e1000 VLAN offload emulation

PATCH Applied Alex Williamson 2 participants 4 messages Nov 19, 2008 – Nov 21, 2008 KVM

Share & tools

Alex Williamson

author

2008-11-19 20:37 UTC #0 lore raw

PATCH 1/1 e1000 VLAN offload emulation

We're currently ignoring the e1000 VLAN tagging, stripping and filtering features in the e1000 emulation. This patch adds backing for the relevant registers and provides a software implementation of the acceleration, such that a guest can make use of VLANs.

This is mostly (only?) useful for a guest on a bridge (not user mode networking). The only caveat beyond that is that you need to make sure the host NIC isn't doing it's own tagging, stripping, or filtering. This generally means the host NIC on the bridge should not be part of a VLAN. Thanks,

Alex

Signed-off-by Alex Williamson <[email protected]>

Patch against KVM, but applies w/ only one line of offset to qemu.

qemu/hw/e1000.c +73 -7

index e2cc7f1..7be87e4 100644
@@ -89,9 +89,12 @@ typedef struct E1000State_st {
     int check_rxov;
     struct e1000_tx {
         unsigned char header[256];
+        unsigned char vlan_header[4];
+        unsigned char vlan[4];
         unsigned char data[0x10000];
         uint16_t size;
         unsigned char sum_needed;
+        unsigned char vlan_needed;
         uint8_t ipcss;
         uint8_t ipcso;
         uint16_t ipcse;

@@ -128,7 +131,8 @@ enum {
     defreg(TDBAL),	defreg(TDH),	defreg(TDLEN),	defreg(TDT),
     defreg(TORH),	defreg(TORL),	defreg(TOTH),	defreg(TOTL),
     defreg(TPR),	defreg(TPT),	defreg(TXDCTL),	defreg(WUFC),
-    defreg(RA),		defreg(MTA),	defreg(CRCERRS),
+    defreg(RA),		defreg(MTA),	defreg(CRCERRS),defreg(VFTA),
+    defreg(VET),
 };
 
 enum { PHY_R = 1, PHY_W = 2, PHY_RW = PHY_R | PHY_W };

@@ -294,6 +298,31 @@ putsum(uint8_t *data, uint32_t n, uint32_t sloc, uint32_t css, uint32_t cse)
     }
 }
 
+static inline int
+vlan_enabled(E1000State *s)
+{
+    return ((s->mac_reg[CTRL] & E1000_CTRL_VME) != 0);
+}
+
+static inline int
+vlan_rx_filter_enabled(E1000State *s)
+{
+    return ((s->mac_reg[RCTL] & E1000_RCTL_VFE) != 0);
+}
+
+static inline int
+is_vlan_packet(E1000State *s, const uint8_t *buf)
+{
+    return (be16_to_cpup((uint16_t *)(buf + 12)) ==
+                le16_to_cpup((uint16_t *)(s->mac_reg + VET)));
+}
+
+static inline int
+is_vlan_txd(uint32_t txd_lower)
+{
+    return ((txd_lower & E1000_TXD_CMD_VLE) != 0);
+}
+
 static void
 xmit_seg(E1000State *s)
 {

@@ -336,7 +365,12 @@ xmit_seg(E1000State *s)
         putsum(tp->data, tp->size, tp->tucso, tp->tucss, tp->tucse);
     if (tp->sum_needed & E1000_TXD_POPTS_IXSM)
         putsum(tp->data, tp->size, tp->ipcso, tp->ipcss, tp->ipcse);
-    qemu_send_packet(s->vc, tp->data, tp->size);
+    if (tp->vlan_needed) {
+        memmove(tp->vlan, tp->data, 12);
+        memcpy(tp->data + 8, tp->vlan_header, 4);
+        qemu_send_packet(s->vc, tp->vlan, tp->size + 4);
+    } else
+        qemu_send_packet(s->vc, tp->data, tp->size);
     s->mac_reg[TPT]++;
     s->mac_reg[GPTC]++;
     n = s->mac_reg[TOTL];

@@ -383,6 +417,15 @@ process_tx_desc(E1000State *s, struct e1000_tx_desc *dp)
         // legacy descriptor
         tp->cptse = 0;
 
