| Message ID | 20210719145317.79692-1-stephan@gerhold.net |
|---|---|
| Headers | show |
| Series | net: wwan: Add Qualcomm BAM-DMUX WWAN network driver | expand |
On Mon, Jul 19, 2021 at 09:43:27AM -0600, Jeffrey Hugo wrote: > On Mon, Jul 19, 2021 at 9:01 AM Stephan Gerhold <stephan@gerhold.net> wrote: > > > > The BAM Data Multiplexer provides access to the network data channels > > of modems integrated into many older Qualcomm SoCs, e.g. Qualcomm MSM8916 > > or MSM8974. This series adds a driver that allows using it. > > > > For more information about BAM-DMUX, see PATCH 4/4. > > > > Shortly said, BAM-DMUX is built using a simple protocol layer on top of > > a DMA engine (Qualcomm BAM DMA). For BAM-DMUX, the BAM DMA engine runs in > > a quite strange mode that I call "remote power collapse", where the > > modem/remote side is responsible for powering on the BAM when needed but we > > are responsible to initialize it. The BAM is power-collapsed when unneeded > > by coordinating power control via bidirectional interrupts from the > > BAM-DMUX driver. > > The hardware is physically located on the modem, and tied to the modem > regulators, etc. The modem has the ultimate "off" switch. However, > due to the BAM architecture (which is complicated), configuration uses > cooperation on both ends. > What I find strange is that it wasn't done similarly to e.g. Slimbus which has a fairly similar setup. (I used that driver as inspiration for how to use the mainline qcom_bam driver instead of the "SPS" from downstream.) Slimbus uses qcom,controlled-remotely together with the LPASS remoteproc, so it looks like there LPASS does both power-collapse and initialization of the BAM. Whereas here the modem does the power-collapse but we're supposed to do the initialization. > > > > The series first adds one possible solution for handling this "remote power > > collapse" mode in the bam_dma driver, then it adds the BAM-DMUX driver to > > the WWAN subsystem. Note that the BAM-DMUX driver does not actually make > > use of the WWAN subsystem yet, since I'm not sure how to fit it in there > > yet (see PATCH 4/4). > > > > Please note that all of the changes in this patch series are based on > > a fairly complicated driver from Qualcomm [1]. > > I do not have access to any documentation about "BAM-DMUX". :( > > I'm pretty sure I still have the internal docs. > > Are there specific things you want to know? Oh, thanks a lot for asking! I mainly mentioned this here to avoid in-depth questions about the hardware (since I can't answer those). I can probably think of many, many questions, but I'll try to limit myself to the two I'm most confused about. :-) It's somewhat unrelated to this initial patch set since I'm not using QMAP at the moment, but I'm quite confused about the "MTU negotiation feature" that you added support for in [1]. (I *think* that is you, right?) :) The part that I somewhat understand is the "signal" sent in the "OPEN" command from the modem. It tells us the maximum buffer size the modem is willing to accept for TX packets ("ul_mtu" in that commit). Similarly, if we send "OPEN" to the modem we make the modem aware of our maximum RX buffer size plus the number of RX buffers. (create_open_signal() function). The part that is confusing me is the way the "dynamic MTU" is enabled/disabled based on the "signal" in "DATA" commands as well. (process_dynamic_mtu() function). When would that happen? The code suggests that the modem might just suddenly announce that the large MTU should be used from now on. But the "buffer_size" is only changed for newly queued RX buffers so I'm not even sure how the modem knows that it can now send more data at once. Any chance you could clarify how this should work exactly? And a second question if you don't mind: What kind of hardware block am I actually talking to here? I say "modem" above but I just know about the BAM and the DMUX protocol layer. I have also seen assertion failures of the modem DSP firmware if I implement something incorrectly. Is the DMUX protocol just some firmware concept or actually something understood by some hardware block? I've also often seen mentions of some "A2" hardware block but I have no idea what that actually is. What's even worse, in a really old kernel A2/BAM-DMUX also appears as part of the IPA driver [2], and I thought IPA is the new thing after BAM-DMUX... Not sure how much you can reveal about this. :) Thanks a lot! Stephan [1]: https://source.codeaurora.org/quic/la/kernel/msm-3.10/commit/?h=LA.BR.1.2.9.1-02310-8x16.0&id=c7001b82388129ee02ac9ae1a1ef9993eafbcb26 [2]: https://source.codeaurora.org/quic/la/kernel/msm/tree/drivers/platform/msm/ipa/a2_service.c?h=LA.BF.1.1.3-01610-8x74.0
Hi Sergey, On Tue, Jul 20, 2021 at 12:10:42PM +0300, Sergey Ryazanov wrote: > On Mon, Jul 19, 2021 at 6:01 PM Stephan Gerhold <stephan@gerhold.net> wrote: > > The BAM Data Multiplexer provides access to the network data channels of > > modems integrated into many older Qualcomm SoCs, e.g. Qualcomm MSM8916 or > > MSM8974. It is built using a simple protocol layer on top of a DMA engine > > (Qualcomm BAM) and bidirectional interrupts to coordinate power control. > > > > The modem announces a fixed set of channels by sending an OPEN command. > > The driver exports each channel as separate network interface so that > > a connection can be established via QMI from userspace. The network > > interface can work either in Ethernet or Raw-IP mode (configurable via > > QMI). However, Ethernet mode seems to be broken with most firmwares > > (network packets are actually received as Raw-IP), therefore the driver > > only supports Raw-IP mode. > > > > The driver uses runtime PM to coordinate power control with the modem. > > TX/RX buffers are put in a kind of "ring queue" and submitted via > > the bam_dma driver of the DMAEngine subsystem. > > > > The basic architecture looks roughly like this: > > > > +------------+ +-------+ > > [IPv4/6] | BAM-DMUX | | | > > [Data...] | | | | > > ---------->|rmnet0 | [DMUX chan: x] | | > > [IPv4/6] | (chan: 0) | [IPv4/6] | | > > [Data...] | | [Data...] | | > > ---------->|rmnet1 |--------------->| Modem | > > | (chan: 1) | BAM | | > > [IPv4/6] | ... | (DMA Engine) | | > > [Data...] | | | | > > ---------->|rmnet7 | | | > > | (chan: 7) | | | > > +------------+ +-------+ > > > > However, on newer SoCs/firmware versions Qualcomm began gradually moving > > to QMAP (rmnet driver) as backend-independent protocol for multiplexing > > and data aggegration. Some firmware versions allow using QMAP on top of > > BAM-DMUX (effectively resulting in a second multiplexing layer plus data > > aggregation). The architecture with QMAP would look roughly like this: > > > > +-------------+ +------------+ +-------+ > > [IPv4/6] | RMNET | | BAM-DMUX | | | > > [Data...] | | | | [DMUX chan: 0] | | > > ---------->|rmnet_data1 | ----->|rmnet0 | [QMAP mux-id: x] | | > > | (mux-id: 1) | | | (chan: 0) | [IPv4/6] | | > > | | | | | [Data...] | | > > [IPv4/6] | ... |------ | |----------------->| Modem | > > [Data...] | | | | BAM | | > > ---------->|rmnet_data42 | [QMAP: x] |[rmnet1] | (DMA Engine) | | > > | (mux-id: 42)| [IPv4/6] |... unused! | | | > > | | [Data...] |[rmnet7] | | | > > | | | | | | > > +-------------+ +------------+ +-------+ > > > > In this case, rmnet1-7 would remain unused. The firmware used on the most > > recent SoCs with BAM-DMUX even seems to announce only a single BAM-DMUX > > channel (rmnet0), which makes QMAP the only option for multiplexing there. > > > > So far the driver is mainly tested on various smartphones/tablets based on > > Qualcomm MSM8916/MSM8974 without QMAP. It looks like QMAP depends on a MTU > > negotiation feature in BAM-DMUX which is not yet supported by the driver. > > > > Signed-off-by: Stephan Gerhold <stephan@gerhold.net> > > --- > > Note that this is my first network driver, so I apologize in advance > > if I made some obvious mistakes. :) > > > > I'm not sure how to integrate the driver with the WWAN subsystem yet. > > At the moment the driver creates network interfaces for all channels > > announced by the modem, it does not make use of the WWAN link management > > yet. Unfortunately, this is a bit complicated: > > > > Both QMAP and the built-in multiplexing layer might be needed at some point. > > There are firmware versions that do not support QMAP and the other way around > > (the built-in multiplexing was disabled on very recent firmware versions). > > Only userspace can check if QMAP is supported in the firmware (via QMI). > > I am not very familiar with the Qualcomm protocols and am just curious > whether BAM-DMUX has any control (management) channels or only IPv4/v6 > data channels? > > The WWAN subsystem began as a framework for exporting management > interfaces (MBIM, AT, etc.) to user space. And then the network > interfaces (data channels) management interface was added to > facilitate management of devices with multiple data channels. That is > why I am curious about the BAM-DMUX device management interface or in > other words, how a user space application could control the modem > work? > Sorry for the confusion! It's briefly mentioned in the Kconfig option but I should have made this more clear in the commit message. It was so long already that I wasn't sure where to put it. :) BAM-DMUX does not have any control channels. Instead I use it together with the rpmsg_wwan_ctrl driver [1] that I already submitted for 5.14. The control/data channels are pretty much separate in this setup and don't have much to do with each other. I also had a short overview of some of the many different modem protocols Qualcomm has come up with in a related RFC for that driver, see [2] if you are curious. I hope that clarifies some things, please let me know if I should explain something better! :) Thanks! Stephan [1]: https://lore.kernel.org/netdev/20210618173611.134685-3-stephan@gerhold.net/ [2]: https://lore.kernel.org/netdev/YLfL9Q+4860uqS8f@gerhold.net/
