Skip to contentSkip to conversation

hw/usb/ehci: Add 64-bit descriptor addressing support

PATCH Superseded Jamin Lin 5 participants 69 messages Apr 16 – Apr 23 Arm/Aspeed machine support

v3 of 7 Superseded Latest is v7 →

lore
Share & tools
lore
author
#1 lore raw

PATCH v3 1/17 hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

In reply to Jamin Lin

The EHCIfstn structure, defined according to EHCI spec section 3.7, is currently unused in the EHCI implementation. Remove it to reduce unused code and improve maintainability.

Additionally, drop a block of disabled (#if 0) code in ehci_state_advqueue() that was never implemented. This code path is incomplete and has no functional impact.

No functional change.

---
 hw/usb/hcd-ehci.h |  7 -------
 hw/usb/hcd-ehci.c | 11 -----------
 2 files changed, 18 deletions(-)
hw/usb/hcd-ehci.h +0 -7
index 0ae8c06331..3aec0e047c 100644
@@ -202,13 +202,6 @@ typedef struct EHCIqh {
 #define BUFPTR_SBYTES_SH              5
 } EHCIqh;
 
-/*  EHCI spec version 1.0 Section 3.7
- */
-typedef struct EHCIfstn {
-    uint32_t next;                    /* Standard next link pointer */
-    uint32_t backptr;                 /* Standard next link pointer */
-} EHCIfstn;
-
 enum async_state {
     EHCI_ASYNC_NONE = 0,
     EHCI_ASYNC_INITIALIZED,
hw/usb/hcd-ehci.c +0 -11
index 57f930b099..5ea8461f70 100644
@@ -1756,17 +1756,6 @@ static int ehci_state_fetchsitd(EHCIState *ehci, int async)
 /* Section 4.10.2 - paragraph 3 */
 static int ehci_state_advqueue(EHCIQueue *q)
 {
-#if 0
-    /* TO-DO: 4.10.2 - paragraph 2
-     * if I-bit is set to 1 and QH is not active
-     * go to horizontal QH
-     */
-    if (I-bit set) {
-        ehci_set_state(ehci, async, EST_HORIZONTALQH);
-        goto out;
-    }
-#endif
-
     /*
      * want data and alt-next qTD is valid
      */
Signature · 1 line
2.43.0

Jamin Lin

author
#0 lore raw

v3 cover hw/usb/ehci: Add 64-bit descriptor addressing support

EHCI supports 64-bit addressing through the CTRLDSSEGMENT register, which provides the upper 32 bits of descriptor addresses when the controller advertises 64-bit capability.

Currently QEMU EHCI model only partially supports this functionality and descriptor addresses are effectively treated as 32-bit. This becomes problematic on systems where system memory is located above the 4GB boundary.

The Linux EHCI driver enables 64-bit addressing if the controller advertises the capability. During initialization it programs the segment register to zero:

https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci-hcd.c#L600

The driver also notes that descriptor structures allocated from the DMA pool use segment zero semantics. Descriptor memory is allocated using the DMA API and platforms may configure a 64-bit DMA mask, allowing descriptor memory to be placed above 4GB.

On AST2700 platforms, system DRAM is mapped at 0x400000000. As a result, descriptor addresses constructed directly from the EHCI registers do not match the actual system addresses used by the controller when accessing queue heads (QH) and queue element transfer descriptors (qTD).

This patch series implements full 64-bit descriptor addressing support in the EHCI emulation. Descriptor address handling is updated to use 64-bit values and the descriptor structures (QH, qTD, iTD and siTD) are extended to support the upper address bits provided by the segment register.

In addition, a descriptor-addr-offset property is introduced so platforms can apply an address translation when descriptor memory resides above the 4GB boundary.

The AST2700 machine uses this property to account for its DRAM mapping at 0x400000000 and enables 64-bit EHCI DMA addressing.

Test Result:
1. EHCI 32bits with ast2600-evb machine
Command line:
./build/qemu-system-arm \
-machine ast2600-evb \
-m 1G \
-drive file=image-bmc,if=mtd,format=raw \
-nographic \
-device usb-kbd,bus=usb-bus.1,id=mykbd \
-drive id=usbdisk,if=none,file=image0.ext4,format=raw \
-device usb-storage,bus=usb-bus.1,id=mystorage,drive=usbdisk
-snapshot \
-nographic
Result:
unable to initialize usb specBus 001 Device 001: ID 1d6b:0002 Linux 6.18.3-v00.08.01-g172b7e27a30d ehci_hcd EHCI Host Controller Bus 001 Device 002: ID 0627:0001 QEMU QEMU USB Keyboard
Bus 001 Device 003: ID 46f4:0001 QEMU QEMU USB HARDDRIVE
Bus 002 Device 001: ID 1d6b:0001 Linux 6.18.3-v00.08.01-g172b7e27a30d uhci_hcd Generic UHCI Host Controller

2. EHCI 64bits with ast2700a2-evb machine
Command line:
./build/qemu-system-aarch64 -M ast2700a2-evb -nographic\
-bios ast27x0_bootrom.bin \
-drive file=image-bmc,format=raw,if=mtd \
-snapshot \
-device usb-kbd,bus=usb-bus.3,id=mykbd \
-drive id=usbdisk,if=none,file=image0.ext4,format=raw \
-device usb-storage,bus=usb-bus.3,id=mystorage,drive=usbdisk
Result:

root@ast2700-default:~# lsusb
unable to initialize usb specBus 001 Device 001: ID 1d6b:0001 Linux 6.18.3-v00.08.01-g172b7e27a30d uhci_hcd Generic UHCI Host Controller
Bus 002 Device 001: ID 1d6b:0002 Linux 6.18.3-v00.08.01-g172b7e27a30d ehci_hcd EHCI Host Controller
Bus 002 Device 002: ID 0627:0001 QEMU QEMU USB Keyboard
Bus 002 Device 003: ID 46f4:0001 QEMU QEMU USB HARDDRIVE

v1
1. Fix checkpatch coding style issues
2. Implement 64-bit addressing for QH/qTD/iTD/siTD descriptors
3. Add descriptor address offset property
4. Enable 64-bit EHCI DMA addressing on AST2700
5. Configure descriptor address offset for AST2700

v2
1. Remove unused EHCIfstn structure and dead code
2. Replace fprintf(stderr, ...) with qemu_log_mask(LOG_GUEST_ERROR)
3. Replace DPRINTF debug logs with trace events
4. Add functional tests for USB EHCI on AST2600 and AST2700 A1/A2
5. Fix review issue

v3:
1. Add Migration version test function
2. Add EHCI 64-bit buffer pointer fields description in commit log

Jamin Lin (17):
hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code hw/usb/hcd-ehci.h: Fix coding style issues reported by checkpatch hw/usb/hcd-ehci.c: Fix coding style issues reported by checkpatch hw/usb/hcd-ehci.c: Replace fprintf(stderr, ...) with
qemu_log_mask(LOG_GUEST_ERROR)
hw/usb/hcd-ehci: Replace DPRINTF debug logs with trace events hw/usb/hcd-ehci: Change descriptor addresses to 64-bit hw/usb/hcd-ehci: Add property to advertise 64-bit addressing
capability
hw/usb/hcd-ehci: Reject CTRLDSSEGMENT writes without 64-bit capability hw/usb/hcd-ehci: Implement 64-bit QH descriptor addressing hw/usb/hcd-ehci: Implement 64-bit qTD descriptor addressing hw/usb/hcd-ehci: Implement 64-bit iTD descriptor addressing hw/usb/hcd-ehci: Implement 64-bit siTD descriptor addressing hw/usb/hcd-ehci: Add descriptor address offset property hw/arm/aspeed_ast27x0: Enable 64-bit EHCI DMA addressing hw/arm/aspeed_ast27x0: Set EHCI descriptor address offset tests/functional/arm/test_aspeed_ast2600_sdk: Add USB EHCI test for

    AST2600 SDK
  tests/functional/aarch64/test_aspeed_ast2700: Add USB EHCI test for
    AST2700 A1/A2
 hw/usb/hcd-ehci.h                             |  55 ++-
 hw/arm/aspeed_ast27x0.c                       |   5 +
 hw/usb/hcd-ehci-pci.c                         |   4 +
 hw/usb/hcd-ehci-sysbus.c                      |   4 +
 hw/usb/hcd-ehci.c                             | 378 +++++++++++-------
 hw/usb/trace-events                           |  25 +-
 .../aarch64/test_aspeed_ast2700a1.py          |   7 +
 .../aarch64/test_aspeed_ast2700a2.py          |   7 +
 .../functional/arm/test_aspeed_ast2600_sdk.py |   7 +
 9 files changed, 301 insertions(+), 191 deletions(-)
Signature · 1 line
2.43.0
author
#8 lore raw

PATCH v3 2/17 hw/usb/hcd-ehci.h: Fix coding style issues reported by checkpatch

No functional change.

---
 hw/usb/hcd-ehci.h | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)
hw/usb/hcd-ehci.h +11 -6
index 3aec0e047c..736407fd52 100644
@@ -43,7 +43,8 @@ typedef struct EHCIPacket EHCIPacket;
 typedef struct EHCIQueue EHCIQueue;
 typedef struct EHCIState EHCIState;
 
-/*  EHCI spec version 1.0 Section 3.3
+/*
+ * EHCI spec version 1.0 Section 3.3
  */
 typedef struct EHCIitd {
     uint32_t next;
@@ -74,7 +75,8 @@ typedef struct EHCIitd {
 #define ITD_BUFPTR_MULT_SH       0
 } EHCIitd;
 
-/*  EHCI spec version 1.0 Section 3.4
+/*
+ * EHCI spec version 1.0 Section 3.4
  */
 typedef struct EHCIsitd {
     uint32_t next;                  /* Standard next link pointer */
@@ -118,7 +120,8 @@ typedef struct EHCIsitd {
     uint32_t backptr;                 /* Standard next link pointer */
 } EHCIsitd;
 
-/*  EHCI spec version 1.0 Section 3.5
+/*
+ * EHCI spec version 1.0 Section 3.5
  */
 typedef struct EHCIqtd {
     uint32_t next;                    /* Standard next link pointer */
@@ -148,7 +151,8 @@ typedef struct EHCIqtd {
 #define QTD_BUFPTR_SH                 12
 } EHCIqtd;
 
-/*  EHCI spec version 1.0 Section 3.6
+/*
+ * EHCI spec version 1.0 Section 3.6
  */
 typedef struct EHCIqh {
     uint32_t next;                    /* Standard next link pointer */
@@ -230,7 +234,8 @@ struct EHCIQueue {
     int async;
     int transact_ctr;
 
-    /* cached data from guest - needs to be flushed
+    /*
+     * cached data from guest - needs to be flushed
      * when guest removes an entry (doorbell, handshake sequence)
      */
     EHCIqh qh;             /* copy of current QH (being worked on) */
@@ -268,7 +273,7 @@ struct EHCIState {
      */
     uint8_t caps[CAPA_SIZE];
     union {
-        uint32_t opreg[0x44/sizeof(uint32_t)];
+        uint32_t opreg[0x44 / sizeof(uint32_t)];
         struct {
             uint32_t usbcmd;
             uint32_t usbsts;
Signature · 1 line
2.43.0
author
#32 lore raw

PATCH v3 8/17 hw/usb/hcd-ehci: Reject CTRLDSSEGMENT writes without 64-bit capability

In reply to Jamin Lin

The EHCI CTRLDSSEGMENT register provides the upper 32 bits [63:32] used to form 64-bit addresses for EHCI control data structures. Per EHCI 1.0 spec section 2.3.5, when the HCCPARAMS 64-bit Addressing Capability bit is zero, CTRLDSSEGMENT is not used: software cannot write it and reads must return zero.

Add a capability check in the operational register write handler and reject guest writes to CTRLDSSEGMENT when 64-bit addressing is not enabled.

---
 hw/usb/hcd-ehci.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.c +8 -1
index 6409457186..b9e4251d62 100644
@@ -1108,7 +1108,14 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,
               "      is enabled and HC is enabled\n");
         }
         break;
-
+    case CTRLDSSEGMENT:
+        if (!s->caps_64bit_addr) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "ehci: write to CTRLDSSEGMENT while "
+                          "      64-bit addressing capability is disabled\n");
+            return;
+        }
+        break;
     case ASYNCLISTADDR:
         if (ehci_async_enabled(s)) {
             qemu_log_mask(LOG_GUEST_ERROR,
Signature · 1 line
2.43.0
author
#30 lore raw

PATCH v3 7/17 hw/usb/hcd-ehci: Add property to advertise 64-bit addressing capability

In reply to Jamin Lin

Introduce a new boolean property, "caps-64bit-addr", to control HCCPARAMS[0] (64-bit Addressing Capability).

