Re: [PATCH 7/8] VMCI: dma dg: add support for DMA datagrams sends
Hi Jorgen, I love your patch! Perhaps something to improve: [auto build test WARNING on char-misc/char-misc-testing] [also build test WARNING on linux/master linus/master v5.17-rc2 next-20220202] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in https://git-scm.com/docs/git-format-patch] url: https://github.com/0day-ci/linux/commits/Jorgen-Hansen/VMCI-dma-dg-Add-support-for-DMA-datagrams/20220202-230034 base: https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc.git 7ab004dbcbee38b8a70798835d3ffcd97a985a5e config: i386-randconfig-s002 (https://download.01.org/0day-ci/archive/20220203/202202031004.smrs3z6m-...@intel.com/config) compiler: gcc-9 (Debian 9.3.0-22) 9.3.0 reproduce: # apt-get install sparse # sparse version: v0.6.4-dirty # https://github.com/0day-ci/linux/commit/303777a2a8daa11d529827395318bb698ddee57e git remote add linux-review https://github.com/0day-ci/linux git fetch --no-tags linux-review Jorgen-Hansen/VMCI-dma-dg-Add-support-for-DMA-datagrams/20220202-230034 git checkout 303777a2a8daa11d529827395318bb698ddee57e # save the config file to linux build tree mkdir build_dir make W=1 C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' O=build_dir ARCH=i386 SHELL=/bin/bash drivers/misc/vmw_vmci/ If you fix the issue, kindly add following tag as appropriate Reported-by: kernel test robot sparse warnings: (new ones prefixed by >>) drivers/misc/vmw_vmci/vmci_guest.c:106:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void const volatile [noderef] __iomem *addr @@ got char * @@ drivers/misc/vmw_vmci/vmci_guest.c:106:45: sparse: expected void const volatile [noderef] __iomem *addr drivers/misc/vmw_vmci/vmci_guest.c:106:45: sparse: got char * drivers/misc/vmw_vmci/vmci_guest.c:103:14: sparse: sparse: symbol 'vmci_read_reg' was not declared. Should it be static? drivers/misc/vmw_vmci/vmci_guest.c:113:44: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void volatile [noderef] __iomem *addr @@ got char * @@ drivers/misc/vmw_vmci/vmci_guest.c:113:44: sparse: expected void volatile [noderef] __iomem *addr drivers/misc/vmw_vmci/vmci_guest.c:113:44: sparse: got char * drivers/misc/vmw_vmci/vmci_guest.c:110:6: sparse: sparse: symbol 'vmci_write_reg' was not declared. Should it be static? >> drivers/misc/vmw_vmci/vmci_guest.c:118:5: sparse: sparse: symbol >> 'vmci_write_data' was not declared. Should it be static? drivers/misc/vmw_vmci/vmci_guest.c:554:27: sparse: sparse: incorrect type in assignment (different address spaces) @@ expected char *mmio_base @@ got void [noderef] __iomem * @@ drivers/misc/vmw_vmci/vmci_guest.c:554:27: sparse: expected char *mmio_base drivers/misc/vmw_vmci/vmci_guest.c:554:27: sparse: got void [noderef] __iomem * drivers/misc/vmw_vmci/vmci_guest.c:909:43: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void [noderef] __iomem * @@ got char *mmio_base @@ drivers/misc/vmw_vmci/vmci_guest.c:909:43: sparse: expected void [noderef] __iomem * drivers/misc/vmw_vmci/vmci_guest.c:909:43: sparse: got char *mmio_base Please review and possibly fold the followup patch. --- 0-DAY CI Kernel Test Service, Intel Corporation https://lists.01.org/hyperkitty/list/kbuild-...@lists.01.org ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
Re: [PATCH 7/8] VMCI: dma dg: add support for DMA datagrams sends