+    if (vlan_enabled(s) && is_vlan_txd(txd_lower) &&
+        (tp->cptse || txd_lower & E1000_TXD_CMD_EOP)) {
+        tp->vlan_needed = 1;
+        cpu_to_be16wu((uint16_t *)(tp->vlan_header),
+                      le16_to_cpup((uint16_t *)(s->mac_reg + VET)));
+        cpu_to_be16wu((uint16_t *)(tp->vlan_header + 2),
+                      le16_to_cpu(dp->upper.fields.special));
+    }
+        
     addr = le64_to_cpu(dp->buffer_addr);
     if (tp->tse && tp->cptse) {
         hdr = tp->hdr_len;

@@ -416,6 +459,7 @@ process_tx_desc(E1000State *s, struct e1000_tx_desc *dp)
         xmit_seg(s);
     tp->tso_frames = 0;
     tp->sum_needed = 0;
+    tp->vlan_needed = 0;
     tp->size = 0;
     tp->cptse = 0;
 }

@@ -482,6 +526,14 @@ receive_filter(E1000State *s, const uint8_t *buf, int size)
     static int mta_shift[] = {4, 3, 2, 0};
     uint32_t f, rctl = s->mac_reg[RCTL], ra[2], *rp;
 
+    if (is_vlan_packet(s, buf) && vlan_rx_filter_enabled(s)) {
+        uint16_t vid = be16_to_cpup((uint16_t *)(buf + 14));
+        uint32_t vfta = le32_to_cpup((uint32_t *)(s->mac_reg + VFTA) +
+                                     ((vid >> 5) & 0x7f));
+        if ((vfta & (1 << (vid & 0x1f))) == 0)
+            return 0;
+    }
+
     if (rctl & E1000_RCTL_UPE)			// promiscuous
         return 1;

@@ -536,6 +588,8 @@ e1000_receive(void *opaque, const uint8_t *buf, int size)
     target_phys_addr_t base;
     unsigned int n, rdt;
     uint32_t rdh_start;
+    uint16_t vlan_special = 0;
+    uint8_t vlan_status = 0, vlan_offset = 0;
 
     if (!(s->mac_reg[RCTL] & E1000_RCTL_EN))
         return;

@@ -549,6 +603,14 @@ e1000_receive(void *opaque, const uint8_t *buf, int size)
     if (!receive_filter(s, buf, size))
         return;
 
+    if (vlan_enabled(s) && is_vlan_packet(s, buf)) {
+        vlan_special = cpu_to_le16(be16_to_cpup((uint16_t *)(buf + 14)));
+        memmove((void *)(buf + 4), buf, 12);
+        vlan_status = E1000_RXD_STAT_VP;
+        vlan_offset = 4;
+        size -= 4;
+    }
+
     rdh_start = s->mac_reg[RDH];
     size += 4; // for the header
     do {

@@ -559,10 +621,11 @@ e1000_receive(void *opaque, const uint8_t *buf, int size)
         base = ((uint64_t)s->mac_reg[RDBAH] << 32) + s->mac_reg[RDBAL] +
                sizeof(desc) * s->mac_reg[RDH];
         cpu_physical_memory_read(base, (void *)&desc, sizeof(desc));
-        desc.status |= E1000_RXD_STAT_DD;
+        desc.special = vlan_special;
+        desc.status |= (vlan_status | E1000_RXD_STAT_DD);
         if (desc.buffer_addr) {
             cpu_physical_memory_write(le64_to_cpu(desc.buffer_addr),
-                                      (void *)buf, size);
+                                      (void *)(buf + vlan_offset), size);
             desc.length = cpu_to_le16(size);
             desc.status |= E1000_RXD_STAT_EOP|E1000_RXD_STAT_IXSM;
         } else // as per intel docs; skip descriptors with null buf addr

@@ -692,7 +755,7 @@ static uint32_t (*macreg_readops[])(E1000State *, int) = {
     getreg(WUFC),	getreg(TDT),	getreg(CTRL),	getreg(LEDCTL),
     getreg(MANC),	getreg(MDIC),	getreg(SWSM),	getreg(STATUS),
     getreg(TORL),	getreg(TOTL),	getreg(IMS),	getreg(TCTL),
-    getreg(RDH),	getreg(RDT),
+    getreg(RDH),	getreg(RDT),	getreg(VET),
 
     [TOTH] = mac_read_clr8,	[TORH] = mac_read_clr8,	[GPRC] = mac_read_clr4,
     [GPTC] = mac_read_clr4,	[TPR] = mac_read_clr4,	[TPT] = mac_read_clr4,