On Mon, Jul 19, 2021 at 05:13:32PM -0600, Jeffrey Hugo wrote: > On 7/19/2021 12:23 PM, Stephan Gerhold wrote: > > On Mon, Jul 19, 2021 at 09:43:27AM -0600, Jeffrey Hugo wrote: > > > On Mon, Jul 19, 2021 at 9:01 AM Stephan Gerhold <stephan@gerhold.net> wrote: > > > > > > > > The BAM Data Multiplexer provides access to the network data channels > > > > of modems integrated into many older Qualcomm SoCs, e.g. Qualcomm MSM8916 > > > > or MSM8974. This series adds a driver that allows using it. > > > > > > > > For more information about BAM-DMUX, see PATCH 4/4. > > > > > > > > Shortly said, BAM-DMUX is built using a simple protocol layer on top of > > > > a DMA engine (Qualcomm BAM DMA). For BAM-DMUX, the BAM DMA engine runs in > > > > a quite strange mode that I call "remote power collapse", where the > > > > modem/remote side is responsible for powering on the BAM when needed but we > > > > are responsible to initialize it. The BAM is power-collapsed when unneeded > > > > by coordinating power control via bidirectional interrupts from the > > > > BAM-DMUX driver. > > > > > > The hardware is physically located on the modem, and tied to the modem > > > regulators, etc. The modem has the ultimate "off" switch. However, > > > due to the BAM architecture (which is complicated), configuration uses > > > cooperation on both ends. > > > > > > > What I find strange is that it wasn't done similarly to e.g. Slimbus > > which has a fairly similar setup. (I used that driver as inspiration for > > how to use the mainline qcom_bam driver instead of the "SPS" from > > downstream.) > > > > Slimbus uses qcom,controlled-remotely together with the LPASS > > remoteproc, so it looks like there LPASS does both power-collapse > > and initialization of the BAM. Whereas here the modem does the > > power-collapse but we're supposed to do the initialization. > > I suspect I don't have a satisfactory answer for you. The teams that did > slimbus were not the teams involved in the bam_dmux, and the two didn't talk > to each-other. The bam_dmux side wasn't aware of the slimbus situation, at > the time. I don't know if the slimbus folks knew about bam_dmux. If you > have two silos working independently, its unlikely they will create exactly > the same solution. > Fair enough :) > > > > It's somewhat unrelated to this initial patch set since I'm not using > > QMAP at the moment, but I'm quite confused about the "MTU negotiation > > feature" that you added support for in [1]. (I *think* that is you, > > right?) :) > > Yes. Do I owe you for some brain damage? :) > A bit to be absolutely honest. :D But I was able to ignore this feature so far so it was not much of a problem. ;) > > > > The part that I somewhat understand is the "signal" sent in the "OPEN" > > command from the modem. It tells us the maximum buffer size the modem > > is willing to accept for TX packets ("ul_mtu" in that commit). > > > > Similarly, if we send "OPEN" to the modem we make the modem aware > > of our maximum RX buffer size plus the number of RX buffers. > > (create_open_signal() function). > > > > The part that is confusing me is the way the "dynamic MTU" is > > enabled/disabled based on the "signal" in "DATA" commands as well. > > (process_dynamic_mtu() function). When would that happen? The code > > suggests that the modem might just suddenly announce that the large > > MTU should be used from now on. But the "buffer_size" is only changed > > for newly queued RX buffers so I'm not even sure how the modem knows > > that it can now send more data at once. > > > > Any chance you could clarify how this should work exactly? > > So, I think some of this might make more sense after my response to question > #2. > Indeed, I was worried that you wouldn't be able to answer the second one, otherwise I would probably have asked it first. I'll reorder the mail because it's clearer: > > And a second question if you don't mind: What kind of hardware block > > am I actually talking to here? I say "modem" above but I just know about > > the BAM and the DMUX protocol layer. I have also seen assertion failures > > of the modem DSP firmware if I implement something incorrectly. > > > > Is the DMUX protocol just some firmware concept or actually something > > understood by some hardware block? I've also often seen mentions of some > > "A2" hardware block but I have no idea what that actually is. What's > > even worse, in a really old kernel A2/BAM-DMUX also appears as part of > > the IPA driver [2], and I thought IPA is the new thing after BAM-DMUX... > > A2 predates IPA. IPA is essentially an evolution of A2. > > Sit down son, let me tell you the history of the world :) > > A long time ago, there was only a single processor that did both the "modem" > and the "apps". We generally would call these the 6K days as that was the > number of the chips (6XXX). Then it was decided that the roles of Apps and > Modem should be separated into two different cores. The modem, handling more > "real time" things, and apps, being more "general purpose". This started > with the 7K series. > > However, this created a problem as data from a data call may need to be > consumed by the modem, or the apps, and it wouldn't be clear until the > packet headers were inspected, where the packet needed to be routed to. > Sometimes this was handled on apps, sometimes on modem. Usually via a fully > featured IP stack. > > With LTE, software couldn't really keep up, and so a hardware engine to > parse the fields and route the package based on programmed filters was > implemented. This is the "Algorithm Accelerator", aka AA, aka A2. > > The A2 first appeared on the 9600 chip, which was originally intended for > Gobi- those dongles you could plug into your laptop to give it a data > connection on the go when there was no wifi. It was then coupled with both > 7x30 and 8660 in what we would call "fusion" to create the first LTE capable > phones (HTC thunderbolt is the product I recall) until an integrated > solution could come along. > > That integrated solution was 8960. > > Back to the fusion solution for a second, the 9600 was connected to the > 7x30/8660 via SDIO. Prior to this, the data call control and data path was > all in chip via SMD. Each rmnet instance had its own SMD channel, so > essentially its own physical pipe. With SDIO and 9600, there were not > enough lanes, so we invented SDIO_CMUX and SDIO_DMUX - the Control and Data > multiplexers over SDIO. > > With 8960, everything was integrated again, so we could run the control path > over SMD and didn't need a mux. However, the A2 moved from the 9600 modem > to the 8960 integrated modem, and now we had a direct connection to its BAM. > Again, the BAM had a limited number of physical pipes, so we needed a data > multiplexer again. Thus SDIO_DMUX evolved into BAM_DMUX. > > The A2 is a hardware block with an attached BAM, that "hangs off" the modem. > There is a software component that also runs on the modem, but in general is > limited to configuration. Processing of data is expected to be all in > hardware. As I think I mentioned, the A2 is a hardware engine that routes > IP packets based on programmed filters. > > BAM instances (as part of the smart peripheral subsystem or SPS) can either > be out in the system, or attached to a peripheral. The A2 BAM is attached > to the A2 peripheral. BAM instances can run in one of 3 modes - BAM-to-BAM, > BAM-to-System, or System-to-System. BAM-to-BAM is two BAM instances talking > to eachother. If the USB controller has a BAM, and the A2 has a BAM, those > two BAMS could talk directly to copy data between the A2 and USB