When enabled, the EHCI controller advertises support for 64-bit address memory pointers as defined in the EHCI specification (Table 2-7, HCCPARAMS). This allows software to use the 64-bit data structure formats described in Appendix B.

When disabled (default), the controller reports 32-bit addressing capability and uses the standard 32-bit data structures.

---
 hw/usb/hcd-ehci.h        | 1 +
 hw/usb/hcd-ehci-pci.c    | 2 ++
 hw/usb/hcd-ehci-sysbus.c | 2 ++
 hw/usb/hcd-ehci.c        | 5 ++++-
 4 files changed, 9 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.h +1 -0
index 24eea9d23b..fe4c508c32 100644
@@ -256,6 +256,7 @@ struct EHCIState {
 
     /* properties */
     uint32_t maxframes;
+    bool caps_64bit_addr;
 
     /*
      *  EHCI spec version 1.0 Section 2.3
hw/usb/hcd-ehci-pci.c +2 -0
index 9febcc1031..2ea8549db9 100644
@@ -137,6 +137,8 @@ static void usb_ehci_pci_write_config(PCIDevice *dev, uint32_t addr,
 
 static const Property ehci_pci_properties[] = {
     DEFINE_PROP_UINT32("maxframes", EHCIPCIState, ehci.maxframes, 128),
+    DEFINE_PROP_BOOL("caps-64bit-addr", EHCIPCIState, ehci.caps_64bit_addr,
+                     false),
 };
 
 static const VMStateDescription vmstate_ehci_pci = {
hw/usb/hcd-ehci-sysbus.c +2 -0
index b31032bbf3..61215e9f3d 100644
@@ -34,6 +34,8 @@ static const Property ehci_sysbus_properties[] = {
     DEFINE_PROP_UINT32("maxframes", EHCISysBusState, ehci.maxframes, 128),
     DEFINE_PROP_BOOL("companion-enable", EHCISysBusState, ehci.companion_enable,
                      false),
+    DEFINE_PROP_BOOL("caps-64bit-addr", EHCISysBusState, ehci.caps_64bit_addr,
+                     false),
 };
 
 static void usb_ehci_sysbus_realize(DeviceState *dev, Error **errp)
hw/usb/hcd-ehci.c +4 -1
index 35173d27b4..6409457186 100644
@@ -2539,6 +2539,9 @@ void usb_ehci_realize(EHCIState *s, DeviceState *dev, Error **errp)
                    s->maxframes);
         return;
     }
+    if (s->caps_64bit_addr) {
+        s->caps[0x08] |= BIT(0);
+    }
 
     memory_region_add_subregion(&s->mem, s->capsbase, &s->mem_caps);
     memory_region_add_subregion(&s->mem, s->opregbase, &s->mem_opreg);
@@ -2598,7 +2601,7 @@ void usb_ehci_init(EHCIState *s, DeviceState *dev)
     s->caps[0x05] = 0x00;        /* No companion ports at present */
     s->caps[0x06] = 0x00;
     s->caps[0x07] = 0x00;
-    s->caps[0x08] = 0x80;        /* We can cache whole frame, no 64-bit */
+    s->caps[0x08] = 0x80;        /* We can cache whole frame */
     s->caps[0x0a] = 0x00;
     s->caps[0x0b] = 0x00;
 
Signature · 1 line
2.43.0
author
#58 lore raw

PATCH v3 15/17 hw/arm/aspeed_ast27x0: Set EHCI descriptor address offset

In reply to Jamin Lin

On AST2700 platforms, system DRAM is mapped above 4GB with the base address at 0x400000000.

The Linux EHCI driver programs the segment register to zero when 64-bit addressing is supported. As a result, descriptor addresses derived from the EHCI registers do not include the DRAM base address.

Descriptor memory is allocated through the DMA API with a 64-bit DMA mask, which allows descriptors to be placed in DRAM above 4GB. When running on AST2700, this means EHCI queue heads (QH) and queue element transfer descriptors (qTD) reside at addresses starting from 0x400000000.

Set the descriptor-addr-offset property to the DRAM base so the emulated EHCI controller can construct the correct descriptor addresses when accessing system memory.

---
 hw/arm/aspeed_ast27x0.c | 3 +++
 1 file changed, 3 insertions(+)
hw/arm/aspeed_ast27x0.c +3 -0
index 4a1f7cad73..3cdbf78ac1 100644
@@ -858,6 +858,9 @@ static void aspeed_soc_ast2700_realize(DeviceState *dev, Error **errp)
     for (i = 0; i < sc->ehcis_num; i++) {
         object_property_set_bool(OBJECT(&s->ehci[i]), "caps-64bit-addr", true,
                                  &error_abort);
+        object_property_set_int(OBJECT(&s->ehci[i]), "descriptor-addr-offset",
+                                sc->memmap[ASPEED_DEV_SDRAM],
+                                &error_abort);
         if (!sysbus_realize(SYS_BUS_DEVICE(&s->ehci[i]), errp)) {
             return;
         }
Signature · 1 line
2.43.0
author
#18 lore raw

PATCH v3 6/17 hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

In reply to Jamin Lin

Change internal EHCI descriptor addresses from uint32_t to uint64_t.

The following fields are updated:
- EHCIPacket::qtdaddr
- EHCIQueue::{qhaddr, qtdaddr}
- EHCIState::{a_fetch_addr, p_fetch_addr}

Update get_dwords() and put_dwords() to take 64-bit addresses and propagate the type change through the descriptor traversal paths.

Adjust NLPTR_GET() to operate on 64-bit values:

    #define NLPTR_GET(x) ((x) & ~0x1fULL)

so that link pointer masking works correctly when descriptor addresses exceed 32-bit space. The previous mask (0xffffffe0) implicitly truncated addresses to 32 bits.

This patch does not change the on-wire descriptor layout yet. It only removes the internal 32-bit address limit and prepares for later patches that will add full 64-bit QH/qTD/iTD/siTD support.

Update the EHCI trace-events prototypes for QH, qTD, iTD, and siTD to use uint64_t for the address argument and print it with PRIx64. This ensures full 64-bit addresses are shown in trace output and improves debugging of queue heads and transfer descriptors.

Migration compatibility:

The fetch address fields in EHCIState are extended from 32-bit to 64-bit, requiring a VMState version bump (v2 -> v3).

Backward compatibility is preserved by keeping the legacy 32-bit fields (a_fetch_addr_pre_v3, p_fetch_addr_pre_v3) for pre-v3 migration streams via VMSTATE_UINT32_TEST(), and introducing new 64-bit fields for v3+.

In post_load, pre-v3 state is promoted to the 64-bit fields.

No functional change.

---
 hw/usb/hcd-ehci.h   | 12 +++++----
 hw/usb/hcd-ehci.c   | 64 +++++++++++++++++++++++++++++----------------
 hw/usb/trace-events | 24 ++++++++---------
 3 files changed, 60 insertions(+), 40 deletions(-)
hw/usb/hcd-ehci.h +7 -5
index 4234591cb4..24eea9d23b 100644
@@ -208,7 +208,7 @@ struct EHCIPacket {
     QTAILQ_ENTRY(EHCIPacket) next;
 
     EHCIqtd qtd;           /* copy of current QTD (being worked on) */
-    uint32_t qtdaddr;      /* address QTD read from                 */
+    uint64_t qtdaddr;      /* address QTD read from                 */
 
     USBPacket packet;
     QEMUSGList sgl;
@@ -229,8 +229,8 @@ struct EHCIQueue {
      * when guest removes an entry (doorbell, handshake sequence)
      */
     EHCIqh qh;             /* copy of current QH (being worked on) */
-    uint32_t qhaddr;       /* address QH read from                 */
-    uint32_t qtdaddr;      /* address QTD read from                */
+    uint64_t qhaddr;       /* address QH read from                 */
+    uint64_t qtdaddr;      /* address QTD read from                */
     int last_pid;          /* pid of last packet executed          */
     USBDevice *dev;
     QTAILQ_HEAD(, EHCIPacket) packets;
@@ -294,8 +294,10 @@ struct EHCIState {
     EHCIQueueHead pqueues;
 
     /* which address to look at next */
-    uint32_t a_fetch_addr;
-    uint32_t p_fetch_addr;
+    uint32_t a_fetch_addr_pre_v3;
+    uint32_t p_fetch_addr_pre_v3;
+    uint64_t a_fetch_addr;
+    uint64_t p_fetch_addr;
 