Hi Jorgen, I love your patch! Perhaps something to improve: [auto build test WARNING on char-misc/char-misc-testing] [also build test WARNING on linux/master linus/master v5.17-rc2 next-20220202] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in https://git-scm.com/docs/git-format-patch] url: https://github.com/0day-ci/linux/commits/Jorgen-Hansen/VMCI-dma-dg-Add-support-for-DMA-datagrams/20220202-230034 base: https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc.git 7ab004dbcbee38b8a70798835d3ffcd97a985a5e config: i386-randconfig-a013-20220131 (https://download.01.org/0day-ci/archive/20220203/202202030803.bfgzqtfl-...@intel.com/config) compiler: clang version 14.0.0 (https://github.com/llvm/llvm-project 6b1e844b69f15bb7dffaf9365cd2b355d2eb7579) reproduce (this is a W=1 build): wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross # https://github.com/0day-ci/linux/commit/303777a2a8daa11d529827395318bb698ddee57e git remote add linux-review https://github.com/0day-ci/linux git fetch --no-tags linux-review Jorgen-Hansen/VMCI-dma-dg-Add-support-for-DMA-datagrams/20220202-230034 git checkout 303777a2a8daa11d529827395318bb698ddee57e # save the config file to linux build tree mkdir build_dir COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=i386 SHELL=/bin/bash drivers/misc/vmw_vmci/ If you fix the issue, kindly add following tag as appropriate Reported-by: kernel test robot All warnings (new ones prefixed by >>): drivers/misc/vmw_vmci/vmci_guest.c:103:14: warning: no previous prototype for function 'vmci_read_reg' [-Wmissing-prototypes] unsigned int vmci_read_reg(struct vmci_guest_device *dev, u32 reg) ^ drivers/misc/vmw_vmci/vmci_guest.c:103:1: note: declare 'static' if the function is not intended to be used outside of this translation unit unsigned int vmci_read_reg(struct vmci_guest_device *dev, u32 reg) ^ static drivers/misc/vmw_vmci/vmci_guest.c:110:6: warning: no previous prototype for function 'vmci_write_reg' [-Wmissing-prototypes] void vmci_write_reg(struct vmci_guest_device *dev, u32 val, u32 reg) ^ drivers/misc/vmw_vmci/vmci_guest.c:110:1: note: declare 'static' if the function is not intended to be used outside of this translation unit void vmci_write_reg(struct vmci_guest_device *dev, u32 val, u32 reg) ^ static >> drivers/misc/vmw_vmci/vmci_guest.c:118:5: warning: no previous prototype for >> function 'vmci_write_data' [-Wmissing-prototypes] int vmci_write_data(struct vmci_guest_device *dev, struct vmci_datagram *dg) ^ drivers/misc/vmw_vmci/vmci_guest.c:118:1: note: declare 'static' if the function is not intended to be used outside of this translation unit int vmci_write_data(struct vmci_guest_device *dev, struct vmci_datagram *dg) ^ static 3 warnings generated. vim +/vmci_write_data +118 drivers/misc/vmw_vmci/vmci_guest.c 109 > 110 void vmci_write_reg(struct vmci_guest_device *dev, u32 val, u32 reg) 111 { 112 if (dev->mmio_base != NULL) 113 writel(val, dev->mmio_base + reg); 114 else 115 iowrite32(val, dev->iobase + reg); 116 } 117 > 118 int vmci_write_data(struct vmci_guest_device *dev, struct vmci_datagram > *dg) 119 { 120 int result; 121 122 if (dev->mmio_base != NULL) { 123 struct vmci_data_in_out_header *buffer_header = dev->tx_buffer; 124 u8 *dg_out_buffer = (u8 *)(buffer_header + 1); 125 126 if (VMCI_DG_SIZE(dg) > VMCI_MAX_DG_SIZE) 127 return VMCI_ERROR_INVALID_ARGS; 128 129 /* 130 * Initialize send buffer with outgoing datagram 131 * and set up header for inline data. Device will 132 * not access buffer asynchronously - only after 133 * the write to VMCI_DATA_OUT_LOW_ADDR. 134 */ 135 memcpy(dg_out_buffer, dg, VMCI_DG_SIZE(dg)); 136 buffer_header->opcode = 0; 137 buffer_header->size = VMCI_DG_SIZE(dg); 138 buffer_header->busy = 1; 139 140 vmci_write_reg(dev, lower_32_bits(dev->tx_buffer_base), 141 VMCI_DATA_OUT_LOW_ADDR); 142 143 /* Caller holds a spinlock, so cannot block. */ 144 spin_until_cond(buffer_header->busy == 0); 145 146 result = vmci_read_reg(vmci_dev_g, VMCI_RESULT_LOW_ADDR); 147 if (result == VMCI_SUCCESS) 148
Re: [PATCH 7/8] VMCI: dma dg: add support for DMA datagrams sends
