Re: [PATCH 00/14] Modifications for DWC OTG since v13
Hi, В Wed, 31 Aug 2011 10:18:42 +0900 Kyungmin Park kmp...@infradead.org пишет: On Wed, Aug 31, 2011 at 12:46 AM, Pratyush Anand pratyush.an...@gmail.com wrote: On Tue, Aug 30, 2011 at 8:57 PM, Tirumala Marri tma...@apm.com wrote: -Original Message- From: Pratyush Anand [mailto:pratyush.an...@st.com] Sent: Tuesday, August 30, 2011 4:58 AM To: linux-...@vger.kernel.org Cc: tma...@apm.com; linuxppc-dev@lists.ozlabs.org; fc...@apm.com; mmiesf...@apm.com; shiraz.has...@st.com; deepak.si...@st.com; vipulkumar.sa...@st.com; rajeev-dlh.ku...@st.com; vipin.ku...@st.com; bhupesh.sha...@st.com; viresh.ku...@st.com; amit.vi...@st.com; Pratyush Anand Subject: [PATCH 00/14] Modifications for DWC OTG since v13 These patches are based on:http://patchwork.ozlabs.org/patch/89560/ After not getting any reply from developers, I started to do modifications for my platform (SPEAr1340). I have done modifications in such a way that all the code in driver/usb/dwc/ would be platform independent. I have tested this code for host/device/dma/slave mode. My fifo configuration is dedicated and dynamic. [Tirumala Marri] We are working on our next release of patches. They should be coming out soon. Oh, thats great !! Actually, I did not get any reply of my previous mail on your patch release. I thought no one is maintaining., and so I sent them with my modifications, after testing them in all dev/host/otg mode. Regards Pratyush Hi, Can you provide the git repo to test? and I wonder what's the difference between dwc (from you) and dwc3 (from Felipe Balbi). I think it's dwc targets for usb 2.0 and dwc3 for usb 3.0. It doesn't look the same. At least, register definitions are quite different. -- Alexander signature.asc Description: PGP signature ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH 00/14] Modifications for DWC OTG since v13
On Wed, Aug 31, 2011 at 6:48 AM, Kyungmin Park kmp...@infradead.org wrote: On Wed, Aug 31, 2011 at 12:46 AM, Pratyush Anand pratyush.an...@gmail.com wrote: On Tue, Aug 30, 2011 at 8:57 PM, Tirumala Marri tma...@apm.com wrote: -Original Message- From: Pratyush Anand [mailto:pratyush.an...@st.com] Sent: Tuesday, August 30, 2011 4:58 AM To: linux-...@vger.kernel.org Cc: tma...@apm.com; linuxppc-dev@lists.ozlabs.org; fc...@apm.com; mmiesf...@apm.com; shiraz.has...@st.com; deepak.si...@st.com; vipulkumar.sa...@st.com; rajeev-dlh.ku...@st.com; vipin.ku...@st.com; bhupesh.sha...@st.com; viresh.ku...@st.com; amit.vi...@st.com; Pratyush Anand Subject: [PATCH 00/14] Modifications for DWC OTG since v13 These patches are based on:http://patchwork.ozlabs.org/patch/89560/ After not getting any reply from developers, I started to do modifications for my platform (SPEAr1340). I have done modifications in such a way that all the code in driver/usb/dwc/ would be platform independent. I have tested this code for host/device/dma/slave mode. My fifo configuration is dedicated and dynamic. [Tirumala Marri] We are working on our next release of patches. They should be coming out soon. Oh, thats great !! Actually, I did not get any reply of my previous mail on your patch release. I thought no one is maintaining., and so I sent them with my modifications, after testing them in all dev/host/otg mode. Regards Pratyush Hi, Can you provide the git repo to test? Currently, these patches are also in circulation for internal reviews. Once reviewed internally, will push them to git repo and will let you know. and I wonder what's the difference between dwc (from you) and dwc3 (from Felipe Balbi). I think it's dwc targets for usb 2.0 and dwc3 for usb 3.0. dwc3 is for DesignWare USB3 DRD Controller Core these patches are for DesignWare USB2.0 High Speed OTG Controller Regards Pratyush Thank you, Kyungmin Park -- To unsubscribe from this list: send the line unsubscribe linux-usb in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line unsubscribe linux-usb in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH 00/14] Modifications for DWC OTG since v13
Hi, (you did not add Greg in Cc, he's the USB maintainer. You also didn't Cc myself as maintainer of the gadget framework) On Tue, Aug 30, 2011 at 05:27:47PM +0530, Pratyush Anand wrote: These patches are based on:http://patchwork.ozlabs.org/patch/89560/ After not getting any reply from developers, I started to do modifications for my platform (SPEAr1340). I have done modifications in such a way that all the code in driver/usb/dwc/ would be platform independent. I have tested this code for host/device/dma/slave mode. My fifo configuration is dedicated and dynamic. Looking at that driver code, I believe this is pretty much Synopsys reference code with a few tweaks. If you want some tip, don't do that. The reference driver, while functional, has lots of unnecessary layering which should be avoided. See how we wrote the DWC3 driver. It's very small code which does the