     USBPacket ipacket;
     QEMUSGList isgl;
hw/usb/hcd-ehci.c +41 -23

196 lines not on this page · raw message

hw/usb/trace-events +12 -12
index 0d4318dcf1..8c90688bb3 100644
@@ -86,15 +86,15 @@ usb_ehci_portsc_write(uint32_t addr, uint32_t port, uint32_t val) "wr mmio 0x%04
 usb_ehci_portsc_change(uint32_t addr, uint32_t port, uint32_t new, uint32_t old) "ch mmio 0x%04x [port %d] = 0x%x (old: 0x%x)"
 usb_ehci_usbsts(const char *sts, int state) "usbsts %s %d"
 usb_ehci_state(const char *schedule, const char *state) "%s schedule %s"
-usb_ehci_qh_ptrs(void *q, uint32_t addr, uint32_t nxt, uint32_t c_qtd, uint32_t n_qtd, uint32_t a_qtd) "q %p - QH @ 0x%08x: next 0x%08x qtds 0x%08x,0x%08x,0x%08x"
-usb_ehci_qh_fields(uint32_t addr, int rl, int mplen, int eps, int ep, int devaddr) "QH @ 0x%08x - rl %d, mplen %d, eps %d, ep %d, dev %d"
-usb_ehci_qh_bits(uint32_t addr, int c, int h, int dtc, int i) "QH @ 0x%08x - c %d, h %d, dtc %d, i %d"
+usb_ehci_qh_ptrs(void *q, uint64_t addr, uint32_t nxt, uint32_t c_qtd, uint32_t n_qtd, uint32_t a_qtd) "q %p - QH @ 0x%" PRIx64 ": next 0x%08x qtds 0x%08x,0x%08x,0x%08x"
+usb_ehci_qh_fields(uint64_t addr, int rl, int mplen, int eps, int ep, int devaddr) "QH @ 0x%" PRIx64 " - rl %d, mplen %d, eps %d, ep %d, dev %d"
+usb_ehci_qh_bits(uint64_t addr, int c, int h, int dtc, int i) "QH @ 0x%" PRIx64 " - c %d, h %d, dtc %d, i %d"
 usb_ehci_qh_tbytes(uint32_t tbytes) "updating tbytes to %d"
-usb_ehci_qtd_ptrs(void *q, uint32_t addr, uint32_t nxt, uint32_t altnext) "q %p - QTD @ 0x%08x: next 0x%08x altnext 0x%08x"
-usb_ehci_qtd_fields(uint32_t addr, int tbytes, int cpage, int cerr, int pid) "QTD @ 0x%08x - tbytes %d, cpage %d, cerr %d, pid %d"
-usb_ehci_qtd_bits(uint32_t addr, int ioc, int active, int halt, int babble, int xacterr) "QTD @ 0x%08x - ioc %d, active %d, halt %d, babble %d, xacterr %d"
-usb_ehci_itd(uint32_t addr, uint32_t nxt, uint32_t mplen, uint32_t mult, uint32_t ep, uint32_t devaddr) "ITD @ 0x%08x: next 0x%08x - mplen %d, mult %d, ep %d, dev %d"
-usb_ehci_sitd(uint32_t addr, uint32_t nxt, uint32_t active) "ITD @ 0x%08x: next 0x%08x - active %d"
+usb_ehci_qtd_ptrs(void *q, uint64_t addr, uint32_t nxt, uint32_t altnext) "q %p - QTD @ 0x%" PRIx64 ": next 0x%08x altnext 0x%08x"
+usb_ehci_qtd_fields(uint64_t addr, int tbytes, int cpage, int cerr, int pid) "QTD @ 0x%" PRIx64 " - tbytes %d, cpage %d, cerr %d, pid %d"
+usb_ehci_qtd_bits(uint64_t addr, int ioc, int active, int halt, int babble, int xacterr) "QTD @ 0x%" PRIx64 " - ioc %d, active %d, halt %d, babble %d, xacterr %d"
+usb_ehci_itd(uint64_t addr, uint32_t nxt, uint32_t mplen, uint32_t mult, uint32_t ep, uint32_t devaddr) "ITD @ 0x%" PRIx64 ": next 0x%08x - mplen %d, mult %d, ep %d, dev %d"
+usb_ehci_sitd(uint64_t addr, uint32_t nxt, uint32_t active) "SITD @ 0x%" PRIx64 ": next 0x%08x - active %d"
 usb_ehci_port_attach(uint32_t port, const char *owner, const char *device) "attach port #%d, owner %s, device %s"
 usb_ehci_port_detach(uint32_t port, const char *owner) "detach port #%d, owner %s"
 usb_ehci_port_reset(uint32_t port, int enable) "reset port #%d - %d"
@@ -104,15 +104,15 @@ usb_ehci_port_resume(uint32_t port) "port #%d"
 usb_ehci_port_disable(uint32_t port) "port #%d"
 usb_ehci_queue_action(void *q, const char *action) "q %p: %s"
 usb_ehci_packet_action(void *q, void *p, const char *action) "q %p p %p: %s"
-usb_ehci_packet_submit(uint32_t qhaddr, uint32_t next, uint32_t qtdaddr, int pid, size_t len, int endp, int status, int actual_length) "qh=0x%x, next=0x%x, qtd=0x%x, pid=0x%x, len=%zd, endp=0x%x, status=%d, actual_length=%d"
+usb_ehci_packet_submit(uint64_t qhaddr, uint32_t next, uint64_t qtdaddr, int pid, size_t len, int endp, int status, int actual_length) "qh=0x%" PRIx64 ", next=0x%x, qtd=0x%" PRIx64 ", pid=0x%x, len=%zd, endp=0x%x, status=%d, actual_length=%d"
 usb_ehci_irq(uint32_t level, uint32_t frindex, uint32_t sts, uint32_t mask) "level %d, frindex 0x%04x, sts 0x%x, mask 0x%x"
 usb_ehci_guest_bug(const char *reason) "%s"
 usb_ehci_doorbell_ring(void) ""
 usb_ehci_doorbell_ack(void) ""
 usb_ehci_dma_error(void) ""
-usb_ehci_execute_complete(uint32_t qhaddr, uint32_t next, uint32_t qtdaddr, int status, int actual_length) "qhaddr=0x%x, next=0x%x, qtdaddr=0x%x, status=%d, actual_length=%d"
-usb_ehci_fetchqh_reclaim_done(uint32_t qhaddr) "QH 0x%08x H-bit set, reclamation status reset - done processing"
-usb_ehci_fetchqh_dbg(uint32_t qhaddr, uint32_t h, uint32_t halt, uint32_t active, uint32_t next) "QH 0x%08x (h 0x%x halt 0x%x active 0x%x) next 0x%08x"
+usb_ehci_execute_complete(uint64_t qhaddr, uint32_t next, uint64_t qtdaddr, int status, int actual_length) "qhaddr=0x%" PRIx64 ", next=0x%x, qtdaddr=0x%" PRIx64 ", status=%d, actual_length=%d"
+usb_ehci_fetchqh_reclaim_done(uint64_t qhaddr) "QH 0x%" PRIx64 " H-bit set, reclamation status reset - done processing"
+usb_ehci_fetchqh_dbg(uint64_t qhaddr, uint32_t h, uint32_t halt, uint32_t active, uint32_t next) "QH 0x%" PRIx64 " (h 0x%x halt 0x%x active 0x%x) next 0x%08x"
 usb_ehci_periodic_state_advance(uint32_t frame, uint32_t list, uint32_t entry) "frame=%d, list=0x%x, entry=0x%x"
 usb_ehci_skipped_uframes(uint64_t skipped_uframes) "skipped %" PRIu64 " uframes"
 usb_ehci_log(const char *msg) "%s"
Signature · 1 line
2.43.0
author
#35 lore raw

PATCH v3 9/17 hw/usb/hcd-ehci: Implement 64-bit QH descriptor addressing

In reply to Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register supplies the upper 32 bits which are concatenated with 32-bit link pointer fields to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

The current implementation assumes 32-bit QH descriptor addresses and directly uses link pointer values without applying the CTRLDSSEGMENT upper dword.

Introduce a helper, ehci_get_desc_addr(), to construct full 64-bit descriptor addresses when 64-bit capability is enabled. Update QH traversal paths (async list walk, horizontal QH link, and periodic schedule entry handling) to use the translated 64-bit addresses.

EHCI 64-bit buffer pointer fields are defined in Appendix B as split 32-bit low/high parts located at separate offsets, rather than a single contiguous 64-bit field. Therefore, the buffer pointers cannot be represented as uint64_t bufptr[5] without violating the descriptor layout defined by the specification.

Introduce ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] fields. Use this helper when calculating transfer buffer addresses so that data buffers above 4GB are correctly handled.

Also add bufptr_hi[5] to EHCIqh to support 64-bit buffer pointer fields as defined in Appendix B.

When 64-bit capability is disabled, descriptor addresses remain 32-bit and existing behaviour is unchanged.

Note: Similar split 64-bit buffer pointer handling is required for qTD, iTD and siTD descriptors, which will be addressed in follow-up changes.

---
 hw/usb/hcd-ehci.h   |  4 ++++
 hw/usb/hcd-ehci.c   | 48 ++++++++++++++++++++++++++++++++++-----------
 hw/usb/trace-events |  2 +-
 3 files changed, 42 insertions(+), 12 deletions(-)
hw/usb/hcd-ehci.h +4 -0
index fe4c508c32..773f376834 100644
@@ -141,6 +141,9 @@ typedef struct EHCIqtd {
 #define QTD_BUFPTR_SH                 12
 } EHCIqtd;
 
+/* QH overlay: altnext_qtd, token, bufptr[5], bufptr_hi[5] */
+#define EHCI_QH_OVERLAY_COUNT 12
+
 /*
  * EHCI spec version 1.0 Section 3.6
  */
@@ -194,6 +197,7 @@ typedef struct EHCIqh {
 #define BUFPTR_FRAMETAG_MASK          0x0000001f
 #define BUFPTR_SBYTES_MASK            0x00000fe0
 #define BUFPTR_SBYTES_SH              5
+    uint32_t bufptr_hi[5];
 } EHCIqh;
 
 enum async_state {
hw/usb/hcd-ehci.c +37 -11

100 lines not on this page · raw message

hw/usb/trace-events +1 -1
index 8c90688bb3..67249d69c2 100644
@@ -113,7 +113,7 @@ usb_ehci_dma_error(void) ""
 usb_ehci_execute_complete(uint64_t qhaddr, uint32_t next, uint64_t qtdaddr, int status, int actual_length) "qhaddr=0x%" PRIx64 ", next=0x%x, qtdaddr=0x%" PRIx64 ", status=%d, actual_length=%d"
 usb_ehci_fetchqh_reclaim_done(uint64_t qhaddr) "QH 0x%" PRIx64 " H-bit set, reclamation status reset - done processing"
 usb_ehci_fetchqh_dbg(uint64_t qhaddr, uint32_t h, uint32_t halt, uint32_t active, uint32_t next) "QH 0x%" PRIx64 " (h 0x%x halt 0x%x active 0x%x) next 0x%08x"
-usb_ehci_periodic_state_advance(uint32_t frame, uint32_t list, uint32_t entry) "frame=%d, list=0x%x, entry=0x%x"
+usb_ehci_periodic_state_advance(uint32_t frame, uint64_t list, uint64_t entry) "frame=%d, list=0x%" PRIx64 ", entry=0x%" PRIx64
 usb_ehci_skipped_uframes(uint64_t skipped_uframes) "skipped %" PRIu64 " uframes"
 usb_ehci_log(const char *msg) "%s"
 
Signature · 1 line
2.43.0
author
#13 lore raw

PATCH v3 4/17 hw/usb/hcd-ehci.c: Replace fprintf(stderr, ...) with qemu_log_mask(LOG_GUEST_ERROR)

In reply to Jamin Lin

Replace direct fprintf(stderr, ...) calls with qemu_log_mask() using LOG_GUEST_ERROR in the EHCI controller implementation.

Direct writes to stderr are discouraged in QEMU as they bypass the logging framework and cannot be filtered or controlled at runtime. Using qemu_log_mask() ensures that guest-visible errors are properly categorized and can be managed via QEMU logging options.

All affected messages correspond to guest-triggerable error conditions, so LOG_GUEST_ERROR is used consistently.

This change improves integration with QEMU's logging infrastructure and aligns with upstream coding guidelines.

---
 hw/usb/hcd-ehci.c | 43 +++++++++++++++++++++++++------------------
 1 file changed, 25 insertions(+), 18 deletions(-)
hw/usb/hcd-ehci.c +25 -18
index 9b92384bac..bebd137f94 100644
@@ -36,6 +36,7 @@
 #include "qemu/error-report.h"
 #include "qemu/main-loop.h"
 #include "system/runstate.h"
+#include "qemu/log.h"
 
 #define FRAME_TIMER_FREQ 1000
 #define FRAME_TIMER_NS   (NANOSECONDS_PER_SECOND / FRAME_TIMER_FREQ)
@@ -425,7 +426,7 @@ static int ehci_get_pid(EHCIqtd *qtd)
     case 2:
         return USB_TOKEN_SETUP;
     default:
-        fprintf(stderr, "bad token\n");
+        qemu_log_mask(LOG_GUEST_ERROR, "bad token\n");
         return 0;
     }
 }
@@ -532,7 +533,7 @@ static void ehci_free_packet(EHCIPacket *p)
     }
     if (p->async == EHCI_ASYNC_FINISHED &&
             p->packet.status == USB_RET_SUCCESS) {
-        fprintf(stderr,
+        qemu_log_mask(LOG_GUEST_ERROR,
                 "EHCI: Dropping completed packet from halted %s ep %02X\n",
                 (p->pid == USB_TOKEN_IN) ? "in" : "out",
                 get_field(p->queue->qh.epchar, QH_EPCHAR_EP));
@@ -1042,8 +1043,9 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,
 
         /* not supporting dynamic frame list size at the moment */
         if ((val & USBCMD_FLS) && !(s->usbcmd & USBCMD_FLS)) {
-            fprintf(stderr, "attempt to set frame list size -- value %d\n",
-                    (int)val & USBCMD_FLS);
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "attempt to set frame list size -- value %d\n",
+                          (int)val & USBCMD_FLS);
             val &= ~USBCMD_FLS;
         }
 
@@ -1101,7 +1103,7 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,
 
     case PERIODICLISTBASE:
         if (ehci_periodic_enabled(s)) {
-            fprintf(stderr,
+            qemu_log_mask(LOG_GUEST_ERROR,
               "ehci: PERIODIC list base register set while periodic schedule\n"
               "      is enabled and HC is enabled\n");
         }
@@ -1109,7 +1111,7 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,
 
     case ASYNCLISTADDR:
         if (ehci_async_enabled(s)) {
-            fprintf(stderr,
+            qemu_log_mask(LOG_GUEST_ERROR,
               "ehci: ASYNC list address register set while async schedule\n"
               "      is enabled and HC is enabled\n");
         }
@@ -1200,7 +1202,7 @@ static int ehci_init_transfer(EHCIPacket *p)
 
     while (bytes > 0) {
         if (cpage > 4) {
-            fprintf(stderr, "cpage out of range (%u)\n", cpage);
+            qemu_log_mask(LOG_GUEST_ERROR, "cpage out of range (%u)\n", cpage);
             qemu_sglist_destroy(&p->sgl);
             return -1;
         }
@@ -1306,7 +1308,8 @@ static void ehci_execute_complete(EHCIQueue *q)
         break;
     default:
         /* should not be triggerable */
-        fprintf(stderr, "USB invalid response %d\n", p->packet.status);
+        qemu_log_mask(LOG_GUEST_ERROR, "USB invalid response %d\n",
+                      p->packet.status);
         g_assert_not_reached();
     }
 
@@ -1354,7 +1357,7 @@ static int ehci_execute(EHCIPacket *p, const char *action)
            p->async == EHCI_ASYNC_INITIALIZED);
 
     if (!(p->qtd.token & QTD_TOKEN_ACTIVE)) {
-        fprintf(stderr, "Attempting to execute inactive qtd\n");
+        qemu_log_mask(LOG_GUEST_ERROR, "Attempting to execute inactive qtd\n");
         return -1;
     }
 