@@ -700,6 +763,7 @@ static uint32_t (*macreg_readops[])(E1000State *, int) = {
     [CRCERRS ... MPC] = &mac_readreg,
     [RA ... RA+31] = &mac_readreg,
     [MTA ... MTA+127] = &mac_readreg,
+    [VFTA ... VFTA+127] = &mac_readreg,
 };
 enum { NREADOPS = sizeof(macreg_readops) / sizeof(*macreg_readops) };

@@ -707,7 +771,7 @@ enum { NREADOPS = sizeof(macreg_readops) / sizeof(*macreg_readops) };
 static void (*macreg_writeops[])(E1000State *, int, uint32_t) = {
     putreg(PBA),	putreg(EERD),	putreg(SWSM),	putreg(WUFC),
     putreg(TDBAL),	putreg(TDBAH),	putreg(TXDCTL),	putreg(RDBAH),
-    putreg(RDBAL),	putreg(LEDCTL),
+    putreg(RDBAL),	putreg(LEDCTL), putreg(CTRL),	putreg(VET),
     [TDLEN] = set_dlen,	[RDLEN] = set_dlen,	[TCTL] = set_tctl,
     [TDT] = set_tctl,	[MDIC] = set_mdic,	[ICS] = set_ics,
     [TDH] = set_16bit,	[RDH] = set_16bit,	[RDT] = set_rdt,

@@ -715,6 +779,7 @@ static void (*macreg_writeops[])(E1000State *, int, uint32_t) = {
     [EECD] = set_eecd,	[RCTL] = set_rx_control,
     [RA ... RA+31] = &mac_writereg,
     [MTA ... MTA+127] = &mac_writereg,
+    [VFTA ... VFTA+127] = &mac_writereg,
 };
 enum { NWRITEOPS = sizeof(macreg_writeops) / sizeof(*macreg_writeops) };

@@ -789,13 +854,14 @@ static const int mac_regtosave[] = {
     LEDCTL,	MANC,	MDIC,	MPC,	PBA,	RCTL,	RDBAH,	RDBAL,	RDH,
     RDLEN,	RDT,	STATUS,	SWSM,	TCTL,	TDBAH,	TDBAL,	TDH,	TDLEN,
     TDT,	TORH,	TORL,	TOTH,	TOTL,	TPR,	TPT,	TXDCTL,	WUFC,
+    VET,
 };
 enum { MAC_NSAVE = sizeof mac_regtosave/sizeof *mac_regtosave };
 
 static const struct {
     int size;
     int array0;
-} mac_regarraystosave[] = { {32, RA}, {128, MTA} };
+} mac_regarraystosave[] = { {32, RA}, {128, MTA}, {128, VFTA} };
 enum { MAC_NARRAYS = sizeof mac_regarraystosave/sizeof *mac_regarraystosave };
 
 static void

Anthony Liguori

2008-11-20 23:11 UTC #1 lore raw

Quoted 11 lines · Alex Williamson

We're currently ignoring the e1000 VLAN tagging, stripping and filtering features in the e1000 emulation. This patch adds backing for the relevant registers and provides a software implementation of the acceleration, such that a guest can make use of VLANs.
This is mostly (only?) useful for a guest on a bridge (not user mode networking). The only caveat beyond that is that you need to make sure the host NIC isn't doing it's own tagging, stripping, or filtering. This generally means the host NIC on the bridge should not be part of a VLAN. Thanks,

Quoted 7 lines

Alex
Signed-off-by Alex Williamson <[email protected]>
--
Patch against KVM, but applies w/ only one line of offset to qemu.

Hi Alex,

The patch looks good but I don't like the idea of taking patches against kvm-userspace. It suggests that people have only tested in kvm-userspace which means it's not guaranteed to work (or even compile) in upstream QEMU. Please rebase against QEMU and resubmit.