hardware > blocks without software interaction (after some configuration). "System" > means system memory, or DDR. Bam-to-System is the mode the A2 BAM runs in > where it takes data to/from DDR and gives/takes that data with the A2. > System-to-System would be used by a BAM instance not associated with any > peripheral to transfer data say from Apps DDR to Modem DDR. > > The A2 can get data from the RF interface, and determine if that needs to go > to some modem consumer, the apps processor, or on some chips to the wifi > processor. All in hardware, much faster than software for multiple reasons, > but mainly because multiple filters can be evaluated in parallel, each > filter looking at multiple fields in parallel. In a nutshell, the IPA is a > revised A2 that is not associated with any processor (like the modem), which > allows it to route data better (think wifi and audio usecases). > > Hope that all helps. I'm "around" for more questions. > Wow, I can't thank you enough for all the detailed explanations! I've seen many small hints of this in various places but I could never really understand how they all relate to each other. This is much clearer now. :) > I don't know how much of this translates to modern platforms. I don't > really work on MSMs anymore, but I can convey what I recall and how things > were "back then" > > So, essentially the change you are looking at is the bam_dmux portion of an > overall feature for improving the performance of what was known as "tethered > rmnet". > > Per my understanding (which the documentation of this feature reinforces), > teathered rmnet was chiefly a test feature. Your "data" (websites, email, > etc) could be consumed by the device itself, or exported off, if you > teathered your phone to a laptop so that the laptop could use the phone's > data connection. There ends up being 3 implementations for this. > > Consuming the data on the phone would route it to the IP stack via the rmnet > driver. > > Consuming the data on an external device could take one of 2 routes. > > Android would use the "native" routing of the Linux IP stack to essentially > NAT the laptop. The data would go to the rmnet driver, to the IP stack, and > the IP stack would route it to USB. > > The other route is that the data could be routed directly to USB. This is > "teathered rmnet". In the case of bam_dmux platforms, the USB stack is a > client of bam_dmux. > > Teathered rmnet was never an end-user usecase. > I'm pretty sure it's actively used now on typical USB modems based on MDM9607. As far as I know that one has BAM-DMUX and "forwards" it via USB (without NAT). > It was essentially a validation feature for both internal testing, and > also qualifying the device with the carriers. The carriers knew that > Android teathering involved NAT based routing on the phone, and wanted > to figure out if the phone could meet the raw performance specs of the > RF technology (LTE Category 4 in this case) in a tethered scenario, > without the routing. > > For tethered rmnet, USB (at the time) was having issues consistently meeting > those data rates (50mbps UL, 100mbps DL concurrently, if I recall > correctly). So, the decided solution was to implement QMAP aggregation. > > A QMAP "call" over tethered rmnet would be negotiated between the app on the > PC, and "dataservices" or "DS" on the modem. One of the initial steps of > that negotitation causes DS to tell A2 software that QMAP over tethered > rmnet is being activated. That would trigger A2 to activate the > process_dynamic_mtu() code path. Now bam_dmux would allocate future RX > buffers of the increased size which could handle the aggregated packets. I > think the part that is confusing you is, what about the already queued > buffers that are of the old size? Well, essentially those get consumed by > the rest of the QMAP call negotiation, so by the time actual aggregated data > is going to be sent from Modem to bam_dmux, the pool has been consumed and > refilled. > > When the tethered rmnet connection is "brought down", DS notifies A2, and A2 > stops requesting the larger buffers. > Hmm, is this "DS" on the modem something special I don't know about? It sounds like the part of the modem that I talk to via QMI to establish new connections. However, since QMI does not go through BAM-DMUX (RPMSG/SMD or QRTR instead) there should be only very few packets sent via BAM-DMUX during negotation of QMAP. To be sure I just tried QMAP with my BAM-DMUX driver again. It's been quite some time since I tried it and it turns out this causes even more "brain damage" than I could even remember. :D For reference: 1. First I need to set the modem to QMAP mode, this works e.g. with qmicli -pd /dev/wwan0qmi0 \ --wda-set-data-format="link-layer-protocol=raw-ip,ul-protocol=qmap,dl-protocol=qmap,dl-datagram-max-size=4096" However, it's important that my BAM-DMUX driver OPENs the channel before doing this (together with announcing support for the "dynamic MTU" feature). Otherwise the modem hangs forever and stops responding to any QMI messages. This doesn't happen when switching to Raw-IP mode. 2. With QMAP, the struct bam_dmux_hdr->len is always set to 0xffff (65535) instead of the actual packet length, which means my current driver just drops those packets ("Data larger than buffer? (65535 > 4088)"). This is also handled in your commit (you get the size from the SPS driver instead), but the bam_dma driver in mainline currently does not have this feature. :/ 3. I sent some ping packets but never got the signal to "enable large MTU" from the modem. Something is still strange here. :/ Given all these complications (that are not present when ignoring QMAP) I would generally agree with you that it's not worth supporting this: > Since this not something an end user should ever exercise, you may want to > consider dropping it. > Personally, I have indeed no need for it. I just suspect someone might want this eventually for one of the following two use cases: 1. Multiplexing on new firmwares: AFAICT there is only one BAM-DMUX channel on recent firmware versions (e.g. MSM8937/MDM9607). In that case multiple connections are only possible through the multiplexing layer in QMAP. I've been told the multiplexing is actually useful and necessary in some cases (maybe it was for some MMS configurations, I don't remember exactly). 2. USB tethering: I know some people are working on mainline Linux for some MDM9607-based USB modems and they will probably want the weird USB tethering feature at some point. But all in all given all the trouble involved when making QMAP work I think I will just ignore that feature for now and wait until someone shows up who absolutely needs this feature... Thanks again for all the explanations! Stephan
Hi Loic, On Mon, Jul 19, 2021 at 06:01:33PM +0200, Loic Poulain wrote: > On Mon, 19 Jul 2021 at 17:01, Stephan Gerhold <stephan@gerhold.net> wrote: > > > > I'm not sure how to integrate the driver with the WWAN subsystem yet. > > At the moment the driver creates network interfaces for all channels > > announced by the modem, it does not make use of the WWAN link management > > yet. Unfortunately, this is a bit complicated: > > > > Both QMAP and the built-in multiplexing layer might be needed at some point. > > There are firmware versions that do not support QMAP and the other way around > > (the built-in multiplexing was disabled on very recent firmware versions). > > Only userspace can check if QMAP is supported in the firmware (via QMI). > > > > I could ignore QMAP completely for now but I think someone will show up > > who will need this eventually. And if there is going to be common code for > > QMAP/rmnet link management it would be nice if BAM-DMUX could also make > > use of it. > > I have this on my TODO list for mhi-net QMAP. > Great, thanks! > > But the question is, how could this look like? How do we know if we should > > create a link for QMAP or a BAM-DMUX channel? Does it even make sense > > to manage the 1-8 channels via the WWAN link management? > > Couldn't it be specified via dts (property or different compatible > string)? It would probably work in most cases, but I have to admit that I would prefer to avoid this for the following reason: This driver is used on some smartphones that have different variants for different parts of the world. As far as Linux is concerned the hardware is pretty much identical, but the modem firmware is often somewhat device-specific. This means that the same device tree is often used with different firmware versions. Perhaps we are lucky enough that the firmware versions have the same capabilities, but I'm not fully sure about that. I think at the end the situation is fairly similar to qmi_wwan/USB. There the kernel also does not know if the modem supports QMAP or not. The