@@ -1395,7 +1398,8 @@ static int ehci_execute(EHCIPacket *p, const char *action)
             p->packet.actual_length);
 
     if (p->packet.actual_length > BUFF_SIZE) {
-        fprintf(stderr, "ret from usb_handle_packet > BUFF_SIZE\n");
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "ret from usb_handle_packet > BUFF_SIZE\n");
         return -1;
     }
 
@@ -1479,8 +1483,9 @@ static int ehci_process_itd(EHCIState *ehci,
             case USB_RET_SUCCESS:
                 break;
             default:
-                fprintf(stderr, "Unexpected iso usb result: %d\n",
-                        ehci->ipacket.status);
+                qemu_log_mask(LOG_GUEST_ERROR,
+                              "Unexpected iso usb result: %d\n",
+                              ehci->ipacket.status);
                 /* Fall through */
             case USB_RET_IOERROR:
             case USB_RET_NODEV:
@@ -1584,7 +1589,8 @@ static int ehci_state_fetchentry(EHCIState *ehci, int async)
 
     /* section 4.8, only QH in async schedule */
     if (async && (NLPTR_TYPE_GET(entry) != NLPTR_TYPE_QH)) {
-        fprintf(stderr, "non queue head request in async schedule\n");
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "non queue head request in async schedule\n");
         return -1;
     }
 
@@ -1606,7 +1612,7 @@ static int ehci_state_fetchentry(EHCIState *ehci, int async)
 
     default:
         /* TODO: handle FSTN type */
-        fprintf(stderr, "FETCHENTRY: entry at %X is of type %u "
+        qemu_log_mask(LOG_GUEST_ERROR, "FETCHENTRY: entry at %X is of type %u "
                 "which is not supported yet\n", entry, NLPTR_TYPE_GET(entry));
         return -1;
     }
@@ -2112,13 +2118,14 @@ static void ehci_advance_state(EHCIState *ehci, int async)
             break;
 
         default:
-            fprintf(stderr, "Bad state!\n");
+            qemu_log_mask(LOG_GUEST_ERROR, "Bad state!\n");
             g_assert_not_reached();
         }
 
         if (again < 0 || itd_count > 16) {
             /* TODO: notify guest (raise HSE irq?) */
-            fprintf(stderr, "processing error - resetting ehci HC\n");
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "processing error - resetting ehci HC\n");
             ehci_reset(ehci);
             again = 0;
         }
@@ -2175,7 +2182,7 @@ static void ehci_advance_async_state(EHCIState *ehci)
 
     default:
         /* this should only be due to a developer mistake */
-        fprintf(stderr, "ehci: Bad asynchronous state %d. "
+        qemu_log_mask(LOG_GUEST_ERROR, "ehci: Bad asynchronous state %d. "
                 "Resetting to active\n", ehci->astate);
         g_assert_not_reached();
     }
@@ -2225,7 +2232,7 @@ static void ehci_advance_periodic_state(EHCIState *ehci)
 
     default:
         /* this should only be due to a developer mistake */
-        fprintf(stderr, "ehci: Bad periodic state %d. "
+        qemu_log_mask(LOG_GUEST_ERROR, "ehci: Bad periodic state %d. "
                 "Resetting to active\n", ehci->pstate);
         g_assert_not_reached();
     }
Signature · 1 line
2.43.0
author
#61 lore raw

PATCH v3 16/17 tests/functional/arm/test_aspeed_ast2600_sdk: Add USB EHCI test for AST2600 SDK

In reply to Jamin Lin

Add a functional test to verify USB EHCI support on the AST2600 SDK machine by attaching a USB keyboard device and checking its enumeration via lsusb.

This introduces a helper routine that runs lsusb in the guest and validates that the emulated "QEMU USB Keyboard" is detected.

---
 tests/functional/arm/test_aspeed_ast2600_sdk.py | 7 +++++++
 1 file changed, 7 insertions(+)
tests/functional/arm/test_aspeed_ast2600_sdk.py +7 -0
index d787e90d10..01548dd135 100755
@@ -40,6 +40,11 @@ def do_ast2600_i3c_test(self):
             'i3ctransfer -d /dev/bus/i3c/5-1234567890ab -r 8 | grep 0x | xargs',
             '0x12 0x34 0x56 0x78 0x90 0xab 0xcd 0xef')
 
+    def do_ast2600_usb_ehci_test(self):
+        exec_command_and_wait_for_pattern(self,
+            'lsusb',
+            'QEMU QEMU USB Keyboard')
+
     def test_arm_ast2600_evb_sdk(self):
         self.set_machine('ast2600-evb')
         self.require_netdev('user')
@@ -54,6 +59,7 @@ def test_arm_ast2600_evb_sdk(self):
         self.vm.add_args('-netdev', 'user,id=net1')
         self.vm.add_args('-device',
             'mock-i3c-target,bus=dw.i3c.5,pid=0xab9078563412')
+        self.vm.add_args('-device', 'usb-kbd,bus=usb-bus.1')
         self.do_test_arm_aspeed_sdk_start(
             self.scratch_file("ast2600-default-image", "image-bmc"))
 
@@ -81,6 +87,7 @@ def test_arm_ast2600_evb_sdk(self):
              '/sbin/hwclock -f /dev/rtc1', year)
         self.do_ast2600_pcie_test()
         self.do_ast2600_i3c_test()
+        self.do_ast2600_usb_ehci_test()
 
 
 if __name__ == '__main__':
Signature · 1 line
2.43.0
author
#66 lore raw

PATCH v3 17/17 tests/functional/aarch64/test_aspeed_ast2700: Add USB EHCI test for AST2700 A1/A2

In reply to Jamin Lin

Add a functional test to verify USB EHCI support on the AST2700 A2/A1 by attaching a USB keyboard device and checking its enumeration via lsusb.

This introduces a helper routine that runs lsusb in the guest and validates that the emulated "QEMU USB Keyboard" is detected.

---
 tests/functional/aarch64/test_aspeed_ast2700a1.py | 7 +++++++
 tests/functional/aarch64/test_aspeed_ast2700a2.py | 7 +++++++
 2 files changed, 14 insertions(+)
tests/functional/aarch64/test_aspeed_ast2700a1.py +7 -0
index b0c08854da..42b8b200cf 100755
@@ -118,6 +118,11 @@ def do_ast2700_pcie_test(self):
             'ip addr show dev eth2',
             'inet 10.0.2.15/24')
 
+    def do_ast2700_usb_ehci_test(self):
+        exec_command_and_wait_for_pattern(self,
+            'lsusb',
+            'QEMU QEMU USB Keyboard')
+
     def start_ast2700_test(self, name, bus_id):
         num_cpu = 4
         load_images_list = [
@@ -161,10 +166,12 @@ def test_aarch64_ast2700a1_evb_sdk_v11_01(self):
         self.archive_extract(self.ASSET_SDK_V1101_AST2700A1)
         self.vm.add_args('-device', 'e1000e,netdev=net1,bus=pcie.2')
         self.vm.add_args('-netdev', 'user,id=net1')
+        self.vm.add_args('-device', 'usb-kbd,bus=usb-bus.3')
         self.start_ast2700_test('ast2700-a1-image', 1)
         self.verify_openbmc_boot_and_login('ast2700-a1')
         self.do_ast2700_i2c_test(1)
         self.do_ast2700_pcie_test()
+        self.do_ast2700_usb_ehci_test()
 
     def test_aarch64_ast2700a1_evb_sdk_vbootrom_v11_01(self):
         self.set_machine('ast2700a1-evb')
tests/functional/aarch64/test_aspeed_ast2700a2.py +7 -0
index ed414999f4..40e0ad81ed 100755
@@ -121,6 +121,11 @@ def do_ast2700_pcie_test(self):
             'ip addr show dev eth2',
             'inet 10.0.2.15/24')
 
+    def do_ast2700_usb_ehci_test(self):
+        exec_command_and_wait_for_pattern(self,
+            'lsusb',
+            'QEMU QEMU USB Keyboard')
+
     def start_ast2700_test(self, name, bus_id):
         num_cpu = 4
         load_images_list = [
@@ -164,10 +169,12 @@ def test_aarch64_ast2700a2_evb_sdk_v11_01(self):
         self.archive_extract(self.ASSET_SDK_V1101_AST2700A2)
         self.vm.add_args('-device', 'e1000e,netdev=net1,bus=pcie.2')
         self.vm.add_args('-netdev', 'user,id=net1')
+        self.vm.add_args('-device', 'usb-kbd,bus=usb-bus.3')
         self.start_ast2700_test('ast2700-default-image', 1)
         self.verify_openbmc_boot_and_login('ast2700-default')
         self.do_ast2700_i2c_test(1)
         self.do_ast2700_pcie_test()
+        self.do_ast2700_usb_ehci_test()
 
     def test_aarch64_ast2700a2_evb_sdk_vbootrom_v11_01(self):
         self.set_machine('ast2700a2-evb')
Signature · 1 line
2.43.0
author
#56 lore raw

PATCH v3 14/17 hw/arm/aspeed_ast27x0: Enable 64-bit EHCI DMA addressing

In reply to Jamin Lin

AST2700 supports a 64-bit DRAM address space. Therefore, DMA transactions must be capable of accessing 64-bit addresses.

Enable the "caps-64bit-addr" property for the EHCI controllers on AST2700 so that USB DMA operations can correctly handle 64-bit memory addresses.

---
 hw/arm/aspeed_ast27x0.c | 2 ++
 1 file changed, 2 insertions(+)
hw/arm/aspeed_ast27x0.c +2 -0
index 87dcb82e1b..4a1f7cad73 100644
@@ -856,6 +856,8 @@ static void aspeed_soc_ast2700_realize(DeviceState *dev, Error **errp)
 
     /* EHCI */
     for (i = 0; i < sc->ehcis_num; i++) {
+        object_property_set_bool(OBJECT(&s->ehci[i]), "caps-64bit-addr", true,
+                                 &error_abort);
         if (!sysbus_realize(SYS_BUS_DEVICE(&s->ehci[i]), errp)) {
             return;
         }
Signature · 1 line
2.43.0
author
#10 lore raw

PATCH v3 3/17 hw/usb/hcd-ehci.c: Fix coding style issues reported by checkpatch

In reply to Jamin Lin

No functional change.

---
 hw/usb/hcd-ehci.c | 126 +++++++++++++++++++++++++---------------------
 1 file changed, 69 insertions(+), 57 deletions(-)
hw/usb/hcd-ehci.c +69 -57
index 5ea8461f70..9b92384bac 100644
@@ -41,21 +41,23 @@
 #define FRAME_TIMER_NS   (NANOSECONDS_PER_SECOND / FRAME_TIMER_FREQ)
 #define UFRAME_TIMER_NS  (FRAME_TIMER_NS / 8)
 
-#define NB_MAXINTRATE    8        // Max rate at which controller issues ints
-#define BUFF_SIZE        5*4096   // Max bytes to transfer per transaction
-#define MAX_QH           100      // Max allowable queue heads in a chain
+#define NB_MAXINTRATE    8      /* Max rate at which controller issues ints */
+#define BUFF_SIZE        (5 * 4096) /* Max bytes to transfer per transaction */
+#define MAX_QH           100      /* Max allowable queue heads in a chain */
 #define MIN_UFR_PER_TICK 24       /* Min frames to process when catching up */
 #define PERIODIC_ACTIVE  512      /* Micro-frames */
 
-/*  Internal periodic / asynchronous schedule state machine states
+/*
+ * Internal periodic / asynchronous schedule state machine states
  */
 typedef enum {
     EST_INACTIVE = 1000,
     EST_ACTIVE,
     EST_EXECUTING,
     EST_SLEEPING,
-    /*  The following states are internal to the state machine function
-    */
+    /*
+     * The following states are internal to the state machine function
+     */
     EST_WAITLISTHEAD,
     EST_FETCHENTRY,
     EST_FETCHQH,
@@ -71,13 +73,13 @@ typedef enum {
 /* macros for accessing fields within next link pointer entry */
 #define NLPTR_GET(x)             ((x) & 0xffffffe0)
 #define NLPTR_TYPE_GET(x)        (((x) >> 1) & 3)
-#define NLPTR_TBIT(x)            ((x) & 1)  // 1=invalid, 0=valid
+#define NLPTR_TBIT(x)            ((x) & 1)  /* 1=invalid, 0=valid */
 