Regards,

Anthony Liguori

Quoted 207 lines

qemu/hw/e1000.c +73 -8

index e2cc7f1..7be87e4 100644
@@ -89,9 +89,12 @@ typedef struct E1000State_st {
     int check_rxov;
     struct e1000_tx {
         unsigned char header[256];
+        unsigned char vlan_header[4];
+        unsigned char vlan[4];
         unsigned char data[0x10000];
         uint16_t size;
         unsigned char sum_needed;
+        unsigned char vlan_needed;
         uint8_t ipcss;
         uint8_t ipcso;
         uint16_t ipcse;

@@ -128,7 +131,8 @@ enum {
     defreg(TDBAL),	defreg(TDH),	defreg(TDLEN),	defreg(TDT),
     defreg(TORH),	defreg(TORL),	defreg(TOTH),	defreg(TOTL),
     defreg(TPR),	defreg(TPT),	defreg(TXDCTL),	defreg(WUFC),
-    defreg(RA),		defreg(MTA),	defreg(CRCERRS),
+    defreg(RA),		defreg(MTA),	defreg(CRCERRS),defreg(VFTA),
+    defreg(VET),
 };
 
 enum { PHY_R = 1, PHY_W = 2, PHY_RW = PHY_R | PHY_W };

@@ -294,6 +298,31 @@ putsum(uint8_t *data, uint32_t n, uint32_t sloc, uint32_t css, uint32_t cse)
     }
 }
 
+static inline int
+vlan_enabled(E1000State *s)
+{
+    return ((s->mac_reg[CTRL] & E1000_CTRL_VME) != 0);
+}
+
+static inline int
+vlan_rx_filter_enabled(E1000State *s)
+{
+    return ((s->mac_reg[RCTL] & E1000_RCTL_VFE) != 0);
+}
+
+static inline int
+is_vlan_packet(E1000State *s, const uint8_t *buf)
+{
+    return (be16_to_cpup((uint16_t *)(buf + 12)) ==
+                le16_to_cpup((uint16_t *)(s->mac_reg + VET)));
+}
+
+static inline int
+is_vlan_txd(uint32_t txd_lower)
+{
+    return ((txd_lower & E1000_TXD_CMD_VLE) != 0);
+}
+
 static void
 xmit_seg(E1000State *s)
 {

@@ -336,7 +365,12 @@ xmit_seg(E1000State *s)
         putsum(tp->data, tp->size, tp->tucso, tp->tucss, tp->tucse);
     if (tp->sum_needed & E1000_TXD_POPTS_IXSM)
         putsum(tp->data, tp->size, tp->ipcso, tp->ipcss, tp->ipcse);
-    qemu_send_packet(s->vc, tp->data, tp->size);
+    if (tp->vlan_needed) {
+        memmove(tp->vlan, tp->data, 12);
+        memcpy(tp->data + 8, tp->vlan_header, 4);
+        qemu_send_packet(s->vc, tp->vlan, tp->size + 4);
+    } else
+        qemu_send_packet(s->vc, tp->data, tp->size);
     s->mac_reg[TPT]++;
     s->mac_reg[GPTC]++;
     n = s->mac_reg[TOTL];

@@ -383,6 +417,15 @@ process_tx_desc(E1000State *s, struct e1000_tx_desc *dp)
         // legacy descriptor
         tp->cptse = 0;
 
+    if (vlan_enabled(s) && is_vlan_txd(txd_lower) &&
+        (tp->cptse || txd_lower & E1000_TXD_CMD_EOP)) {
+        tp->vlan_needed = 1;
+        cpu_to_be16wu((uint16_t *)(tp->vlan_header),
+                      le16_to_cpup((uint16_t *)(s->mac_reg + VET)));
+        cpu_to_be16wu((uint16_t *)(tp->vlan_header + 2),
+                      le16_to_cpu(dp->upper.fields.special));
+    }
+        
     addr = le64_to_cpu(dp->buffer_addr);
     if (tp->tse && tp->cptse) {
         hdr = tp->hdr_len;

@@ -416,6 +459,7 @@ process_tx_desc(E1000State *s, struct e1000_tx_desc *dp)
         xmit_seg(s);
     tp->tso_frames = 0;
     tp->sum_needed = 0;
+    tp->vlan_needed = 0;
     tp->size = 0;
     tp->cptse = 0;
 }

@@ -482,6 +526,14 @@ receive_filter(E1000State *s, const uint8_t *buf, int size)
     static int mta_shift[] = {4, 3, 2, 0};
     uint32_t f, rctl = s->mac_reg[RCTL], ra[2], *rp;
 
+    if (is_vlan_packet(s, buf) && vlan_rx_filter_enabled(s)) {
+        uint16_t vid = be16_to_cpup((uint16_t *)(buf + 14));
+        uint32_t vfta = le32_to_cpup((uint32_t *)(s->mac_reg + VFTA) +
+                                     ((vid >> 5) & 0x7f));
+        if ((vfta & (1 << (vid & 0x1f))) == 0)
+            return 0;
+    }
+
     if (rctl & E1000_RCTL_UPE)			// promiscuous
         return 1;