way it's solved there at the moment is that ModemManager tries to enable it from user space and then the mode of the network interface can be switched through a sysfs file ("qmi/pass_through"). Something like this should probably also work in my case. This should also allow me to ignore QMAP for now and deal with it if someone really needs it at some point since it's quite complicated for BAM-DMUX. (I tried QMAP again today and listed the problems in [1] for reference, but it's all BAM-DMUX specific...) [1] https://lore.kernel.org/netdev/YPmF8bzevuabO2K9@gerhold.net/ > > would it make sense to have two drivers (with common core) to > manage either the multi-bam channel or newer QMAP based single > bam-channel modems. > There should be fairly little difference between those two usage modes, so I don't think it's worth splitting the driver for this. Actually right now (ignoring the link management of the WWAN subsystem), it's already possible to use both. I can use the network interfaces as-is in Raw-IP mode or I do "sudo ip link add link rmnet0 name rmnet0_qmap type rmnet mux_id 1" on top and use QMAP. The BAM-DMUX driver does not care, because it just hands over sent/received packets as-is and the modem data format must be always configured via QMI from user space. > > > > Another problem is that the WWAN subsystem currently creates all network > > interfaces below the common WWAN device. This means that userspace like > > ModemManager has no way to check which driver provides them. This is > > necessary though to decide how to set it up via QMI (ModemManager uses it). > > Well, I have quite a similar concern since I'm currently porting > mhi-net mbim to wwan framework, and I was thinking about not making > wwan device parent of the network link/netdev (in the same way as > wlan0 is not child of ieee80211 device), but not sure if it's a good > idea or not since we can not really consider driver name part of the > uapi. > Hm, I think the main disadvantage of that would be that the network interface is no longer directly related to the WWAN device, right? Userspace would then need some special matching to find the network interfaces that belong to a certain control port. With the current setup, e.g. ModemManager can simply match the WWAN device and then look at its children and find the control port and network interfaces. How would it find the network interfaces if they are no longer below the WWAN device? > The way links are created is normally abstracted, so if you know which > bam variant you have from wwan network driver side (e.g. via dts), you > should have nothing to check on the user side, except the session id. In a perfect world it would probably be like this, but I'm afraid the Qualcomm firmware situation isn't as simple. User space needs to know which setup it is dealing with because all the setup happens via QMI. Let's take the BAM-DMUX channels vs QMAP mux-IDs for example: First, user space needs to configure the data format. This happens with the QMI WDA (Wireless Data Administrative Service) "Set Data Format" message. Parameter would be link layer format (Raw-IP in both cases) but also the uplink/downlink data aggregation protocol. This is either one of many QMAP versions (qmap|qmapv2|qmapv3|qmapv4|qmapv5), or simply "none" when using BAM-DMUX without QMAP. Then, the "session ID" (= BAM-DMUX channel or QMAP mux-ID) must be bound to a WDS (Wireless Data Service) session. The QMI message for that is different for BAM-DMUX and QMAP: - BAM-DMUX: WDS "Bind Data Port" (Parameter: SIO port number, can be derived from channel ID) - QMAP: WDS "Bind MUX Data Port" (note the "MUX", different message!) (Parameter: MUX ID, port type (USB/embedded/...), port number) My point here: Since userspace is responsible for QMI at the moment we will definitely need to make it aware of the setup that it needs to apply. Just having an abstract "session ID" won't be enough to set up the connection properly. :/ Thanks! Stephan
Hello Stephan, On Wed, Jul 21, 2021 at 3:17 PM Stephan Gerhold <stephan@gerhold.net> wrote: > On Tue, Jul 20, 2021 at 12:10:42PM +0300, Sergey Ryazanov wrote: >> On Mon, Jul 19, 2021 at 6:01 PM Stephan Gerhold <stephan@gerhold.net> wrote: >>> The BAM Data Multiplexer provides access to the network data channels of >>> modems integrated into many older Qualcomm SoCs, e.g. Qualcomm MSM8916 or >>> MSM8974. It is built using a simple protocol layer on top of a DMA engine >>> (Qualcomm BAM) and bidirectional interrupts to coordinate power control. >>> >>> The modem announces a fixed set of channels by sending an OPEN command. >>> The driver exports each channel as separate network interface so that >>> a connection can be established via QMI from userspace. The network >>> interface can work either in Ethernet or Raw-IP mode (configurable via >>> QMI). However, Ethernet mode seems to be broken with most firmwares >>> (network packets are actually received as Raw-IP), therefore the driver >>> only supports Raw-IP mode. >>> >>> The driver uses runtime PM to coordinate power control with the modem. >>> TX/RX buffers are put in a kind of "ring queue" and submitted via >>> the bam_dma driver of the DMAEngine subsystem. >>> >>> The basic architecture looks roughly like this: >>> >>> +------------+ +-------+ >>> [IPv4/6] | BAM-DMUX | | | >>> [Data...] | | | | >>> ---------->|rmnet0 | [DMUX chan: x] | | >>> [IPv4/6] | (chan: 0) | [IPv4/6] | | >>> [Data...] | | [Data...] | | >>> ---------->|rmnet1 |--------------->| Modem | >>> | (chan: 1) | BAM | | >>> [IPv4/6] | ... | (DMA Engine) | | >>> [Data...] | | | | >>> ---------->|rmnet7 | | | >>> | (chan: 7) | | | >>> +------------+ +-------+ >>> >>> However, on newer SoCs/firmware versions Qualcomm began gradually moving >>> to QMAP (rmnet driver) as backend-independent protocol for multiplexing >>> and data aggegration. Some firmware versions allow using QMAP on top of >>> BAM-DMUX (effectively resulting in a second multiplexing layer plus data >>> aggregation). The architecture with QMAP would look roughly like this: >>> >>> +-------------+ +------------+ +-------+ >>> [IPv4/6] | RMNET | | BAM-DMUX | | | >>> [Data...] | | | | [DMUX chan: 0] | | >>> ---------->|rmnet_data1 | ----->|rmnet0 | [QMAP mux-id: x] | | >>> | (mux-id: 1) | | | (chan: 0) | [IPv4/6] | | >>> | | | | | [Data...] | | >>> [IPv4/6] | ... |------ | |----------------->| Modem | >>> [Data...] | | | | BAM | | >>> ---------->|rmnet_data42 | [QMAP: x] |[rmnet1] | (DMA Engine) | | >>> | (mux-id: 42)| [IPv4/6] |... unused! | | | >>> | | [Data...] |[rmnet7] | | | >>> | | | | | | >>> +-------------+ +------------+ +-------+ >>> >>> In this case, rmnet1-7 would remain unused. The firmware used on the most >>> recent SoCs with BAM-DMUX even seems to announce only a single BAM-DMUX >>> channel (rmnet0), which makes QMAP the only option for multiplexing there. >>> >>> So far the driver is mainly tested on various smartphones/tablets based on >>> Qualcomm MSM8916/MSM8974 without QMAP. It looks like QMAP depends on a MTU >>> negotiation feature in BAM-DMUX which is not yet supported by the driver. >>> >>> Signed-off-by: Stephan Gerhold <stephan@gerhold.net> >>> --- >>> Note that this is my first network driver, so I apologize in advance >>> if I made some obvious mistakes. :) >>> >>> I'm not sure how to integrate the driver with the WWAN subsystem yet. >>> At the moment the driver creates network interfaces for all channels >>> announced by the modem, it does not make use of the WWAN link management >>> yet. Unfortunately, this is a bit complicated: >>> >>> Both QMAP and the built-in multiplexing layer might be needed at some point. >>> There are firmware versions that do not support QMAP and the other way around >>> (the built-in multiplexing was disabled on very recent firmware versions). >>> Only userspace can check if QMAP is supported in the firmware (via QMI). >> >> I am not very familiar with the Qualcomm protocols and am just curious >> whether BAM-DMUX has any control (management) channels or only IPv4/v6 >> data channels? >> >> The WWAN subsystem began as a framework for exporting management >> interfaces (MBIM, AT, etc.) to user space. And then the network >> interfaces (data channels) management interface was added to >> facilitate management of devices with multiple data channels. That is >> why I am curious about the BAM-DMUX device management interface or in >> other words, how a user space application could control the modem >> work? > > Sorry for the confusion! It's briefly mentioned in the Kconfig option > but I should have made this more clear in the commit message. It was so > long already that I wasn't sure where to put it. :) > > BAM-DMUX does not have any control channels. Instead I use it together > with the rpmsg_wwan_ctrl driver [1] that I already submitted for 5.14. > The control/data channels are pretty much separate in this setup and > don't have much to do with each other. > > I also had a short overview of some of the many different modem > protocols Qualcomm has come up with in a related RFC for that driver, > see [2] if you are curious. > > I hope that clarifies some things, please let me know if I should > explain something better! :) > > [1]: https://lore.kernel.org/netdev/20210618173611.134685-3-stephan@gerhold.net/ > [2]: https://lore.kernel.org/netdev/YLfL9Q+4860uqS8f@gerhold.net/ Many thanks for such informative clarification, especially for pointing me to the rpmsg_wwan_ctrl driver. I saw it, but by a some reason I did not link it to BAM-DMUX. Reading these links in conjunction with your parallel talks make the situation much more clear. I could not say that "I know kung fu", but I can say that now I know how complex kung fu is. -- Sergey