 /* link pointer types */
-#define NLPTR_TYPE_ITD           0     // isoc xfer descriptor
-#define NLPTR_TYPE_QH            1     // queue head
-#define NLPTR_TYPE_STITD         2     // split xaction, isoc xfer descriptor
-#define NLPTR_TYPE_FSTN          3     // frame span traversal node
+#define NLPTR_TYPE_ITD           0     /* isoc xfer descriptor */
+#define NLPTR_TYPE_QH            1     /* queue head */
+#define NLPTR_TYPE_STITD         2     /* split xaction, isoc xfer descriptor */
+#define NLPTR_TYPE_FSTN          3     /* frame span traversal node */
 
 #define SET_LAST_RUN_CLOCK(s) \
     (s)->last_run_ns = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
@@ -88,10 +90,10 @@ typedef enum {
 
 #define set_field(data, newval, field) do { \
     uint32_t val = *data; \
-    val &= ~ field##_MASK; \
+    val &= ~field##_MASK; \
     val |= ((newval) << field##_SH) & field##_MASK; \
     *data = val; \
-    } while(0)
+    } while (0)
 
 static const char *ehci_state_names[] = {
     [EST_INACTIVE]     = "INACTIVE",
@@ -472,8 +474,10 @@ static bool ehci_verify_pid(EHCIQueue *q, EHCIqtd *qtd)
     }
 }
 
-/* Finish executing and writeback a packet outside of the regular
-   fetchqh -> fetchqtd -> execute -> writeback cycle */
+/*
+ * Finish executing and writeback a packet outside of the regular
+ * fetchqh -> fetchqtd -> execute -> writeback cycle
+ */
 static void ehci_writeback_async_complete_packet(EHCIPacket *p)
 {
     EHCIQueue *q = p->queue;
@@ -733,7 +737,7 @@ static void ehci_detach(USBPort *port)
     ehci_queues_rip_device(s, port->dev, 0);
     ehci_queues_rip_device(s, port->dev, 1);
 
-    *portsc &= ~(PORTSC_CONNECT|PORTSC_PED|PORTSC_SUSPEND);
+    *portsc &= ~(PORTSC_CONNECT | PORTSC_PED | PORTSC_SUSPEND);
     *portsc |= PORTSC_CSC;
 
     ehci_raise_irq(s, USBSTS_PCD);
@@ -858,7 +862,7 @@ void ehci_reset(void *opaque)
      * Do the detach before touching portsc, so that it correctly gets send to
      * us or to our companion based on PORTSC_POWNER before the reset.
      */
-    for(i = 0; i < EHCI_PORTS; i++) {
+    for (i = 0; i < EHCI_PORTS; i++) {
         devs[i] = s->ports[i].dev;
         if (devs[i] && devs[i]->attached) {
             usb_detach(&s->ports[i]);
@@ -877,7 +881,7 @@ void ehci_reset(void *opaque)
     s->astate = EST_INACTIVE;
     s->pstate = EST_INACTIVE;
 
-    for(i = 0; i < EHCI_PORTS; i++) {
+    for (i = 0; i < EHCI_PORTS; i++) {
         if (s->companion_ports[i]) {
             s->portsc[i] = PORTSC_POWNER | PORTSC_PPOWER;
         } else {
@@ -942,8 +946,9 @@ static void handle_port_owner_write(EHCIState *s, int port, uint32_t owner)
     uint32_t *portsc = &s->portsc[port];
     uint32_t orig;
 
-    if (s->companion_ports[port] == NULL)
+    if (s->companion_ports[port] == NULL) {
         return;
+    }
 
     owner = owner & PORTSC_POWNER;
     orig  = *portsc & PORTSC_POWNER;
@@ -988,7 +993,7 @@ static void ehci_port_write(void *ptr, hwaddr addr,
         trace_usb_ehci_port_reset(port, 1);
     }
 
-    if (!(val & PORTSC_PRESET) &&(*portsc & PORTSC_PRESET)) {
+    if (!(val & PORTSC_PRESET) && (*portsc & PORTSC_PRESET)) {
         trace_usb_ehci_port_reset(port, 0);
         if (dev && dev->attached) {
             usb_port_reset(&s->ports[port]);
@@ -1065,8 +1070,10 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,
         break;
 
     case USBSTS:
-        val &= USBSTS_RO_MASK;              // bits 6 through 31 are RO
-        ehci_clear_usbsts(s, val);          // bits 0 through 5 are R/WC
+        /* bits 6 through 31 are RO */
+        val &= USBSTS_RO_MASK;
+        /* bits 0 through 5 are R/WC */
+        ehci_clear_usbsts(s, val);
         val = s->usbsts;
         ehci_update_irq(s);
         break;
@@ -1131,8 +1138,7 @@ static void ehci_flush_qh(EHCIQueue *q)
     put_dwords(q->ehci, addr + 3 * sizeof(uint32_t), qh + 3, dwords - 3);
 }
 
-// 4.10.2
-
+/* 4.10.2 */
 static int ehci_qh_do_overlay(EHCIQueue *q)
 {
     EHCIPacket *p = QTAILQ_FIRST(&q->packets);
@@ -1145,8 +1151,7 @@ static int ehci_qh_do_overlay(EHCIQueue *q)
     assert(p != NULL);
     assert(p->qtdaddr == q->qtdaddr);
 
-    // remember values in fields to preserve in qh after overlay
-
+    /*  remember values in fields to preserve in qh after overlay */
     dtoggle = q->qh.token & QTD_TOKEN_DTOGGLE;
     ping    = q->qh.token & QTD_TOKEN_PING;
 
@@ -1170,7 +1175,7 @@ static int ehci_qh_do_overlay(EHCIQueue *q)
     }
 
     if (!(q->qh.epchar & QH_EPCHAR_DTC)) {
-        // preserve QH DT bit
+        /* preserve QH DT bit */
         q->qh.token &= ~QTD_TOKEN_DTOGGLE;
         q->qh.token |= dtoggle;
     }
@@ -1397,9 +1402,7 @@ static int ehci_execute(EHCIPacket *p, const char *action)
     return 1;
 }
 
-/*  4.7.2
- */
-
+/* 4.7.2 */
 static int ehci_process_itd(EHCIState *ehci,
                             EHCIitd *itd,
                             uint32_t addr)
@@ -1411,13 +1414,13 @@ static int ehci_process_itd(EHCIState *ehci,
 
     ehci->periodic_sched_active = PERIODIC_ACTIVE;
 
-    dir =(itd->bufptr[1] & ITD_BUFPTR_DIRECTION);
+    dir = (itd->bufptr[1] & ITD_BUFPTR_DIRECTION);
     devaddr = get_field(itd->bufptr[0], ITD_BUFPTR_DEVADDR);
     endp = get_field(itd->bufptr[0], ITD_BUFPTR_EP);
     max = get_field(itd->bufptr[1], ITD_BUFPTR_MAXPKT);
     mult = get_field(itd->bufptr[2], ITD_BUFPTR_MULT);
 
-    for(i = 0; i < 8; i++) {
+    for (i = 0; i < 8; i++) {
         if (itd->transact[i] & ITD_XACT_ACTIVE) {
             pg   = get_field(itd->transact[i], ITD_XACT_PGSEL);
             off  = itd->transact[i] & ITD_XACT_OFFSET_MASK;
@@ -1513,8 +1516,9 @@ static int ehci_process_itd(EHCIState *ehci,
 }
 
 
-/*  This state is the entry point for asynchronous schedule
- *  processing.  Entry here constitutes a EHCI start event state (4.8.5)
+/*
+ * This state is the entry point for asynchronous schedule
+ * processing.  Entry here constitutes a EHCI start event state (4.8.5)
  */
 static int ehci_state_waitlisthead(EHCIState *ehci,  int async)
 {
@@ -1531,7 +1535,7 @@ static int ehci_state_waitlisthead(EHCIState *ehci,  int async)
     ehci_queues_rip_unused(ehci, async);
 
     /*  Find the head of the list (4.9.1.1) */
-    for(i = 0; i < MAX_QH; i++) {
+    for (i = 0; i < MAX_QH; i++) {
         if (get_dwords(ehci, NLPTR_GET(entry), (uint32_t *) &qh,
                        sizeof(EHCIqh) >> 2) < 0) {
             return 0;
@@ -1564,8 +1568,9 @@ out:
 }
 
 
-/*  This state is the entry point for periodic schedule processing as
- *  well as being a continuation state for async processing.
+/*
+ * This state is the entry point for periodic schedule processing as
+ * well as being a continuation state for async processing.
  */
 static int ehci_state_fetchentry(EHCIState *ehci, int async)
 {
@@ -1674,7 +1679,7 @@ static EHCIQueue *ehci_state_fetchqh(EHCIState *ehci, int async)
 
 #if EHCI_DEBUG
     if (q->qhaddr != q->qh.next) {
-    DPRINTF("FETCHQH:  QH 0x%08x (h %x halt %x active %x) next 0x%08x\n",
+        DPRINTF("FETCHQH:  QH 0x%08x (h %x halt %x active %x) next 0x%08x\n",
                q->qhaddr,
                q->qh.epchar & QH_EPCHAR_H,
                q->qh.token & QTD_TOKEN_HALT,
@@ -1924,8 +1929,10 @@ static int ehci_state_execute(EHCIQueue *q)
         return -1;
     }
 
-    // TODO verify enough time remains in the uframe as in 4.4.1.1
-    // TODO write back ptr to async list when done or out of time
+    /*
+     * TODO verify enough time remains in the uframe as in 4.4.1.1
+     * TODO write back ptr to async list when done or out of time
+     */
 
     /* 4.10.3, bottom of page 82, go horizontal on transaction counter == 0 */
     if (!q->async && q->transact_ctr == 0) {
@@ -2036,7 +2043,7 @@ static void ehci_advance_state(EHCIState *ehci, int async)
     int again;
 
     do {
-        switch(ehci_get_state(ehci, async)) {
+        switch (ehci_get_state(ehci, async)) {
         case EST_WAITLISTHEAD:
             again = ehci_state_waitlisthead(ehci, async);
             break;
@@ -2115,21 +2122,20 @@ static void ehci_advance_state(EHCIState *ehci, int async)
             ehci_reset(ehci);
             again = 0;
         }
-    }
-    while (again);
+    } while (again);
 }
 
 static void ehci_advance_async_state(EHCIState *ehci)
 {
     const int async = 1;
 
-    switch(ehci_get_state(ehci, async)) {
+    switch (ehci_get_state(ehci, async)) {
     case EST_INACTIVE:
         if (!ehci_async_enabled(ehci)) {
             break;
         }
         ehci_set_state(ehci, async, EST_ACTIVE);
-        // No break, fall through to ACTIVE
+        /* No break, fall through to ACTIVE */
 
     case EST_ACTIVE:
         if (!ehci_async_enabled(ehci)) {
@@ -2153,7 +2159,8 @@ static void ehci_advance_async_state(EHCIState *ehci)
         ehci_set_state(ehci, async, EST_WAITLISTHEAD);
         ehci_advance_state(ehci, async);
 
-        /* If the doorbell is set, the guest wants to make a change to the
+        /*
+         * If the doorbell is set, the guest wants to make a change to the
          * schedule. The host controller needs to release cached data.
          * (section 4.8.2)
          */
@@ -2180,13 +2187,13 @@ static void ehci_advance_periodic_state(EHCIState *ehci)
     uint32_t list;
     const int async = 0;
 
-    // 4.6
+    /* 4.6 */
 
-    switch(ehci_get_state(ehci, async)) {
+    switch (ehci_get_state(ehci, async)) {
     case EST_INACTIVE:
         if (!(ehci->frindex & 7) && ehci_periodic_enabled(ehci)) {
             ehci_set_state(ehci, async, EST_ACTIVE);
-            // No break, fall through to ACTIVE
+            /* No break, fall through to ACTIVE */
         } else
             break;
 
@@ -2210,7 +2217,7 @@ static void ehci_advance_periodic_state(EHCIState *ehci)
 
         DPRINTF("PERIODIC state adv fr=%d.  [%08X] -> %08X\n",
                 ehci->frindex / 8, list, entry);
-        ehci_set_fetch_addr(ehci, async,entry);
+        ehci_set_fetch_addr(ehci, async, entry);
         ehci_set_state(ehci, async, EST_FETCHENTRY);
         ehci_advance_state(ehci, async);
         ehci_queues_rip_unused(ehci, async);
@@ -2235,7 +2242,8 @@ static void ehci_update_frindex(EHCIState *ehci, int uframes)
         ehci_raise_irq(ehci, USBSTS_FLR);
     }
 
-    /* How many times will frindex roll over 0x4000 with this frame count?
+    /*
+     * How many times will frindex roll over 0x4000 with this frame count?
      * usbsts_frindex is decremented by 0x4000 on rollover until it reaches 0
      */
     int rollovers = (ehci->frindex + uframes) / 0x4000;
@@ -2315,8 +2323,9 @@ static void ehci_work_bh(void *opaque)
         ehci->async_stepdown++;
     }
 
-    /*  Async is not inside loop since it executes everything it can once
-     *  called
+    /*
+     * Async is not inside loop since it executes everything it can once
+     * called
      */
     if (ehci_async_enabled(ehci) || ehci->astate != EST_INACTIVE) {
         need_timer++;
@@ -2334,15 +2343,18 @@ static void ehci_work_bh(void *opaque)
     }
 
     if (need_timer) {
-        /* If we've raised int, we speed up the timer, so that we quickly
-         * notice any new packets queued up in response */
+        /*
+         * If we've raised int, we speed up the timer, so that we quickly
+         * notice any new packets queued up in response
+         */
         if (ehci->int_req_by_async && (ehci->usbsts & USBSTS_INT)) {
             expire_time = t_now +
                 NANOSECONDS_PER_SECOND / (FRAME_TIMER_FREQ * 4);
             ehci->int_req_by_async = false;
         } else {
-            expire_time = t_now + (NANOSECONDS_PER_SECOND
-                               * (ehci->async_stepdown+1) / FRAME_TIMER_FREQ);
+            expire_time = t_now
+                + (NANOSECONDS_PER_SECOND * (ehci->async_stepdown + 1) /
+                   FRAME_TIMER_FREQ);
         }
         timer_mod(ehci->frame_timer, expire_time);
     }
Signature · 1 line
2.43.0
author
#48 lore raw

PATCH v3 13/17 hw/usb/hcd-ehci: Add descriptor address offset property

In reply to Jamin Lin

When 64-bit addressing is supported, the Linux EHCI driver programs the segment register to zero. See ehci_run function:
https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci-hcd.c

The driver comment also notes that descriptor structures allocated from the DMA pool use segment zero semantics.