@@ -536,6 +588,8 @@ e1000_receive(void *opaque, const uint8_t *buf, int size)
     target_phys_addr_t base;
     unsigned int n, rdt;
     uint32_t rdh_start;
+    uint16_t vlan_special = 0;
+    uint8_t vlan_status = 0, vlan_offset = 0;
 
     if (!(s->mac_reg[RCTL] & E1000_RCTL_EN))
         return;

@@ -549,6 +603,14 @@ e1000_receive(void *opaque, const uint8_t *buf, int size)
     if (!receive_filter(s, buf, size))
         return;
 
+    if (vlan_enabled(s) && is_vlan_packet(s, buf)) {
+        vlan_special = cpu_to_le16(be16_to_cpup((uint16_t *)(buf + 14)));
+        memmove((void *)(buf + 4), buf, 12);
+        vlan_status = E1000_RXD_STAT_VP;
+        vlan_offset = 4;
+        size -= 4;
+    }
+
     rdh_start = s->mac_reg[RDH];
     size += 4; // for the header
     do {

@@ -559,10 +621,11 @@ e1000_receive(void *opaque, const uint8_t *buf, int size)
         base = ((uint64_t)s->mac_reg[RDBAH] << 32) + s->mac_reg[RDBAL] +
                sizeof(desc) * s->mac_reg[RDH];
         cpu_physical_memory_read(base, (void *)&desc, sizeof(desc));
-        desc.status |= E1000_RXD_STAT_DD;
+        desc.special = vlan_special;
+        desc.status |= (vlan_status | E1000_RXD_STAT_DD);
         if (desc.buffer_addr) {
             cpu_physical_memory_write(le64_to_cpu(desc.buffer_addr),
-                                      (void *)buf, size);
+                                      (void *)(buf + vlan_offset), size);
             desc.length = cpu_to_le16(size);
             desc.status |= E1000_RXD_STAT_EOP|E1000_RXD_STAT_IXSM;
         } else // as per intel docs; skip descriptors with null buf addr

@@ -692,7 +755,7 @@ static uint32_t (*macreg_readops[])(E1000State *, int) = {
     getreg(WUFC),	getreg(TDT),	getreg(CTRL),	getreg(LEDCTL),
     getreg(MANC),	getreg(MDIC),	getreg(SWSM),	getreg(STATUS),
     getreg(TORL),	getreg(TOTL),	getreg(IMS),	getreg(TCTL),
-    getreg(RDH),	getreg(RDT),
+    getreg(RDH),	getreg(RDT),	getreg(VET),
 
     [TOTH] = mac_read_clr8,	[TORH] = mac_read_clr8,	[GPRC] = mac_read_clr4,
     [GPTC] = mac_read_clr4,	[TPR] = mac_read_clr4,	[TPT] = mac_read_clr4,

@@ -700,6 +763,7 @@ static uint32_t (*macreg_readops[])(E1000State *, int) = {
     [CRCERRS ... MPC] = &mac_readreg,
     [RA ... RA+31] = &mac_readreg,
     [MTA ... MTA+127] = &mac_readreg,
+    [VFTA ... VFTA+127] = &mac_readreg,
 };
 enum { NREADOPS = sizeof(macreg_readops) / sizeof(*macreg_readops) };

@@ -707,7 +771,7 @@ enum { NREADOPS = sizeof(macreg_readops) / sizeof(*macreg_readops) };
 static void (*macreg_writeops[])(E1000State *, int, uint32_t) = {
     putreg(PBA),	putreg(EERD),	putreg(SWSM),	putreg(WUFC),
     putreg(TDBAL),	putreg(TDBAH),	putreg(TXDCTL),	putreg(RDBAH),
-    putreg(RDBAL),	putreg(LEDCTL),
+    putreg(RDBAL),	putreg(LEDCTL), putreg(CTRL),	putreg(VET),
     [TDLEN] = set_dlen,	[RDLEN] = set_dlen,	[TCTL] = set_tctl,
     [TDT] = set_tctl,	[MDIC] = set_mdic,	[ICS] = set_ics,
     [TDH] = set_16bit,	[RDH] = set_16bit,	[RDT] = set_rdt,

