > Hi Dan.<br>
> > Thanks by your help.<br>
> ><br>
> > I tested it in a different board and the signal level of the APs<br>
> > were correct. At least, not all the same and with a value of -101.<br>
> > May be a problem of the chip wifi of that board, I don't know. I<br>
> > use the same antenna for the 2 tests, and the boards were in the<br>
> > same position (more or less) for the 2 tests (with the same AP<br>
> > locations). I'll continue with the tests. I don't understand how<br>
> > can I connect with these APs that the wifi have detected with that<br>
> > poor signal.<br>
> ><br>
> > I've compiled Wpa-supplicant with bgscan option enabled and tested<br>
> > in a different board, and now I have roaming between that 2 APs<br>
> > (with bgscan:learn). Now I'm trying to reduce the searching time,<br>
> > because when the wireless is searching network it can't transmit<br>
> > data, isn't it?<br>
> ><br>
> > Víctor<br>
> ><br>
> ><br>
> > El 08/10/2014 19:33, Dan Williams escribió:<br>
> ><br>
> > On Wed, 2014-10-08 at 12:35 +0200, Víctor Andrés Andrés wrote:<br>
> ><br>
> > First, roaming works better with "bgscan" options enabled.<br>
> > This option<br>
> > tells the supplicant to periodically scan looking for a better<br>
> > AP. If<br>
> > scans don't happen (either through "bgscan" or manually via<br>
> > the control<br>
> > interface) the supplicant will stick with the current AP until the<br>
> > connection is broken. The only way the supplicant knows which<br>
> > AP is<br>
> > better is if a scan has taken place and it knows the RSSI of<br>
> > each AP.<br>
> ><br>
> > Second, it looks like the driver is broken for signal strength<br>
> > reporting. -100dBm signal level is quite awful, and the fact<br>
> > that it<br>
> > shows *all* APs at that level is pretty much a smoking gun.<br>
> > This is the<br>
> > reason the supplicant won't switch APs even if you do trigger<br>
> > a manual<br>
> > scan, because no AP really has a better signal level than the<br>
> > current<br>
> > one.<br>
> ><br>
> > (one other slight possibility: your antenna isn't connected<br>
> > very well,<br>
> > or isn't connected at all, and you're close to the APs...)<br>
> ><br>
> > Dan<br>
> ><br>
> ><br>
> > _______________________________________________<br>
> > HostAP mailing list<br>
</div></div>> > <a href="mailto:HostAP@lists.shmoo.com">HostAP@lists.shmoo.com</a> <mailto:<a href="mailto:HostAP@lists.shmoo.com">HostAP@lists.shmoo.com</a>><br>
> > <a href="http://lists.shmoo.com/mailman/listinfo/hostap" target="_blank">http://lists.shmoo.com/mailman/listinfo/hostap</a><br>
> ><br>
> ><br>
><br>
<br>
<br>
</blockquote></div><br></div>