Descriptor memory is allocated using the DMA API. The platform driver configures a 64-bit DMA mask so memory can be allocated above 4GB. See ehci_platform_probe function:
https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci-platform.c

On AST2700 platforms, system DRAM is mapped above 4GB at 0x400000000. As a result, descriptor addresses constructed directly from the guest EHCI registers do not match the actual system address used by the controller when fetching queue heads (QH) and queue element transfer descriptors (qTD).

Introduce a descriptor-addr-offset property so platforms can provide an address offset applied when constructing descriptor addresses. This allows systems where DRAM resides above 4GB to correctly access EHCI descriptors while keeping the default behavior unchanged for existing machines.

---
 hw/usb/hcd-ehci.h        | 1 +
 hw/usb/hcd-ehci-pci.c    | 2 ++
 hw/usb/hcd-ehci-sysbus.c | 2 ++
 hw/usb/hcd-ehci.c        | 7 ++++++-
 4 files changed, 11 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.h +1 -0
index 6406f536e8..05d8db4a5d 100644
@@ -264,6 +264,7 @@ struct EHCIState {
     /* properties */
     uint32_t maxframes;
     bool caps_64bit_addr;
+    uint64_t descriptor_addr_offset;
 
     /*
      *  EHCI spec version 1.0 Section 2.3
hw/usb/hcd-ehci-pci.c +2 -0
index 2ea8549db9..115d05ede0 100644
@@ -139,6 +139,8 @@ static const Property ehci_pci_properties[] = {
     DEFINE_PROP_UINT32("maxframes", EHCIPCIState, ehci.maxframes, 128),
     DEFINE_PROP_BOOL("caps-64bit-addr", EHCIPCIState, ehci.caps_64bit_addr,
                      false),
+    DEFINE_PROP_UINT64("descriptor-addr-offset", EHCIPCIState,
+                       ehci.descriptor_addr_offset, 0),
 };
 
 static const VMStateDescription vmstate_ehci_pci = {
hw/usb/hcd-ehci-sysbus.c +2 -0
index 61215e9f3d..df138fb339 100644
@@ -36,6 +36,8 @@ static const Property ehci_sysbus_properties[] = {
                      false),
     DEFINE_PROP_BOOL("caps-64bit-addr", EHCISysBusState, ehci.caps_64bit_addr,
                      false),
+    DEFINE_PROP_UINT64("descriptor-addr-offset", EHCISysBusState,
+                       ehci.descriptor_addr_offset, 0),
 };
 
 static void usb_ehci_sysbus_realize(DeviceState *dev, Error **errp)
hw/usb/hcd-ehci.c +6 -1
index e1d5ecbad7..ea9d5846bc 100644
@@ -162,7 +162,12 @@ static uint64_t ehci_get_buf_addr(EHCIState *s, uint32_t hi, uint32_t low,
 
 static uint64_t ehci_get_desc_addr(EHCIState *s, uint32_t low)
 {
-    return ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX);
+    uint64_t addr;
+
+    addr = ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX);
+    addr += s->descriptor_addr_offset;
+
+    return addr;
 }
 
 static void ehci_trace_usbsts(uint32_t mask, int state)
Signature · 1 line
2.43.0
author
#40 lore raw

PATCH v3 10/17 hw/usb/hcd-ehci: Implement 64-bit qTD descriptor addressing

In reply to Jamin Lin

EHCI supports 64-bit addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register provides the upper 32 bits that are concatenated with 32-bit link pointer values to form 64-bit control data structure addresses (EHCI 1.0, section 2.3.5 and Appendix B).

qTD link pointers (current_qtd/next_qtd/altnext_qtd and qTD.next) are stored as 32-bit values in the data structures and must be expanded to full 64-bit descriptor addresses when 64-bit mode is enabled. Update the qTD traversal paths to use ehci_get_desc_addr() when following link pointers.

Appendix B also defines high dword fields for qTD buffer pointers. Add bufptr_hi[5] to EHCIqtd and extend qTD fetch and QH overlay handling to load and propagate the high buffer pointer fields.

When 64-bit capability is disabled, descriptor and buffer addresses remain 32-bit and existing behaviour is unchanged.

---
 hw/usb/hcd-ehci.h |  1 +
 hw/usb/hcd-ehci.c | 19 ++++++++++++-------
 2 files changed, 13 insertions(+), 7 deletions(-)
hw/usb/hcd-ehci.h +1 -0
index 773f376834..712de9d50e 100644
@@ -139,6 +139,7 @@ typedef struct EHCIqtd {
     uint32_t bufptr[5];               /* Standard buffer pointer */
 #define QTD_BUFPTR_MASK               0xfffff000
 #define QTD_BUFPTR_SH                 12
+    uint32_t bufptr_hi[5];
 } EHCIqtd;
 
 /* QH overlay: altnext_qtd, token, bufptr[5], bufptr_hi[5] */
hw/usb/hcd-ehci.c +12 -7
index 37ae637e08..d93d227e6f 100644
@@ -473,7 +473,8 @@ static bool ehci_verify_qtd(EHCIPacket *p, EHCIqtd *qtd)
             (p->qtd.next != qtd->next)) ||
         (!NLPTR_TBIT(p->qtd.altnext) && (p->qtd.altnext != qtd->altnext)) ||
         p->qtd.token != qtd->token ||
-        p->qtd.bufptr[0] != qtd->bufptr[0]) {
+        p->qtd.bufptr[0] != qtd->bufptr[0] ||
+        p->qtd.bufptr_hi[0] != qtd->bufptr_hi[0]) {
         return false;
     } else {
         return true;
@@ -1199,6 +1200,7 @@ static int ehci_qh_do_overlay(EHCIQueue *q)
 
     for (i = 0; i < 5; i++) {
         q->qh.bufptr[i] = p->qtd.bufptr[i];
+        q->qh.bufptr_hi[i] = p->qtd.bufptr_hi[i];
     }
 
     if (!(q->qh.epchar & QH_EPCHAR_DTC)) {
@@ -1232,7 +1234,8 @@ static int ehci_init_transfer(EHCIPacket *p)
             return -1;
         }
 
-        page  = p->qtd.bufptr[cpage] & QTD_BUFPTR_MASK;
+        page = ehci_get_buf_addr(p->queue->ehci, p->qtd.bufptr_hi[cpage],
+                                 p->qtd.bufptr[cpage], QTD_BUFPTR_MASK);
         page += offset;
         plen  = bytes;
         if (plen > 4096 - offset) {
@@ -1726,7 +1729,7 @@ static EHCIQueue *ehci_state_fetchqh(EHCIState *ehci, int async)
     } else if ((q->qh.token & QTD_TOKEN_ACTIVE) &&
                (NLPTR_TBIT(q->qh.current_qtd) == 0) &&
                (q->qh.current_qtd != 0)) {
-        q->qtdaddr = q->qh.current_qtd;
+        q->qtdaddr = ehci_get_desc_addr(ehci, q->qh.current_qtd);
         ehci_set_state(ehci, async, EST_FETCHQTD);
 
     } else {
@@ -1798,14 +1801,14 @@ static int ehci_state_advqueue(EHCIQueue *q)
      */
     if (((q->qh.token & QTD_TOKEN_TBYTES_MASK) != 0) &&
         (NLPTR_TBIT(q->qh.altnext_qtd) == 0)) {
-        q->qtdaddr = q->qh.altnext_qtd;
+        q->qtdaddr = ehci_get_desc_addr(q->ehci, q->qh.altnext_qtd);
         ehci_set_state(q->ehci, q->async, EST_FETCHQTD);
 
     /*
      *  next qTD is valid
      */
     } else if (NLPTR_TBIT(q->qh.next_qtd) == 0) {
-        q->qtdaddr = q->qh.next_qtd;
+        q->qtdaddr = ehci_get_desc_addr(q->ehci, q->qh.next_qtd);
         ehci_set_state(q->ehci, q->async, EST_FETCHQTD);
 
     /*
@@ -1834,7 +1837,9 @@ static int ehci_state_fetchqtd(EHCIQueue *q)
     if (get_dwords(q->ehci, addr +  0, &qtd.next,    1) < 0 ||
         get_dwords(q->ehci, addr +  4, &qtd.altnext, 1) < 0 ||
         get_dwords(q->ehci, addr + 12, qtd.bufptr,
-                   ARRAY_SIZE(qtd.bufptr)) < 0) {
+                   ARRAY_SIZE(qtd.bufptr)) < 0 ||
+        get_dwords(q->ehci, addr + 32, qtd.bufptr_hi,
+                   ARRAY_SIZE(qtd.bufptr_hi)) < 0) {
         return 0;
     }
     ehci_trace_qtd(q, NLPTR_GET(q->qtdaddr), &qtd);
@@ -1915,7 +1920,7 @@ static int ehci_fill_queue(EHCIPacket *p)
         if (NLPTR_TBIT(qtd.next) != 0) {
             break;
         }
-        qtdaddr = qtd.next;
+        qtdaddr = ehci_get_desc_addr(q->ehci, qtd.next);
         /*
          * Detect circular td lists, Windows creates these, counting on the
          * active bit going low after execution to make the queue stop.
Signature · 1 line
2.43.0
author
#42 lore raw

PATCH v3 11/17 hw/usb/hcd-ehci: Implement 64-bit iTD descriptor addressing

In reply to Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register provides the upper 32 bits that are concatenated with 32-bit link pointer values to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

iTD link pointers are stored as 32-bit values and must be expanded to full 64-bit descriptor addresses when 64-bit mode is enabled. Update the iTD traversal path to use ehci_get_desc_addr() when following link pointers.

Appendix B also defines high dword fields for iTD buffer pointers. Add bufptr_hi[7] to EHCIitd and use ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] fields when processing isochronous transfers. This allows buffers above 4GB to be handled correctly.

When 64-bit capability is disabled, descriptor and buffer addresses remain 32-bit and existing behaviour is unchanged.