@@ -715,6 +779,7 @@ static void (*macreg_writeops[])(E1000State *, int, uint32_t) = {
     [EECD] = set_eecd,	[RCTL] = set_rx_control,
     [RA ... RA+31] = &mac_writereg,
     [MTA ... MTA+127] = &mac_writereg,
+    [VFTA ... VFTA+127] = &mac_writereg,
 };
 enum { NWRITEOPS = sizeof(macreg_writeops) / sizeof(*macreg_writeops) };

@@ -789,13 +854,14 @@ static const int mac_regtosave[] = {
     LEDCTL,	MANC,	MDIC,	MPC,	PBA,	RCTL,	RDBAH,	RDBAL,	RDH,
     RDLEN,	RDT,	STATUS,	SWSM,	TCTL,	TDBAH,	TDBAL,	TDH,	TDLEN,
     TDT,	TORH,	TORL,	TOTH,	TOTL,	TPR,	TPT,	TXDCTL,	WUFC,
+    VET,
 };
 enum { MAC_NSAVE = sizeof mac_regtosave/sizeof *mac_regtosave };
 
 static const struct {
     int size;
     int array0;
-} mac_regarraystosave[] = { {32, RA}, {128, MTA} };
+} mac_regarraystosave[] = { {32, RA}, {128, MTA}, {128, VFTA} };
 enum { MAC_NARRAYS = sizeof mac_regarraystosave/sizeof *mac_regarraystosave };
 
 static void
 
 
--

To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Alex Williamson

author

2008-11-21 15:41 UTC #2 lore raw

Quoted 4 lines · Anthony Liguori

The patch looks good but I don't like the idea of taking patches against kvm-userspace. It suggests that people have only tested in kvm-userspace which means it's not guaranteed to work (or even compile) in upstream QEMU. Please rebase against QEMU and resubmit.

Hi Anthony,

Thanks for reviewing. Here's a new one, created and tested against qemu 5765. Thanks,

Alex

e1000: Add VLAN tagging/stripping/filtering offload emulation

Signed-off-by Alex Williamson <[email protected]>

hw/e1000.c +73 -7

index b351119..71781e6 100644
@@ -88,9 +88,12 @@ typedef struct E1000State_st {
     int check_rxov;
     struct e1000_tx {
         unsigned char header[256];
+        unsigned char vlan_header[4];
+        unsigned char vlan[4];
         unsigned char data[0x10000];
         uint16_t size;
         unsigned char sum_needed;
+        unsigned char vlan_needed;
         uint8_t ipcss;
         uint8_t ipcso;
         uint16_t ipcse;

@@ -127,7 +130,8 @@ enum {
     defreg(TDBAL),	defreg(TDH),	defreg(TDLEN),	defreg(TDT),
     defreg(TORH),	defreg(TORL),	defreg(TOTH),	defreg(TOTL),
     defreg(TPR),	defreg(TPT),	defreg(TXDCTL),	defreg(WUFC),
-    defreg(RA),		defreg(MTA),	defreg(CRCERRS),
+    defreg(RA),		defreg(MTA),	defreg(CRCERRS),defreg(VFTA),
+    defreg(VET),
 };
 
 enum { PHY_R = 1, PHY_W = 2, PHY_RW = PHY_R | PHY_W };

@@ -293,6 +297,31 @@ putsum(uint8_t *data, uint32_t n, uint32_t sloc, uint32_t css, uint32_t cse)
     }
 }
 
+static inline int
+vlan_enabled(E1000State *s)
+{
+    return ((s->mac_reg[CTRL] & E1000_CTRL_VME) != 0);
+}
+
+static inline int
+vlan_rx_filter_enabled(E1000State *s)
+{
+    return ((s->mac_reg[RCTL] & E1000_RCTL_VFE) != 0);
+}
+
+static inline int
+is_vlan_packet(E1000State *s, const uint8_t *buf)
+{
+    return (be16_to_cpup((uint16_t *)(buf + 12)) ==
+                le16_to_cpup((uint16_t *)(s->mac_reg + VET)));
+}
+
+static inline int
+is_vlan_txd(uint32_t txd_lower)
+{
+    return ((txd_lower & E1000_TXD_CMD_VLE) != 0);
+}
+
 static void
 xmit_seg(E1000State *s)
 {