On Thu, Jul 22, 2021 at 6:40 PM Stephan Gerhold <stephan@gerhold.net> wrote: > On Mon, Jul 19, 2021 at 06:01:33PM +0200, Loic Poulain wrote: >> On Mon, 19 Jul 2021 at 17:01, Stephan Gerhold <stephan@gerhold.net> wrote: >>> I'm not sure how to integrate the driver with the WWAN subsystem yet. >>> At the moment the driver creates network interfaces for all channels >>> announced by the modem, it does not make use of the WWAN link management >>> yet. Unfortunately, this is a bit complicated: >>> >>> Both QMAP and the built-in multiplexing layer might be needed at some point. >>> There are firmware versions that do not support QMAP and the other way around >>> (the built-in multiplexing was disabled on very recent firmware versions). >>> Only userspace can check if QMAP is supported in the firmware (via QMI). >>> >>> I could ignore QMAP completely for now but I think someone will show up >>> who will need this eventually. And if there is going to be common code for >>> QMAP/rmnet link management it would be nice if BAM-DMUX could also make >>> use of it. >> >> I have this on my TODO list for mhi-net QMAP. > > Great, thanks! > >>> But the question is, how could this look like? How do we know if we should >>> create a link for QMAP or a BAM-DMUX channel? Does it even make sense >>> to manage the 1-8 channels via the WWAN link management? >> >> Couldn't it be specified via dts (property or different compatible >> string)? > > It would probably work in most cases, but I have to admit that I would > prefer to avoid this for the following reason: This driver is used on > some smartphones that have different variants for different parts of the > world. As far as Linux is concerned the hardware is pretty much > identical, but the modem firmware is often somewhat device-specific. > > This means that the same device tree is often used with different > firmware versions. Perhaps we are lucky enough that the firmware > versions have the same capabilities, but I'm not fully sure about that. > > I think at the end the situation is fairly similar to qmi_wwan/USB. > There the kernel also does not know if the modem supports QMAP or not. > The way it's solved there at the moment is that ModemManager tries to > enable it from user space and then the mode of the network interface > can be switched through a sysfs file ("qmi/pass_through"). > > Something like this should probably also work in my case. This should > also allow me to ignore QMAP for now and deal with it if someone really > needs it at some point since it's quite complicated for BAM-DMUX. > (I tried QMAP again today and listed the problems in [1] for reference, > but it's all BAM-DMUX specific...) > > [1] https://lore.kernel.org/netdev/YPmF8bzevuabO2K9@gerhold.net/ > >> would it make sense to have two drivers (with common core) to >> manage either the multi-bam channel or newer QMAP based single >> bam-channel modems. > > There should be fairly little difference between those two usage modes, > so I don't think it's worth splitting the driver for this. Actually > right now (ignoring the link management of the WWAN subsystem), > it's already possible to use both. > > I can use the network interfaces as-is in Raw-IP mode or I do > "sudo ip link add link rmnet0 name rmnet0_qmap type rmnet mux_id 1" > on top and use QMAP. The BAM-DMUX driver does not care, because it > just hands over sent/received packets as-is and the modem data format > must be always configured via QMI from user space. > >>> Another problem is that the WWAN subsystem currently creates all network >>> interfaces below the common WWAN device. This means that userspace like >>> ModemManager has no way to check which driver provides them. This is >>> necessary though to decide how to set it up via QMI (ModemManager uses it). >> >> Well, I have quite a similar concern since I'm currently porting >> mhi-net mbim to wwan framework, and I was thinking about not making >> wwan device parent of the network link/netdev (in the same way as >> wlan0 is not child of ieee80211 device), but not sure if it's a good >> idea or not since we can not really consider driver name part of the >> uapi. > > Hm, I think the main disadvantage of that would be that the network > interface is no longer directly related to the WWAN device, right? > Userspace would then need some special matching to find the network > interfaces that belong to a certain control port. > > With the current setup, e.g. ModemManager can simply match the WWAN > device and then look at its children and find the control port and > network interfaces. How would it find the network interfaces if they are > no longer below the WWAN device? > > > The way links are created is normally abstracted, so if you know which > > bam variant you have from wwan network driver side (e.g. via dts), you > > should have nothing to check on the user side, except the session id. > > In a perfect world it would probably be like this, but I'm afraid the > Qualcomm firmware situation isn't as simple. User space needs to know > which setup it is dealing with because all the setup happens via QMI. > > Let's take the BAM-DMUX channels vs QMAP mux-IDs for example: > > First, user space needs to configure the data format. This happens with > the QMI WDA (Wireless Data Administrative Service) "Set Data Format" > message. Parameter would be link layer format (Raw-IP in both cases) > but also the uplink/downlink data aggregation protocol. This is either > one of many QMAP versions (qmap|qmapv2|qmapv3|qmapv4|qmapv5), or simply > "none" when using BAM-DMUX without QMAP. > > Then, the "session ID" (= BAM-DMUX channel or QMAP mux-ID) must be bound > to a WDS (Wireless Data Service) session. The QMI message for that is > different for BAM-DMUX and QMAP: > > - BAM-DMUX: WDS "Bind Data Port" > (Parameter: SIO port number, can be derived from channel ID) > > - QMAP: WDS "Bind MUX Data Port" (note the "MUX", different message!) > (Parameter: MUX ID, port type (USB/embedded/...), port number) > > My point here: Since userspace is responsible for QMI at the moment > we will definitely need to make it aware of the setup that it needs to > apply. Just having an abstract "session ID" won't be enough to set up > the connection properly. :/ Stephan, Loic, I have a polemic question related to a drivers model that we should build to smoothly support qualcomm hardware by the kernel. I would depict the situation as I see it and then ask the question. Please correct me if I am misunderstanding something or simply wrong. Or maybe you will be gracious once more and point me to earlier discussions :) We always talk that a userspace software should take care of multiplexing configuration to make data communication possible at all. The motivation here is simple - management protocol (QMI) is complex, userspace software must implement it anyway to manage network connectivity, so why not implement the multiplexing management there too? This way the userspace software that should simply command a "modem" to establish a data connection and poll a "modem" for a signal level became a self contained device manager that knows all modem-to-host interconnection details and even must to perform an initial modem-to-host interfaces negotiation and configuration. The last task is what userspace software usually expects to be performed by an OS kernel. But what if we implement the QMI multiplexing management part in the kernel? This way the kernel will take care about modem-to-host communication protocols and interfaces, and provides userspace with a single WWAN device (possibly with multiple network and network management interfaces). I do not propose to fully implement QMI protocol inside the kernel, but implement only a mux management part, while passing all other messages between a "modem" and a userspace software as-is. What pros and cons of such a design do you see?