---
 hw/usb/hcd-ehci.h | 1 +
 hw/usb/hcd-ehci.c | 9 ++++++---
 2 files changed, 7 insertions(+), 3 deletions(-)
hw/usb/hcd-ehci.h +1 -0
index 712de9d50e..af8c080c60 100644
@@ -63,6 +63,7 @@ typedef struct EHCIitd {
 #define ITD_BUFPTR_MAXPKT_SH     0
 #define ITD_BUFPTR_MULT_MASK     0x00000003
 #define ITD_BUFPTR_MULT_SH       0
+    uint32_t bufptr_hi[7];
 } EHCIitd;
 
 /*
hw/usb/hcd-ehci.c +6 -3
index d93d227e6f..3db1c25e4d 100644
@@ -1463,7 +1463,8 @@ static int ehci_process_itd(EHCIState *ehci,
                 return -1;
             }
 
-            ptr1 = (itd->bufptr[pg] & ITD_BUFPTR_MASK);
+            ptr1 = ehci_get_buf_addr(ehci, itd->bufptr_hi[pg],
+                                     itd->bufptr[pg], ITD_BUFPTR_MASK);
             qemu_sglist_init(&ehci->isgl, ehci->device, 2, ehci->as);
             if (off + len > 4096) {
                 /* transfer crosses page border */
@@ -1471,7 +1472,9 @@ static int ehci_process_itd(EHCIState *ehci,
                     qemu_sglist_destroy(&ehci->isgl);
                     return -1;  /* avoid page pg + 1 */
                 }
-                ptr2 = (itd->bufptr[pg + 1] & ITD_BUFPTR_MASK);
+                ptr2 = ehci_get_buf_addr(ehci, itd->bufptr_hi[pg + 1],
+                                         itd->bufptr[pg + 1],
+                                         ITD_BUFPTR_MASK);
                 uint32_t len2 = off + len - 4096;
                 uint32_t len1 = len - len2;
                 qemu_sglist_add(&ehci->isgl, ptr1 + off, len1);
@@ -1761,7 +1764,7 @@ static int ehci_state_fetchitd(EHCIState *ehci, int async)
 
     put_dwords(ehci, NLPTR_GET(entry), (uint32_t *) &itd,
                sizeof(EHCIitd) >> 2);
-    ehci_set_fetch_addr(ehci, async, itd.next);
+    ehci_set_fetch_addr(ehci, async, ehci_get_desc_addr(ehci, itd.next));
     ehci_set_state(ehci, async, EST_FETCHENTRY);
 
     return 1;
Signature · 1 line
2.43.0
author
#44 lore raw

PATCH v3 12/17 hw/usb/hcd-ehci: Implement 64-bit siTD descriptor addressing

In reply to Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register provides the upper 32 bits that are concatenated with 32-bit link pointer values to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

siTD link pointers are stored as 32-bit values and must be expanded to full 64-bit descriptor addresses when 64-bit mode is enabled. Update the siTD traversal path to use ehci_get_desc_addr() when following link pointers.

Appendix B also defines high dword fields for siTD buffer pointers. Add bufptr_hi[] fields to EHCIsitd and use ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] when processing split isochronous transfers. This allows buffers above 4GB to be handled correctly.

When 64-bit capability is disabled, descriptor and buffer addresses remain 32-bit and existing behaviour is unchanged.

---
 hw/usb/hcd-ehci.h | 1 +
 hw/usb/hcd-ehci.c | 2 +-
 2 files changed, 2 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.h +1 -0
index af8c080c60..6406f536e8 100644
@@ -109,6 +109,7 @@ typedef struct EHCIsitd {
 #define SITD_BUFPTR_TCNT_MASK         0x00000007
 
     uint32_t backptr;                 /* Standard next link pointer */
+    uint32_t bufptr_hi[2];
 } EHCIsitd;
 
 /*
hw/usb/hcd-ehci.c +1 -1
index 3db1c25e4d..e1d5ecbad7 100644
@@ -1791,7 +1791,7 @@ static int ehci_state_fetchsitd(EHCIState *ehci, int async)
         warn_report("Skipping active siTD");
     }
 
-    ehci_set_fetch_addr(ehci, async, sitd.next);
+    ehci_set_fetch_addr(ehci, async, ehci_get_desc_addr(ehci, sitd.next));
     ehci_set_state(ehci, async, EST_FETCHENTRY);
     return 1;
 }
Signature · 1 line
2.43.0
author
#16 lore raw

PATCH v3 5/17 hw/usb/hcd-ehci: Replace DPRINTF debug logs with trace events

In reply to Jamin Lin

Replace legacy DPRINTF() debug logging in the EHCI host controller implementation with QEMU trace events.

The EHCI_DEBUG macro and associated DPRINTF() definitions are removed as they are no longer needed.

Trace events are added for:
- packet submission and completion
- queue head execution state
- periodic schedule advancement
- ITD error conditions
- port state handling
- skipped microframes

No functional change.

---
 hw/usb/hcd-ehci.h   | 10 ---------
 hw/usb/hcd-ehci.c   | 49 +++++++++++++++++++--------------------------
 hw/usb/trace-events |  9 +++++++++
 3 files changed, 30 insertions(+), 38 deletions(-)
hw/usb/hcd-ehci.h +0 -10
index 736407fd52..4234591cb4 100644
@@ -24,16 +24,6 @@
 #include "hw/pci/pci_device.h"
 #include "hw/core/sysbus.h"
 
-#ifndef EHCI_DEBUG
-#define EHCI_DEBUG   0
-#endif
-
-#if EHCI_DEBUG
-#define DPRINTF printf
-#else
-#define DPRINTF(...)
-#endif
-
 #define MMIO_SIZE        0x1000
 #define CAPA_SIZE        0x10
 
hw/usb/hcd-ehci.c +21 -28

110 lines not on this page · raw message

hw/usb/trace-events +9 -0
index dd04f14add..0d4318dcf1 100644
@@ -89,6 +89,7 @@ usb_ehci_state(const char *schedule, const char *state) "%s schedule %s"
 usb_ehci_qh_ptrs(void *q, uint32_t addr, uint32_t nxt, uint32_t c_qtd, uint32_t n_qtd, uint32_t a_qtd) "q %p - QH @ 0x%08x: next 0x%08x qtds 0x%08x,0x%08x,0x%08x"
 usb_ehci_qh_fields(uint32_t addr, int rl, int mplen, int eps, int ep, int devaddr) "QH @ 0x%08x - rl %d, mplen %d, eps %d, ep %d, dev %d"
 usb_ehci_qh_bits(uint32_t addr, int c, int h, int dtc, int i) "QH @ 0x%08x - c %d, h %d, dtc %d, i %d"
+usb_ehci_qh_tbytes(uint32_t tbytes) "updating tbytes to %d"
 usb_ehci_qtd_ptrs(void *q, uint32_t addr, uint32_t nxt, uint32_t altnext) "q %p - QTD @ 0x%08x: next 0x%08x altnext 0x%08x"
 usb_ehci_qtd_fields(uint32_t addr, int tbytes, int cpage, int cerr, int pid) "QTD @ 0x%08x - tbytes %d, cpage %d, cerr %d, pid %d"
 usb_ehci_qtd_bits(uint32_t addr, int ioc, int active, int halt, int babble, int xacterr) "QTD @ 0x%08x - ioc %d, active %d, halt %d, babble %d, xacterr %d"
@@ -100,13 +101,21 @@ usb_ehci_port_reset(uint32_t port, int enable) "reset port #%d - %d"
 usb_ehci_port_suspend(uint32_t port) "port #%d"
 usb_ehci_port_wakeup(uint32_t port) "port #%d"
 usb_ehci_port_resume(uint32_t port) "port #%d"
+usb_ehci_port_disable(uint32_t port) "port #%d"
 usb_ehci_queue_action(void *q, const char *action) "q %p: %s"
 usb_ehci_packet_action(void *q, void *p, const char *action) "q %p p %p: %s"
+usb_ehci_packet_submit(uint32_t qhaddr, uint32_t next, uint32_t qtdaddr, int pid, size_t len, int endp, int status, int actual_length) "qh=0x%x, next=0x%x, qtd=0x%x, pid=0x%x, len=%zd, endp=0x%x, status=%d, actual_length=%d"
 usb_ehci_irq(uint32_t level, uint32_t frindex, uint32_t sts, uint32_t mask) "level %d, frindex 0x%04x, sts 0x%x, mask 0x%x"
 usb_ehci_guest_bug(const char *reason) "%s"
 usb_ehci_doorbell_ring(void) ""
 usb_ehci_doorbell_ack(void) ""
 usb_ehci_dma_error(void) ""
+usb_ehci_execute_complete(uint32_t qhaddr, uint32_t next, uint32_t qtdaddr, int status, int actual_length) "qhaddr=0x%x, next=0x%x, qtdaddr=0x%x, status=%d, actual_length=%d"
+usb_ehci_fetchqh_reclaim_done(uint32_t qhaddr) "QH 0x%08x H-bit set, reclamation status reset - done processing"
+usb_ehci_fetchqh_dbg(uint32_t qhaddr, uint32_t h, uint32_t halt, uint32_t active, uint32_t next) "QH 0x%08x (h 0x%x halt 0x%x active 0x%x) next 0x%08x"
+usb_ehci_periodic_state_advance(uint32_t frame, uint32_t list, uint32_t entry) "frame=%d, list=0x%x, entry=0x%x"
+usb_ehci_skipped_uframes(uint64_t skipped_uframes) "skipped %" PRIu64 " uframes"
+usb_ehci_log(const char *msg) "%s"
 
 # hcd-uhci.c
 usb_uhci_reset(void) "=== RESET ==="
Signature · 1 line
2.43.0
#2 lore raw

Re: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

In reply to Jamin Lin
Quoted 39 lines · Jamin Lin

The EHCIfstn structure, defined according to EHCI spec section 3.7, is currently unused in the EHCI implementation. Remove it to reduce unused code and improve maintainability.

Additionally, drop a block of disabled (#if 0) code in ehci_state_advqueue() that was never implemented. This code path is incomplete and has no functional impact.

No functional change.

---
hw/usb/hcd-ehci.h | 7 -------
hw/usb/hcd-ehci.c | 11 -----------

2 files changed, 18 deletions(-)
hw/usb/hcd-ehci.h +0 -7
index 0ae8c06331..3aec0e047c 100644
@@ -202,13 +202,6 @@ typedef struct EHCIqh {
 #define BUFPTR_SBYTES_SH              5
 } EHCIqh;
 
-/*  EHCI spec version 1.0 Section 3.7
- */
-typedef struct EHCIfstn {
-    uint32_t next;                    /* Standard next link pointer */
-    uint32_t backptr;                 /* Standard next link pointer */
-} EHCIfstn;
-
 enum async_state {
    EHCI_ASYNC_NONE = 0,
    EHCI_ASYNC_INITIALIZED,
hw/usb/hcd-ehci.c +0 -0
index 57f930b099..5ea8461f70 100644

hw/usb/hcd-ehci.c static int ehci_state_fetchsitd(EHCIState *ehci, int async) · line 1763

Quoted 8 lines
hw/usb/hcd-ehci.c +0 -5
@@ -1756,17 +1756,6 @@ static int ehci_state_fetchsitd(EHCIState *ehci, int async)
 /* Section 4.10.2 - paragraph 3 */
 static int ehci_state_advqueue(EHCIQueue *q)
 {
-#if 0
-    /* TO-DO: 4.10.2 - paragraph 2
-     * if I-bit is set to 1 and QH is not active
-     * go to horizontal QH
-     */

Is it worth to keep the comment in case it's still valid to remind we have something missing here?

Regards,
BALATON Zoltan

Quoted 12 lines
hw/usb/hcd-ehci.c +0 -6
-    if (I-bit set) {
-        ehci_set_state(ehci, async, EST_HORIZONTALQH);
-        goto out;
-    }
-#endif
-
    /*
     * want data and alt-next qTD is valid
     */
Signature · 1 line
2.43.0
author
#3 lore raw

RE: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

Hi all,

Quoted 54 lines

Subject: Re: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

Quoted 42 lines · Jamin Lin

The EHCIfstn structure, defined according to EHCI spec section 3.7, is currently unused in the EHCI implementation. Remove it to reduce unused code and improve maintainability.

Additionally, drop a block of disabled (#if 0) code in ehci_state_advqueue() that was never implemented. This code path is incomplete and has no functional impact.

No functional change.

---
hw/usb/hcd-ehci.h | 7 -------
hw/usb/hcd-ehci.c | 11 -----------

2 files changed, 18 deletions(-)
hw/usb/hcd-ehci.h index renamed from hw/usb/hcd-ehci.h +0 -0

0ae8c06331..3aec0e047c 100644

hw/usb/hcd-ehci.h +0 -7
@@ -202,13 +202,6 @@ typedef struct EHCIqh {
 #define BUFPTR_SBYTES_SH              5
 } EHCIqh;
 
-/*  EHCI spec version 1.0 Section 3.7
- */
-typedef struct EHCIfstn {
-    uint32_t next;                    /* Standard next link pointer */
-    uint32_t backptr;                 /* Standard next link pointer */
-} EHCIfstn;
-
 enum async_state {
    EHCI_ASYNC_NONE = 0,
    EHCI_ASYNC_INITIALIZED,
hw/usb/hcd-ehci.c index renamed from hw/usb/hcd-ehci.c +0 -0

57f930b099..5ea8461f70 100644

hw/usb/hcd-ehci.c +0 -0
@@ -1756,17 +1756,6 @@ static int ehci_state_fetchsitd(EHCIState
 *ehci, int async)
 /* Section 4.10.2 - paragraph 3 */

static int ehci_state_advqueue(EHCIQueue *q) { -#if 0

hw/usb/hcd-ehci.c static int ehci_state_fetchsitd(EHCIState *ehci, int async) · line 1763

Quoted 4 lines
hw/usb/hcd-ehci.c +0 -4
-    /* TO-DO: 4.10.2 - paragraph 2
-     * if I-bit is set to 1 and QH is not active
-     * go to horizontal QH
-     */

Is it worth to keep the comment in case it's still valid to remind we have something missing here?