@@ -335,7 +364,12 @@ xmit_seg(E1000State *s)
         putsum(tp->data, tp->size, tp->tucso, tp->tucss, tp->tucse);
     if (tp->sum_needed & E1000_TXD_POPTS_IXSM)
         putsum(tp->data, tp->size, tp->ipcso, tp->ipcss, tp->ipcse);
-    qemu_send_packet(s->vc, tp->data, tp->size);
+    if (tp->vlan_needed) {
+        memmove(tp->vlan, tp->data, 12);
+        memcpy(tp->data + 8, tp->vlan_header, 4);
+        qemu_send_packet(s->vc, tp->vlan, tp->size + 4);
+    } else
+        qemu_send_packet(s->vc, tp->data, tp->size);
     s->mac_reg[TPT]++;
     s->mac_reg[GPTC]++;
     n = s->mac_reg[TOTL];

@@ -382,6 +416,15 @@ process_tx_desc(E1000State *s, struct e1000_tx_desc *dp)
         // legacy descriptor
         tp->cptse = 0;
 
+    if (vlan_enabled(s) && is_vlan_txd(txd_lower) &&
+        (tp->cptse || txd_lower & E1000_TXD_CMD_EOP)) {
+        tp->vlan_needed = 1;
+        cpu_to_be16wu((uint16_t *)(tp->vlan_header),
+                      le16_to_cpup((uint16_t *)(s->mac_reg + VET)));
+        cpu_to_be16wu((uint16_t *)(tp->vlan_header + 2),
+                      le16_to_cpu(dp->upper.fields.special));
+    }
+        
     addr = le64_to_cpu(dp->buffer_addr);
     if (tp->tse && tp->cptse) {
         hdr = tp->hdr_len;

@@ -415,6 +458,7 @@ process_tx_desc(E1000State *s, struct e1000_tx_desc *dp)
         xmit_seg(s);
     tp->tso_frames = 0;
     tp->sum_needed = 0;
+    tp->vlan_needed = 0;
     tp->size = 0;
     tp->cptse = 0;
 }

@@ -481,6 +525,14 @@ receive_filter(E1000State *s, const uint8_t *buf, int size)
     static int mta_shift[] = {4, 3, 2, 0};
     uint32_t f, rctl = s->mac_reg[RCTL], ra[2], *rp;
 
+    if (is_vlan_packet(s, buf) && vlan_rx_filter_enabled(s)) {
+        uint16_t vid = be16_to_cpup((uint16_t *)(buf + 14));
+        uint32_t vfta = le32_to_cpup((uint32_t *)(s->mac_reg + VFTA) +
+                                     ((vid >> 5) & 0x7f));
+        if ((vfta & (1 << (vid & 0x1f))) == 0)
+            return 0;
+    }
+
     if (rctl & E1000_RCTL_UPE)			// promiscuous
         return 1;

@@ -535,6 +587,8 @@ e1000_receive(void *opaque, const uint8_t *buf, int size)
     target_phys_addr_t base;
     unsigned int n, rdt;
     uint32_t rdh_start;
+    uint16_t vlan_special = 0;
+    uint8_t vlan_status = 0, vlan_offset = 0;
 
     if (!(s->mac_reg[RCTL] & E1000_RCTL_EN))
         return;

@@ -548,6 +602,14 @@ e1000_receive(void *opaque, const uint8_t *buf, int size)
     if (!receive_filter(s, buf, size))
         return;
 
+    if (vlan_enabled(s) && is_vlan_packet(s, buf)) {
+        vlan_special = cpu_to_le16(be16_to_cpup((uint16_t *)(buf + 14)));
+        memmove((void *)(buf + 4), buf, 12);
+        vlan_status = E1000_RXD_STAT_VP;
+        vlan_offset = 4;
+        size -= 4;
+    }
+
     rdh_start = s->mac_reg[RDH];
     size += 4; // for the header
     do {

@@ -558,10 +620,11 @@ e1000_receive(void *opaque, const uint8_t *buf, int size)
         base = ((uint64_t)s->mac_reg[RDBAH] << 32) + s->mac_reg[RDBAL] +
                sizeof(desc) * s->mac_reg[RDH];
         cpu_physical_memory_read(base, (void *)&desc, sizeof(desc));
-        desc.status |= E1000_RXD_STAT_DD;
+        desc.special = vlan_special;
+        desc.status |= (vlan_status | E1000_RXD_STAT_DD);
         if (desc.buffer_addr) {
             cpu_physical_memory_write(le64_to_cpu(desc.buffer_addr),
-                                      (void *)buf, size);
+                                      (void *)(buf + vlan_offset), size);
             desc.length = cpu_to_le16(size);
             desc.status |= E1000_RXD_STAT_EOP|E1000_RXD_STAT_IXSM;
         } else // as per intel docs; skip descriptors with null buf addr