Hey! > > But what if we implement the QMI multiplexing management part in the > kernel? This way the kernel will take care about modem-to-host > communication protocols and interfaces, and provides userspace with a > single WWAN device (possibly with multiple network and network > management interfaces). > > I do not propose to fully implement QMI protocol inside the kernel, > but implement only a mux management part, while passing all other > messages between a "modem" and a userspace software as-is. > > What pros and cons of such a design do you see? > The original GobiNet driver already provided some QMI protocol implementation in the driver itself. In addition to initial device setup as you suggest, it also allowed userspace applications to allocate and release QMI clients for the different services that could be used independently by different processes. Not going to say that was the wrong way to do it, but the implementation is definitely not simple. The decision taken in qmi_wwan to make the driver as simple as possible and leave all the QMI management to userspace was quite an important one; it made the driver extremely simple, leaving all the complexity of managing the protocol to userspace, and while it had some initial drawbacks (e.g. only one process could talk QMI at a time) the userspace tools have evolved to avoid them (e.g. the qmi-proxy). I wrote some time ago about this, maybe it's still relevant today: Blogpost https://sigquit.wordpress.com/2014/06/11/qmiwwan-or-gobinet/, Article in PDF https://aleksander.es/data/Qualcomm%20Gobi%20devices%20on%20Linux.pdf Making the driver talk QMI just for device setup would require the kernel to know how the QMI protocol works, how QMI client allocations and releases are done, how errors are reported, how is the format of the requests and responses involved; it would require the kernel to wait until the QMI protocol endpoint in the modem is capable of returning QMI responses (this could be up to 20 or 30 secs after the device is available in the bus), it would require to have possibly some specific rules on how the QMI clients are managed after a suspend/resume operation. It would also require to sync the access to the CTL service, which is the one running QMI service allocations and releases, so that both kernel and userspace can perform operations with that service at the same time. It would need to know how different QMI capable devices behave, because not all devices support the same services, and some don't even support the WDA service that would be the one needed to setup data aggregation. There is definitely some overlap on what the kernel could do and what userspace could do, and I'd say that we have much more flexibility in userspace to do all this leaving all the complexity out of the kernel driver. ModemManager already provides a unified API to e.g. setup multiplexed data sessions, regardless of what the underlying kernel implementation is (qmi_wwan only, qmi_wwan+rmnet, ipa+rmnet, bam-dmux, cdc_mbim...) . The logic doing all that is extremely complex and possibly full of errors, I would definitely not want to have all that logic in the kernel itself, let the errors be in userspace! Unifying stuff in the kernel is a good idea, but if you ask me, it should be done in a way that is as simple as possible, leaving complexity to userspace, even if that means that userspace still needs to know what type of device we have behind the wwan subsystem, because userspace will anyway need to know all that.
On 7/22/2021 8:51 AM, Stephan Gerhold wrote: > On Mon, Jul 19, 2021 at 05:13:32PM -0600, Jeffrey Hugo wrote: >> On 7/19/2021 12:23 PM, Stephan Gerhold wrote: >>> On Mon, Jul 19, 2021 at 09:43:27AM -0600, Jeffrey Hugo wrote: >>>> On Mon, Jul 19, 2021 at 9:01 AM Stephan Gerhold <stephan@gerhold.net> wrote: >>>>> >>>>> The BAM Data Multiplexer provides access to the network data channels >>>>> of modems integrated into many older Qualcomm SoCs, e.g. Qualcomm MSM8916 >>>>> or MSM8974. This series adds a driver that allows using it. >>>>> >>>>> For more information about BAM-DMUX, see PATCH 4/4. >>>>> >>>>> Shortly said, BAM-DMUX is built using a simple protocol layer on top of >>>>> a DMA engine (Qualcomm BAM DMA). For BAM-DMUX, the BAM DMA engine runs in >>>>> a quite strange mode that I call "remote power collapse", where the >>>>> modem/remote side is responsible for powering on the BAM when needed but we >>>>> are responsible to initialize it. The BAM is power-collapsed when unneeded >>>>> by coordinating power control via bidirectional interrupts from the >>>>> BAM-DMUX driver. >>>> >>>> The hardware is physically located on the modem, and tied to the modem >>>> regulators, etc. The modem has the ultimate "off" switch. However, >>>> due to the BAM architecture (which is complicated), configuration uses >>>> cooperation on both ends. >>>> >>> >>> What I find strange is that it wasn't done similarly to e.g. Slimbus >>> which has a fairly similar setup. (I used that driver as inspiration for >>> how to use the mainline qcom_bam driver instead of the "SPS" from >>> downstream.) >>> >>> Slimbus uses qcom,controlled-remotely together with the LPASS >>> remoteproc, so it looks like there LPASS does both power-collapse >>> and initialization of the BAM. Whereas here the modem does the >>> power-collapse but we're supposed to do the initialization. >> >> I suspect I don't have a satisfactory answer for you. The teams that did >> slimbus were not the teams involved in the bam_dmux, and the two didn't talk >> to each-other. The bam_dmux side wasn't aware of the slimbus situation, at >> the time. I don't know if the slimbus folks knew about bam_dmux. If you >> have two silos working independently, its unlikely they will create exactly >> the same solution. >> > > Fair enough :) > >>> >>> It's somewhat unrelated to this initial patch set since I'm not using >>> QMAP at the moment, but I'm quite confused about the "MTU negotiation >>> feature" that you added support for in [1]. (I *think* that is you, >>> right?) :) >> >> Yes. Do I owe you for some brain damage? :) >> > > A bit to be absolutely honest. :D > But I was able to ignore this feature so far so it was not much of > a problem. ;) > >>> >>> The part that I somewhat understand is the "signal" sent in the "OPEN" >>> command from the modem. It tells us the maximum buffer size the modem >>> is willing to accept for TX packets ("ul_mtu" in that commit). >>> >>> Similarly, if we send "OPEN" to the modem we make the modem aware >>> of our maximum RX buffer size plus the number of RX buffers. >>> (create_open_signal() function). >>> >>> The part that is confusing me is the way the "dynamic MTU" is >>> enabled/disabled based on the "signal" in "DATA" commands as well. >>> (process_dynamic_mtu() function). When would that happen? The code >>> suggests that the modem might just suddenly announce that the large >>> MTU should be used from now on. But the "buffer_size" is only changed >>> for newly queued RX buffers so I'm not even sure how the modem knows >>> that it can now send more data at once. >>> >>> Any chance you could clarify how this should work exactly? >> >> So, I think some of this might make more sense after my response to question >> #2. >> > > Indeed, I was worried that you wouldn't be able to answer the second > one, otherwise I would probably have asked it first. I'll reorder the > mail because it's clearer: > >>> And a second question if you don't mind: What kind of hardware block >>> am I actually talking to here? I say "modem" above but I just know about >>> the BAM and the DMUX protocol layer. I have also seen assertion failures >>> of the modem DSP firmware if I