necessary. It's now functionally complete, from now on we only need to improve on performance (for example starting transfers on Transfer Complete event if we already have requests queued). Nevertheless, the driver is small and has been tested with all transfer types. If you really want to use this driver as is, that's your call. But I would suggest to Greg that he only takes this in if there is someone dedicated to maintain it. It's a really big codebase and whenever I change something on the gadget framework I will have to patch this beast too. If there's no maintainer assigned to it, who's willing to put effort in improving this driver in the long run, I think this will become yet another burden on the community. Still, it's your call. -- balbi signature.asc Description: Digital signature ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
[PATCH 00/14] Modifications for DWC OTG since v13
These patches are based on:http://patchwork.ozlabs.org/patch/89560/ After not getting any reply from developers, I started to do modifications for my platform (SPEAr1340). I have done modifications in such a way that all the code in driver/usb/dwc/ would be platform independent. I have tested this code for host/device/dma/slave mode. My fifo configuration is dedicated and dynamic. Changes since v13: 1. All the patches has been renamed from usb/ppc4xx to dwc/otg. Changes has been done, because these files are valid for other architecture also. 2. Parameter passing mechanism has been modified. Now, driver can also receive parameters from a platform device. 3. Next request is started after clear halt for all slave transfer There might be a situation, when there are data into fifo, EP is receiving IN token, but EP is disabled. In this case, EP will just stop and do nothing. So, for each transfer type start_next_request is enabled. 4. Dynamic fifo was configured by reading number of endpoints. Number of endpoints can be different than the number of dynamic fifo in the controller. Number of dynamic fifo is programmed in fifo_number parameter. Program fifos only for fifo_number of fifo. 5. If platform specific functions like phy_init and param_init is defined then call it. 6. Few define which are used for TX, were defined as RX in the code. All such definition has been replaced throughout the code. 7. There are some defined steps to disable an endpoint. It can not be disabled just by setting epdisable bit. Now, it is disabled as specification says. During stall of and endpoint, first stall bit should be set and then endpoint should be disabled. In dwc_otg_pcd_ep_disable, TX fifo must be flushed along with disabling of endpoint. In DMA mode if Fifo is flushed only once then, DMA will write data into it again. So, it must be flushed till DMA transfer is complete. 8. dwc_otg_ep_deactivate is called after dwc_otg_ep_disable. 9. platform_get_irq return -ENXIO , if requested IRQ is not found. So, return must be checked accordingly. 10. There is a struct (core_params) for all possible variable parameters of OTG. Earlier, these parameters were not modifiable for any platform data. It was reading from the HW and then was programming the same value again. Now, it has been modified. A platform code can modify these parameters. Struct definition has been moved from cil.h to a new file include/linux/usb/dwc_otg.h. Some of the parameters have been initialized with default values, while other with -1. A new parameter fifo_number has been added. This parameter is also defined in the specification. Otg code was reading number of endpoint from hwcfg regs for the purpose of this parameter, which was wrong. SW checks for only fifo_number fifos and not MAX_TX_FIFOS. Now , there is only one define for MAX_TX_FIFO and MAX_PERIO_FIFOS. As, per specification they can not be different. 11. In dwc_otg_pcd_handle_np_tx_fifo_empty_intr a different interrupt was cleared. Corrected it to clear NP TX Fifo empty interrupt. Re-enabling of interrupt was not correct when packets were still pending. This patch also corrects this behaviour. 12. Code checks for space availability in fifo during write_epmty_tx_fifo. It was writing only when space available was more than data to be written. So, it was writing 512 bytes only when 516 bytes space was available. Modified condition check from to =. 13. When TX fifo empty interrupt is received, transfer completion may not be true. Next transfer should only be started when transfer is complete, i.e. packet count in dieptsiz register has become zero. 14. Max number of EP increased to 16 15. There can be few devices like SPEAr1340, which issues a connection only after removal of soft disconnect. Default reset value is disconnected. It will not have any side effect even for the device where reset value is connected. 16. As per data sheet, when bit 31 of GRSTCTL is 1 then AHB is IDLE. Earlier function was waiting for this but to be 0, which was wrong. 17. Although name of file suggest that it is specific for PPC. But, all of its content are common for OTG and can be used for any platform. There were some compilation issued on ARM platform with this file. It has been removed by calling by proper functions. Pratyush Anand (4): dwc/otg: Structure declaration for shared data include/linux/usb/gadget.h : include linux/device.h for successful compilation usb/gadget/kconfig: added dwc otg as an option for peripheral controller arm/include/asm/io.h : added macros to read/write big/little endian register Tirumala Marri (10): dwc/otg: Add Register definitions dwc/otg: Add driver framework dwc/otg: Add Core