@@ -691,7 +754,7 @@ static uint32_t (*macreg_readops[])(E1000State *, int) = {
     getreg(WUFC),	getreg(TDT),	getreg(CTRL),	getreg(LEDCTL),
     getreg(MANC),	getreg(MDIC),	getreg(SWSM),	getreg(STATUS),
     getreg(TORL),	getreg(TOTL),	getreg(IMS),	getreg(TCTL),
-    getreg(RDH),	getreg(RDT),
+    getreg(RDH),	getreg(RDT),	getreg(VET),
 
     [TOTH] = mac_read_clr8,	[TORH] = mac_read_clr8,	[GPRC] = mac_read_clr4,
     [GPTC] = mac_read_clr4,	[TPR] = mac_read_clr4,	[TPT] = mac_read_clr4,

@@ -699,6 +762,7 @@ static uint32_t (*macreg_readops[])(E1000State *, int) = {
     [CRCERRS ... MPC] = &mac_readreg,
     [RA ... RA+31] = &mac_readreg,
     [MTA ... MTA+127] = &mac_readreg,
+    [VFTA ... VFTA+127] = &mac_readreg,
 };
 enum { NREADOPS = sizeof(macreg_readops) / sizeof(*macreg_readops) };

@@ -706,7 +770,7 @@ enum { NREADOPS = sizeof(macreg_readops) / sizeof(*macreg_readops) };
 static void (*macreg_writeops[])(E1000State *, int, uint32_t) = {
     putreg(PBA),	putreg(EERD),	putreg(SWSM),	putreg(WUFC),
     putreg(TDBAL),	putreg(TDBAH),	putreg(TXDCTL),	putreg(RDBAH),
-    putreg(RDBAL),	putreg(LEDCTL),
+    putreg(RDBAL),	putreg(LEDCTL), putreg(CTRL),	putreg(VET),
     [TDLEN] = set_dlen,	[RDLEN] = set_dlen,	[TCTL] = set_tctl,
     [TDT] = set_tctl,	[MDIC] = set_mdic,	[ICS] = set_ics,
     [TDH] = set_16bit,	[RDH] = set_16bit,	[RDT] = set_rdt,

@@ -714,6 +778,7 @@ static void (*macreg_writeops[])(E1000State *, int, uint32_t) = {
     [EECD] = set_eecd,	[RCTL] = set_rx_control,
     [RA ... RA+31] = &mac_writereg,
     [MTA ... MTA+127] = &mac_writereg,
+    [VFTA ... VFTA+127] = &mac_writereg,
 };
 enum { NWRITEOPS = sizeof(macreg_writeops) / sizeof(*macreg_writeops) };

@@ -788,13 +853,14 @@ static const int mac_regtosave[] = {
     LEDCTL,	MANC,	MDIC,	MPC,	PBA,	RCTL,	RDBAH,	RDBAL,	RDH,
     RDLEN,	RDT,	STATUS,	SWSM,	TCTL,	TDBAH,	TDBAL,	TDH,	TDLEN,
     TDT,	TORH,	TORL,	TOTH,	TOTL,	TPR,	TPT,	TXDCTL,	WUFC,
+    VET,
 };
 enum { MAC_NSAVE = sizeof mac_regtosave/sizeof *mac_regtosave };
 
 static const struct {
     int size;
     int array0;
-} mac_regarraystosave[] = { {32, RA}, {128, MTA} };
+} mac_regarraystosave[] = { {32, RA}, {128, MTA}, {128, VFTA} };
 enum { MAC_NARRAYS = sizeof mac_regarraystosave/sizeof *mac_regarraystosave };
 
 static void

Anthony Liguori

Applied

2008-11-21 16:25 UTC #3 lore raw

Quoted 10 lines · Alex Williamson

Quoted 2 lines · Anthony Liguori
The patch looks good but I don't like the idea of taking patches against kvm-userspace. It suggests that people have only tested in kvm-userspace which means it's not guaranteed to work (or even compile) in upstream QEMU. Please rebase against QEMU and resubmit.
Hi Anthony,
Thanks for reviewing. Here's a new one, created and tested against qemu 5765. Thanks,

Applied. Thanks.

Regards,

Anthony Liguori