Hi Jorgen, I love your patch! Perhaps something to improve: [auto build test WARNING on char-misc/char-misc-testing] [also build test WARNING on linux/master linus/master v5.17-rc2 next-20220202] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in https://git-scm.com/docs/git-format-patch] url: https://github.com/0day-ci/linux/commits/Jorgen-Hansen/VMCI-dma-dg-Add-support-for-DMA-datagrams/20220202-230034 base: https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc.git 7ab004dbcbee38b8a70798835d3ffcd97a985a5e config: x86_64-allyesconfig (https://download.01.org/0day-ci/archive/20220203/202202030100.5mr551y1-...@intel.com/config) compiler: gcc-9 (Debian 9.3.0-22) 9.3.0 reproduce (this is a W=1 build): # https://github.com/0day-ci/linux/commit/303777a2a8daa11d529827395318bb698ddee57e git remote add linux-review https://github.com/0day-ci/linux git fetch --no-tags linux-review Jorgen-Hansen/VMCI-dma-dg-Add-support-for-DMA-datagrams/20220202-230034 git checkout 303777a2a8daa11d529827395318bb698ddee57e # save the config file to linux build tree mkdir build_dir make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash drivers/misc/vmw_vmci/ If you fix the issue, kindly add following tag as appropriate Reported-by: kernel test robot All warnings (new ones prefixed by >>): drivers/misc/vmw_vmci/vmci_guest.c:103:14: warning: no previous prototype for 'vmci_read_reg' [-Wmissing-prototypes] 103 | unsigned int vmci_read_reg(struct vmci_guest_device *dev, u32 reg) | ^ drivers/misc/vmw_vmci/vmci_guest.c:110:6: warning: no previous prototype for 'vmci_write_reg' [-Wmissing-prototypes] 110 | void vmci_write_reg(struct vmci_guest_device *dev, u32 val, u32 reg) | ^~ >> drivers/misc/vmw_vmci/vmci_guest.c:118:5: warning: no previous prototype for >> 'vmci_write_data' [-Wmissing-prototypes] 118 | int vmci_write_data(struct vmci_guest_device *dev, struct vmci_datagram *dg) | ^~~ vim +/vmci_write_data +118 drivers/misc/vmw_vmci/vmci_guest.c 117 > 118 int vmci_write_data(struct vmci_guest_device *dev, struct vmci_datagram > *dg) 119 { 120 int result; 121 122 if (dev->mmio_base != NULL) { 123 struct vmci_data_in_out_header *buffer_header = dev->tx_buffer; 124 u8 *dg_out_buffer = (u8 *)(buffer_header + 1); 125 126 if (VMCI_DG_SIZE(dg) > VMCI_MAX_DG_SIZE) 127 return VMCI_ERROR_INVALID_ARGS; 128 129 /* 130 * Initialize send buffer with outgoing datagram 131 * and set up header for inline data. Device will 132 * not access buffer asynchronously - only after 133 * the write to VMCI_DATA_OUT_LOW_ADDR. 134 */ 135 memcpy(dg_out_buffer, dg, VMCI_DG_SIZE(dg)); 136 buffer_header->opcode = 0; 137 buffer_header->size = VMCI_DG_SIZE(dg); 138 buffer_header->busy = 1; 139 140 vmci_write_reg(dev, lower_32_bits(dev->tx_buffer_base), 141 VMCI_DATA_OUT_LOW_ADDR); 142 143 /* Caller holds a spinlock, so cannot block. */ 144 spin_until_cond(buffer_header->busy == 0); 145 146 result = vmci_read_reg(vmci_dev_g, VMCI_RESULT_LOW_ADDR); 147 if (result == VMCI_SUCCESS) 148 result = (int)buffer_header->result; 149 } else { 150 iowrite8_rep(dev->iobase + VMCI_DATA_OUT_ADDR, 151 dg, VMCI_DG_SIZE(dg)); 152 result = vmci_read_reg(vmci_dev_g, VMCI_RESULT_LOW_ADDR); 153 } 154 155 return result; 156 } 157 --- 0-DAY CI Kernel Test Service, Intel Corporation https://lists.01.org/hyperkitty/list/kbuild-...@lists.01.org ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
[PATCH 7/8] VMCI: dma dg: add support for DMA datagrams sends