implement something incorrectly. >>> >>> Is the DMUX protocol just some firmware concept or actually something >>> understood by some hardware block? I've also often seen mentions of some >>> "A2" hardware block but I have no idea what that actually is. What's >>> even worse, in a really old kernel A2/BAM-DMUX also appears as part of >>> the IPA driver [2], and I thought IPA is the new thing after BAM-DMUX... >> >> A2 predates IPA. IPA is essentially an evolution of A2. >> >> Sit down son, let me tell you the history of the world :) >> >> A long time ago, there was only a single processor that did both the "modem" >> and the "apps". We generally would call these the 6K days as that was the >> number of the chips (6XXX). Then it was decided that the roles of Apps and >> Modem should be separated into two different cores. The modem, handling more >> "real time" things, and apps, being more "general purpose". This started >> with the 7K series. >> >> However, this created a problem as data from a data call may need to be >> consumed by the modem, or the apps, and it wouldn't be clear until the >> packet headers were inspected, where the packet needed to be routed to. >> Sometimes this was handled on apps, sometimes on modem. Usually via a fully >> featured IP stack. >> >> With LTE, software couldn't really keep up, and so a hardware engine to >> parse the fields and route the package based on programmed filters was >> implemented. This is the "Algorithm Accelerator", aka AA, aka A2. >> >> The A2 first appeared on the 9600 chip, which was originally intended for >> Gobi- those dongles you could plug into your laptop to give it a data >> connection on the go when there was no wifi. It was then coupled with both >> 7x30 and 8660 in what we would call "fusion" to create the first LTE capable >> phones (HTC thunderbolt is the product I recall) until an integrated >> solution could come along. >> >> That integrated solution was 8960. >> >> Back to the fusion solution for a second, the 9600 was connected to the >> 7x30/8660 via SDIO. Prior to this, the data call control and data path was >> all in chip via SMD. Each rmnet instance had its own SMD channel, so >> essentially its own physical pipe. With SDIO and 9600, there were not >> enough lanes, so we invented SDIO_CMUX and SDIO_DMUX - the Control and Data >> multiplexers over SDIO. >> >> With 8960, everything was integrated again, so we could run the control path >> over SMD and didn't need a mux. However, the A2 moved from the 9600 modem >> to the 8960 integrated modem, and now we had a direct connection to its BAM. >> Again, the BAM had a limited number of physical pipes, so we needed a data >> multiplexer again. Thus SDIO_DMUX evolved into BAM_DMUX. >> >> The A2 is a hardware block with an attached BAM, that "hangs off" the modem. >> There is a software component that also runs on the modem, but in general is >> limited to configuration. Processing of data is expected to be all in >> hardware. As I think I mentioned, the A2 is a hardware engine that routes >> IP packets based on programmed filters. >> >> BAM instances (as part of the smart peripheral subsystem or SPS) can either >> be out in the system, or attached to a peripheral. The A2 BAM is attached >> to the A2 peripheral. BAM instances can run in one of 3 modes - BAM-to-BAM, >> BAM-to-System, or System-to-System. BAM-to-BAM is two BAM instances talking >> to eachother. If the USB controller has a BAM, and the A2 has a BAM, those >> two BAMS could talk directly to copy data between the A2 and USB hardware >> blocks without software interaction (after some configuration). "System" >> means system memory, or DDR. Bam-to-System is the mode the A2 BAM runs in >> where it takes data to/from DDR and gives/takes that data with the A2. >> System-to-System would be used by a BAM instance not associated with any >> peripheral to transfer data say from Apps DDR to Modem DDR. >> >> The A2 can get data from the RF interface, and determine if that needs to go >> to some modem consumer, the apps processor, or on some chips to the wifi >> processor. All in hardware, much faster than software for multiple reasons, >> but mainly because multiple filters can be evaluated in parallel, each >> filter looking at multiple fields in parallel. In a nutshell, the IPA is a >> revised A2 that is not associated with any processor (like the modem), which >> allows it to route data better (think wifi and audio usecases). >> >> Hope that all helps. I'm "around" for more questions. >> > > Wow, I can't thank you enough for all the detailed explanations! > I've seen many small hints of this in various places but I could never > really understand how they all relate to each other. > This is much clearer now. :) > >> I don't know how much of this translates to modern platforms. I don't >> really work on MSMs anymore, but I can convey what I recall and how things >> were "back then" >> >> So, essentially the change you are looking at is the bam_dmux portion of an >> overall feature for improving the performance of what was known as "tethered >> rmnet". >> >> Per my understanding (which the documentation of this feature reinforces), >> teathered rmnet was chiefly a test feature. Your "data" (websites, email, >> etc) could be consumed by the device itself, or exported off, if you >> teathered your phone to a laptop so that the laptop could use the phone's >> data connection. There ends up being 3 implementations for this. >> >> Consuming the data on the phone would route it to the IP stack via the rmnet >> driver. >> >> Consuming the data on an external device could take one of 2 routes. >> >> Android would use the "native" routing of the Linux IP stack to essentially >> NAT the laptop. The data would go to the rmnet driver, to the IP stack, and >> the IP stack would route it to USB. >> >> The other route is that the data could be routed directly to USB. This is >> "teathered rmnet". In the case of bam_dmux platforms, the USB stack is a >> client of bam_dmux. >> >> Teathered rmnet was never an end-user usecase. >> > > I'm pretty sure it's actively used now on typical USB modems based on > MDM9607. As far as I know that one has BAM-DMUX and "forwards" it via > USB (without NAT). > >> It was essentially a validation feature for both internal testing, and >> also qualifying the device with the carriers. The carriers knew that >> Android teathering involved NAT based routing on the phone, and wanted >> to figure out if the phone could meet the raw performance specs of the >> RF technology (LTE Category 4 in this case) in a tethered scenario, >> without the routing. >> >> For tethered rmnet, USB (at the time) was having issues consistently meeting >> those data rates (50mbps UL, 100mbps DL concurrently, if I recall >> correctly). So, the decided solution was to implement QMAP aggregation. >> >> A QMAP "call" over tethered rmnet would be negotiated between the app on the >> PC, and "dataservices" or "DS" on the modem. One of the initial steps of >> that negotitation causes DS to tell A2 software that QMAP over tethered >> rmnet is being activated. That would trigger A2 to activate the >> process_dynamic_mtu() code path. Now bam_dmux would allocate future RX >> buffers of the increased size which could handle the aggregated packets. I >> think the part that is confusing you is, what about the already queued >> buffers that are of the old size? Well, essentially those get consumed by >> the rest of the QMAP call negotiation, so by the time actual aggregated data >> is going to be sent from Modem to bam_dmux, the pool has been consumed and >> refilled. >> >> When the tethered rmnet connection is "brought down", DS notifies