RE: [PATCH 00/14] Modifications for DWC OTG since v13
-Original Message- From: Pratyush Anand [mailto:pratyush.an...@st.com] Sent: Tuesday, August 30, 2011 4:58 AM To: linux-...@vger.kernel.org Cc: tma...@apm.com; linuxppc-dev@lists.ozlabs.org; fc...@apm.com; mmiesf...@apm.com; shiraz.has...@st.com; deepak.si...@st.com; vipulkumar.sa...@st.com; rajeev-dlh.ku...@st.com; vipin.ku...@st.com; bhupesh.sha...@st.com; viresh.ku...@st.com; amit.vi...@st.com; Pratyush Anand Subject: [PATCH 00/14] Modifications for DWC OTG since v13 These patches are based on:http://patchwork.ozlabs.org/patch/89560/ After not getting any reply from developers, I started to do modifications for my platform (SPEAr1340). I have done modifications in such a way that all the code in driver/usb/dwc/ would be platform independent. I have tested this code for host/device/dma/slave mode. My fifo configuration is dedicated and dynamic. [Tirumala Marri] We are working on our next release of patches. They should be coming out soon. ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH 00/14] Modifications for DWC OTG since v13
On Tue, Aug 30, 2011 at 8:57 PM, Tirumala Marri tma...@apm.com wrote: -Original Message- From: Pratyush Anand [mailto:pratyush.an...@st.com] Sent: Tuesday, August 30, 2011 4:58 AM To: linux-...@vger.kernel.org Cc: tma...@apm.com; linuxppc-dev@lists.ozlabs.org; fc...@apm.com; mmiesf...@apm.com; shiraz.has...@st.com; deepak.si...@st.com; vipulkumar.sa...@st.com; rajeev-dlh.ku...@st.com; vipin.ku...@st.com; bhupesh.sha...@st.com; viresh.ku...@st.com; amit.vi...@st.com; Pratyush Anand Subject: [PATCH 00/14] Modifications for DWC OTG since v13 These patches are based on:http://patchwork.ozlabs.org/patch/89560/ After not getting any reply from developers, I started to do modifications for my platform (SPEAr1340). I have done modifications in such a way that all the code in driver/usb/dwc/ would be platform independent. I have tested this code for host/device/dma/slave mode. My fifo configuration is dedicated and dynamic. [Tirumala Marri] We are working on our next release of patches. They should be coming out soon. Oh, thats great !! Actually, I did not get any reply of my previous mail on your patch release. I thought no one is maintaining., and so I sent them with my modifications, after testing them in all dev/host/otg mode. Regards Pratyush -- To unsubscribe from this list: send the line unsubscribe linux-usb in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH 00/14] Modifications for DWC OTG since v13
On Wed, Aug 31, 2011 at 12:46 AM, Pratyush Anand pratyush.an...@gmail.com wrote: On Tue, Aug 30, 2011 at 8:57 PM, Tirumala Marri tma...@apm.com wrote: -Original Message- From: Pratyush Anand [mailto:pratyush.an...@st.com] Sent: Tuesday, August 30, 2011 4:58 AM To: linux-...@vger.kernel.org Cc: tma...@apm.com; linuxppc-dev@lists.ozlabs.org; fc...@apm.com; mmiesf...@apm.com; shiraz.has...@st.com; deepak.si...@st.com; vipulkumar.sa...@st.com; rajeev-dlh.ku...@st.com; vipin.ku...@st.com; bhupesh.sha...@st.com; viresh.ku...@st.com; amit.vi...@st.com; Pratyush Anand Subject: [PATCH 00/14] Modifications for DWC OTG since v13 These patches are based on:http://patchwork.ozlabs.org/patch/89560/ After not getting any reply from developers, I started to do modifications for my platform (SPEAr1340). I have done modifications in such a way that all the code in driver/usb/dwc/ would be platform independent. I have tested this code for host/device/dma/slave mode. My fifo configuration is dedicated and dynamic. [Tirumala Marri] We are working on our next release of patches. They should be coming out soon. Oh, thats great !! Actually, I did not get any reply of my previous mail on your patch release. I thought no one is maintaining., and so I sent them with my modifications, after testing them in all dev/host/otg mode. Regards Pratyush Hi, Can you provide the git repo to test? and I wonder what's the difference between dwc (from you) and dwc3 (from Felipe Balbi). I think it's dwc targets for usb 2.0 and dwc3 for usb 3.0. Thank you, Kyungmin Park -- To unsubscribe from this list: send the line unsubscribe linux-usb in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line unsubscribe linux-usb in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