Use DMA based send operation from the transmit buffer instead of the iowrite8_rep based datagram send when DMA datagrams are supported. The outgoing datagram is sent as inline data in the VMCI transmit buffer. Once the header has been configured, the send is initiated by writing the lower 32 bit of the buffer base address to the VMCI_DATA_OUT_LOW_ADDR register. Only then will the device process the header and the datagram itself. Following that, the driver busy waits (it isn't possible to sleep on the send path) for the header busy flag to change - indicating that the send is complete. Reviewed-by: Vishnu Dasa Signed-off-by: Jorgen Hansen --- drivers/misc/vmw_vmci/vmci_guest.c | 44 -- include/linux/vmw_vmci_defs.h | 34 +++ 2 files changed, 76 insertions(+), 2 deletions(-) diff --git a/drivers/misc/vmw_vmci/vmci_guest.c b/drivers/misc/vmw_vmci/vmci_guest.c index d0acb686b464..ca73a1913404 100644 --- a/drivers/misc/vmw_vmci/vmci_guest.c +++ b/drivers/misc/vmw_vmci/vmci_guest.c @@ -13,6 +13,7 @@ #include #include #include +#include #include #include #include @@ -114,6 +115,46 @@ void vmci_write_reg(struct vmci_guest_device *dev, u32 val, u32 reg) iowrite32(val, dev->iobase + reg); } +int vmci_write_data(struct vmci_guest_device *dev, struct vmci_datagram *dg) +{ + int result; + + if (dev->mmio_base != NULL) { + struct vmci_data_in_out_header *buffer_header = dev->tx_buffer; + u8 *dg_out_buffer = (u8 *)(buffer_header + 1); + + if (VMCI_DG_SIZE(dg) > VMCI_MAX_DG_SIZE) + return VMCI_ERROR_INVALID_ARGS; + + /* +* Initialize send buffer with outgoing datagram +* and set up header for inline data. Device will +* not access buffer asynchronously - only after +* the write to VMCI_DATA_OUT_LOW_ADDR. +*/ + memcpy(dg_out_buffer, dg, VMCI_DG_SIZE(dg)); + buffer_header->opcode = 0; + buffer_header->size = VMCI_DG_SIZE(dg); + buffer_header->busy = 1; + + vmci_write_reg(dev, lower_32_bits(dev->tx_buffer_base), + VMCI_DATA_OUT_LOW_ADDR); + + /* Caller holds a spinlock, so cannot block. */ + spin_until_cond(buffer_header->busy == 0); + + result = vmci_read_reg(vmci_dev_g, VMCI_RESULT_LOW_ADDR); + if (result == VMCI_SUCCESS) + result = (int)buffer_header->result; + } else { + iowrite8_rep(dev->iobase + VMCI_DATA_OUT_ADDR, +dg, VMCI_DG_SIZE(dg)); + result = vmci_read_reg(vmci_dev_g, VMCI_RESULT_LOW_ADDR); + } + + return result; +} + /* * VM to hypervisor call mechanism. We use the standard VMware naming * convention since shared code is calling this function as well. @@ -139,8 +180,7 @@ int vmci_send_datagram(struct vmci_datagram *dg) spin_lock_irqsave(&vmci_dev_spinlock, flags); if (vmci_dev_g) { - iowrite8_rep(vmci_dev_g->iobase + VMCI_DATA_OUT_ADDR, -dg, VMCI_DG_SIZE(dg)); + vmci_write_data(vmci_dev_g, dg); result = vmci_read_reg(vmci_dev_g, VMCI_RESULT_LOW_ADDR); } else { result = VMCI_ERROR_UNAVAILABLE; diff --git a/include/linux/vmw_vmci_defs.h b/include/linux/vmw_vmci_defs.h index 8bc37d8244a8..6fb663b36f72 100644 --- a/include/linux/vmw_vmci_defs.h +++ b/include/linux/vmw_vmci_defs.h @@ -110,6 +110,40 @@ enum { #define VMCI_MMIO_ACCESS_OFFSET((size_t)(128 * 1024)) #define VMCI_MMIO_ACCESS_SIZE ((size_t)(64 * 1024)) +/* + * For VMCI devices supporting the VMCI_CAPS_DMA_DATAGRAM capability, the + * sending and receiving of datagrams can be performed using DMA to/from + * a driver allocated buffer. + * Sending and receiving will be handled as follows: + * - when sending datagrams, the driver initializes the buffer where the + * data part will refer to the outgoing VMCI datagram, sets the busy flag + * to 1 and writes the address of the buffer to VMCI_DATA_OUT_HIGH_ADDR + * and VMCI_DATA_OUT_LOW_ADDR. Writing to VMCI_DATA_OUT_LOW_ADDR triggers + * the device processing of the buffer. When the device has processed the + * buffer, it will write the result value to the buffer and then clear the + * busy flag. + * - when receiving datagrams, the driver initializes the buffer where the + * data part will describe the receive buffer, clears the busy flag and + * writes the address of the buffer to VMCI_DATA_IN_HIGH_ADDR and + * VMCI_DATA_IN_LOW_ADDR. Writing to VMCI_DATA_IN_LOW_ADDR triggers the + * device processing of the buffer. The device will copy as many available + * datagrams into the buffer as possible, and then sets the busy flag. + * When the busy flag