A2, and A2 >> stops requesting the larger buffers. >> > > Hmm, is this "DS" on the modem something special I don't know about? > It sounds like the part of the modem that I talk to via QMI to establish > new connections. You have the gist of it. Kinda need to dance around here :( > However, since QMI does not go through BAM-DMUX > (RPMSG/SMD or QRTR instead) there should be only very few packets sent > via BAM-DMUX during negotation of QMAP. > > To be sure I just tried QMAP with my BAM-DMUX driver again. It's been > quite some time since I tried it and it turns out this causes even more > "brain damage" than I could even remember. :D For reference: > > 1. First I need to set the modem to QMAP mode, this works e.g. with > qmicli -pd /dev/wwan0qmi0 \ > --wda-set-data-format="link-layer-protocol=raw-ip,ul-protocol=qmap,dl-protocol=qmap,dl-datagram-max-size=4096" > > However, it's important that my BAM-DMUX driver OPENs the channel > before doing this (together with announcing support for the "dynamic > MTU" feature). Otherwise the modem hangs forever and stops responding > to any QMI messages. This doesn't happen when switching to Raw-IP mode. > > 2. With QMAP, the struct bam_dmux_hdr->len is always set to 0xffff (65535) > instead of the actual packet length, which means my current driver > just drops those packets ("Data larger than buffer? (65535 > 4088)"). > > This is also handled in your commit (you get the size from the SPS > driver instead), but the bam_dma driver in mainline currently does > not have this feature. :/ Huh. It seems really odd to me that the client doesn't get "notified" of the actual length of data transferred. That could easily be less than the buffer provided, so there isn't a way for the client to derive that info. > > 3. I sent some ping packets but never got the signal to "enable large > MTU" from the modem. Something is still strange here. :/ > > Given all these complications (that are not present when ignoring QMAP) > I would generally agree with you that it's not worth supporting this: > >> Since this not something an end user should ever exercise, you may want to >> consider dropping it. >> > > Personally, I have indeed no need for it. I just suspect someone might > want this eventually for one of the following two use cases: > > 1. Multiplexing on new firmwares: AFAICT there is only one BAM-DMUX > channel on recent firmware versions (e.g. MSM8937/MDM9607). In that > case multiple connections are only possible through the multiplexing > layer in QMAP. I've been told the multiplexing is actually useful and > necessary in some cases (maybe it was for some MMS configurations, > I don't remember exactly). > > 2. USB tethering: I know some people are working on mainline Linux > for some MDM9607-based USB modems and they will probably want the > weird USB tethering feature at some point. > > But all in all given all the trouble involved when making QMAP work > I think I will just ignore that feature for now and wait until someone > shows up who absolutely needs this feature... QMAP is useful. It finally gets rid of the need to have multiple physical pipes that exist because of "reasons". Sadly, it came about a decade late IMO. Regardless, I think I agree. Focus on what you really care about, and leave everything else until later. Thankfully layering makes that easier. Otherwise, nothing gets done.
On Mon, Jul 19, 2021 at 04:53:14PM +0200, Stephan Gerhold wrote: > In some configurations, the BAM DMA controller is set up by a remote > processor and the local processor can simply start making use of it > without setting up the BAM. This is already supported using the > "qcom,controlled-remotely" property. > > However, for some reason another possible configuration is that the > remote processor is responsible for powering up the BAM, but we are > still responsible for initializing it (e.g. resetting it etc). Add > a "qcom,remote-power-collapse" property to describe that configuration. > > Signed-off-by: Stephan Gerhold <stephan@gerhold.net> > --- > NOTE: This is *not* a compile-time requirement for the BAM-DMUX driver > so this could also go through the dmaengine tree. > > Also note that there is an ongoing effort to convert these bindings > to DT schema but sadly there were not any updates for a while. :/ > https://lore.kernel.org/linux-arm-msm/20210519143700.27392-2-bhupesh.sharma@linaro.org/ > --- > Documentation/devicetree/bindings/dma/qcom_bam_dma.txt | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/Documentation/devicetree/bindings/dma/qcom_bam_dma.txt b/Documentation/devicetree/bindings/dma/qcom_bam_dma.txt > index cf5b9e44432c..362a4f0905a8 100644 > --- a/Documentation/devicetree/bindings/dma/qcom_bam_dma.txt > +++ b/Documentation/devicetree/bindings/dma/qcom_bam_dma.txt > @@ -15,6 +15,8 @@ Required properties: > the secure world. > - qcom,controlled-remotely : optional, indicates that the bam is controlled by > remote proccessor i.e. execution environment. > +- qcom,remote-power-collapse : optional, indicates that the bam is powered up by > + a remote processor but must be initialized by the local processor. Wouldn't 'qcom,remote-power' or 'qcom,remote-powered' be sufficient? I don't understand what 'collapse' means here. Doesn't sound good though. Rob
On Thu, Jul 29, 2021 at 01:36:31PM -0600, Rob Herring wrote: > On Mon, Jul 19, 2021 at 04:53:14PM +0200, Stephan Gerhold wrote: > > In some configurations, the BAM DMA controller is set up by a remote > > processor and the local processor can simply start making use of it > > without setting up the BAM. This is already supported using the > > "qcom,controlled-remotely" property. > > > > However, for some reason another possible configuration is that the > > remote processor is responsible for powering up the BAM, but we are > > still responsible for initializing it (e.g. resetting it etc). Add > > a "qcom,remote-power-collapse" property to describe that configuration. > > > > Signed-off-by: Stephan Gerhold <stephan@gerhold.net> > > --- > > NOTE: This is *not* a compile-time requirement for the BAM-DMUX driver > > so this could also go through the dmaengine tree. > > > > Also note that there is an ongoing effort to convert these bindings > > to DT schema but sadly there were not any updates for a while. :/ > > https://lore.kernel.org/linux-arm-msm/20210519143700.27392-2-bhupesh.sharma@linaro.org/ > > --- > > Documentation/devicetree/bindings/dma/qcom_bam_dma.txt | 2 ++ > > 1 file changed, 2 insertions(+) > > > > diff --git a/Documentation/devicetree/bindings/dma/qcom_bam_dma.txt b/Documentation/devicetree/bindings/dma/qcom_bam_dma.txt > > index cf5b9e44432c..362a4f0905a8 100644 > > --- a/Documentation/devicetree/bindings/dma/qcom_bam_dma.txt > > +++ b/Documentation/devicetree/bindings/dma/qcom_bam_dma.txt > > @@ -15,6 +15,8 @@ Required properties: > > the secure world. > > - qcom,controlled-remotely : optional, indicates that the bam is controlled by > > remote proccessor i.e. execution environment. > > +- qcom,remote-power-collapse : optional, indicates that the bam is powered up by > > + a remote processor but must be initialized by the local processor. > > Wouldn't 'qcom,remote-power' or 'qcom,remote-powered' be sufficient? I > don't understand what 'collapse' means here. Doesn't sound good though. > Yeah I can't think of any significant meaning of the "collapse" part for the bindings, I probably just picked it up somewhere while trying to find some information about how the BAM DMUX setup works. :) Just one question, would you prefer "qcom,remote-powered" or rather "qcom,powered-remotely" for consistency with the existing "qcom,controlled-remotely"? Both sounds fine to me. Thanks! Stephan