Quoted 3 lines

Regards,
BALATON Zoltan

If we decide to keep these comments as reminders for unimplemented parts, please let me know. I will resend the patch series and drop this patch.

Thanks,
Jamin

Quoted 12 lines
Quoted 12 lines
hw/usb/hcd-ehci.c +0 -7
-    if (I-bit set) {
-        ehci_set_state(ehci, async, EST_HORIZONTALQH);
-        goto out;
-    }
-#endif
-
    /*
     * want data and alt-next qTD is valid
     */
--

2.43.0

author
#4 lore raw

RE: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

Hi all,

Quoted 64 lines

Subject: RE: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

Hi all,

Quoted 28 lines

Subject: Re: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

Quoted 24 lines · Jamin Lin

The EHCIfstn structure, defined according to EHCI spec section 3.7, is currently unused in the EHCI implementation. Remove it to reduce unused code and improve maintainability.

Additionally, drop a block of disabled (#if 0) code in ehci_state_advqueue() that was never implemented. This code path is incomplete and has no functional impact.

No functional change.

---
hw/usb/hcd-ehci.h | 7 -------
hw/usb/hcd-ehci.c | 11 -----------

2 files changed, 18 deletions(-)
hw/usb/hcd-ehci.h index renamed from hw/usb/hcd-ehci.h +0 -0

0ae8c06331..3aec0e047c 100644

hw/usb/hcd-ehci.h +0 -0
@@ -202,13 +202,6 @@ typedef struct EHCIqh {
 #define BUFPTR_SBYTES_SH              5
 } EHCIqh;
 
Quoted 4 lines

hw/usb/hcd-ehci.h typedef struct EHCIqh { · line 208

Quoted 4 lines
hw/usb/hcd-ehci.h +0 -4
-/*  EHCI spec version 1.0 Section 3.7
- */
-typedef struct EHCIfstn {
-    uint32_t next;                    /* Standard next link pointer

*/

Quoted 22 lines
Quoted 14 lines
hw/usb/hcd-ehci.h +0 -3
-    uint32_t backptr;                 /* Standard next link pointer */
-} EHCIfstn;
-
 enum async_state {
    EHCI_ASYNC_NONE = 0,
    EHCI_ASYNC_INITIALIZED,
hw/usb/hcd-ehci.c index renamed from hw/usb/hcd-ehci.c +0 -0

57f930b099..5ea8461f70 100644

hw/usb/hcd-ehci.c +0 -0
@@ -1756,17 +1756,6 @@ static int ehci_state_fetchsitd(EHCIState
 *ehci, int async)
 /* Section 4.10.2 - paragraph 3 */

static int ehci_state_advqueue(EHCIQueue *q) { -#if 0

hw/usb/hcd-ehci.c

Quoted 4 lines
hw/usb/hcd-ehci.c +0 -4
-    /* TO-DO: 4.10.2 - paragraph 2
-     * if I-bit is set to 1 and QH is not active
-     * go to horizontal QH
-     */

Is it worth to keep the comment in case it's still valid to remind we have something missing here?

Quoted 3 lines

Regards,
BALATON Zoltan

Quoted 6 lines

If we decide to keep these comments as reminders for unimplemented parts, please let me know. I will resend the patch series and drop this patch.

Thanks,
Jamin

Another approach:
We can remove the #if 0 dead code and keep only the comments as a reminder that this part is not yet implemented.

Ex:

hw/usb/hcd-ehci.c +0 -5
@@ -1756,16 +1756,10 @@ static int ehci_state_fetchsitd(EHCIState *ehci, int async)
 /* Section 4.10.2 - paragraph 3 */
 static int ehci_state_advqueue(EHCIQueue *q)
 {
-#if 0
     /* TO-DO: 4.10.2 - paragraph 2
      * if I-bit is set to 1 and QH is not active
      * go to horizontal QH
      */
-    if (I-bit set) {
-        ehci_set_state(ehci, async, EST_HORIZONTALQH);
-        goto out;
-    }
 

Thanks,
Jamin

Quoted 12 lines
Quoted 12 lines
Quoted 12 lines
hw/usb/hcd-ehci.c +0 -7
-    if (I-bit set) {
-        ehci_set_state(ehci, async, EST_HORIZONTALQH);
-        goto out;
-    }
-#endif
-
    /*
     * want data and alt-next qTD is valid
     */
--

2.43.0

Reviewed-by on v3 6/17
#19 lore raw

Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

In reply to Jamin Lin
Quoted 39 lines · Jamin Lin

Change internal EHCI descriptor addresses from uint32_t to uint64_t.

The following fields are updated:
- EHCIPacket::qtdaddr
- EHCIQueue::{qhaddr, qtdaddr}
- EHCIState::{a_fetch_addr, p_fetch_addr}

Update get_dwords() and put_dwords() to take 64-bit addresses and propagate the type change through the descriptor traversal paths.

Adjust NLPTR_GET() to operate on 64-bit values:

     #define NLPTR_GET(x) ((x) & ~0x1fULL)

so that link pointer masking works correctly when descriptor addresses exceed 32-bit space. The previous mask (0xffffffe0) implicitly truncated addresses to 32 bits.

This patch does not change the on-wire descriptor layout yet. It only removes the internal 32-bit address limit and prepares for later patches that will add full 64-bit QH/qTD/iTD/siTD support.

Update the EHCI trace-events prototypes for QH, qTD, iTD, and siTD to use uint64_t for the address argument and print it with PRIx64. This ensures full 64-bit addresses are shown in trace output and improves debugging of queue heads and transfer descriptors.

Migration compatibility:

The fetch address fields in EHCIState are extended from 32-bit to 64-bit, requiring a VMState version bump (v2 -> v3).

Backward compatibility is preserved by keeping the legacy 32-bit fields (a_fetch_addr_pre_v3, p_fetch_addr_pre_v3) for pre-v3 migration streams via VMSTATE_UINT32_TEST(), and introducing new 64-bit fields for v3+.

In post_load, pre-v3 state is promoted to the 64-bit fields.

Quoted 8 lines

No functional change.

---
  hw/usb/hcd-ehci.h   | 12 +++++----
  hw/usb/hcd-ehci.c   | 64 +++++++++++++++++++++++++++++----------------
  hw/usb/trace-events | 24 ++++++++---------
  3 files changed, 60 insertions(+), 40 deletions(-)
Reviewed-by on v3 7/17
#31 lore raw

Re: [PATCH v3 07/17] hw/usb/hcd-ehci: Add property to advertise 64-bit addressing capability

In reply to Jamin Lin
Quoted 11 lines · Jamin Lin

Introduce a new boolean property, "caps-64bit-addr", to control HCCPARAMS[0] (64-bit Addressing Capability).

When enabled, the EHCI controller advertises support for 64-bit address memory pointers as defined in the EHCI specification (Table 2-7, HCCPARAMS). This allows software to use the 64-bit data structure formats described in Appendix B.

When disabled (default), the controller reports 32-bit addressing capability and uses the standard 32-bit data structures.

Quoted 8 lines
---
  hw/usb/hcd-ehci.h        | 1 +
  hw/usb/hcd-ehci-pci.c    | 2 ++
  hw/usb/hcd-ehci-sysbus.c | 2 ++
  hw/usb/hcd-ehci.c        | 5 ++++-
  4 files changed, 9 insertions(+), 1 deletion(-)
#36 lore raw

Re: [PATCH v3 09/17] hw/usb/hcd-ehci: Implement 64-bit QH descriptor addressing

In reply to Jamin Lin
Quoted 70 lines · Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register supplies the upper 32 bits which are concatenated with 32-bit link pointer fields to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

The current implementation assumes 32-bit QH descriptor addresses and directly uses link pointer values without applying the CTRLDSSEGMENT upper dword.

Introduce a helper, ehci_get_desc_addr(), to construct full 64-bit descriptor addresses when 64-bit capability is enabled. Update QH traversal paths (async list walk, horizontal QH link, and periodic schedule entry handling) to use the translated 64-bit addresses.

EHCI 64-bit buffer pointer fields are defined in Appendix B as split 32-bit low/high parts located at separate offsets, rather than a single contiguous 64-bit field. Therefore, the buffer pointers cannot be represented as uint64_t bufptr[5] without violating the descriptor layout defined by the specification.

Introduce ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] fields. Use this helper when calculating transfer buffer addresses so that data buffers above 4GB are correctly handled.

Also add bufptr_hi[5] to EHCIqh to support 64-bit buffer pointer fields as defined in Appendix B.

When 64-bit capability is disabled, descriptor addresses remain 32-bit and existing behaviour is unchanged.

Note: Similar split 64-bit buffer pointer handling is required for qTD, iTD and siTD descriptors, which will be addressed in follow-up changes.

---
  hw/usb/hcd-ehci.h   |  4 ++++
  hw/usb/hcd-ehci.c   | 48 ++++++++++++++++++++++++++++++++++-----------
  hw/usb/trace-events |  2 +-
  3 files changed, 42 insertions(+), 12 deletions(-)
hw/usb/hcd-ehci.h +4 -0
index fe4c508c32..773f376834 100644
@@ -141,6 +141,9 @@ typedef struct EHCIqtd {
  #define QTD_BUFPTR_SH                 12
  } EHCIqtd;
 
+/* QH overlay: altnext_qtd, token, bufptr[5], bufptr_hi[5] */
+#define EHCI_QH_OVERLAY_COUNT 12
+
  /*
   * EHCI spec version 1.0 Section 3.6
   */
@@ -194,6 +197,7 @@ typedef struct EHCIqh {
  #define BUFPTR_FRAMETAG_MASK          0x0000001f
  #define BUFPTR_SBYTES_MASK            0x00000fe0
  #define BUFPTR_SBYTES_SH              5
+    uint32_t bufptr_hi[5];
  } EHCIqh;
 
  enum async_state {
hw/usb/hcd-ehci.c +0 -0
index b9e4251d62..37ae637e08 100644

hw/usb/hcd-ehci.c static const char *addr2str(hwaddr addr) · line 151

Quoted 5 lines
hw/usb/hcd-ehci.c +2 -0
@@ -147,6 +147,24 @@ static const char *addr2str(hwaddr addr)
      return nr2str(ehci_mmio_names, ARRAY_SIZE(ehci_mmio_names), addr);
  }
 
+static uint64_t ehci_get_buf_addr(EHCIState *s, uint32_t hi, uint32_t low,
+                                  uint32_t mask)

Rename @mask as @low_mask? Or directly use 64-bit?

static uint64_t ehci_get_buf_addr(const EHCIState *s,

                                  uint32_t hi, uint32_t lo,
                                  uint64_t mask)
{
    uint64_t addr = s->caps_64bit_addr ? deposit64(lo, 0, 32, hi): lo;

    return addr & mask;
}
Quoted 15 lines
hw/usb/hcd-ehci.c +15 -0
+{
+    uint64_t addr;
+
+    addr = (uint64_t)(low & mask);
+    if (s->caps_64bit_addr) {
+        addr |= (uint64_t)hi << 32;
+    }
+
+    return addr;
+}
+
+static uint64_t ehci_get_desc_addr(EHCIState *s, uint32_t low)
+{
+    return ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX);
+}
Reviewed-by on v3 5/17
#17 lore raw

Re: [PATCH v3 05/17] hw/usb/hcd-ehci: Replace DPRINTF debug logs with trace events

In reply to Jamin Lin
Quoted 14 lines · Jamin Lin

Replace legacy DPRINTF() debug logging in the EHCI host controller implementation with QEMU trace events.

The EHCI_DEBUG macro and associated DPRINTF() definitions are removed as they are no longer needed.

Trace events are added for:
- packet submission and completion
- queue head execution state
- periodic schedule advancement
- ITD error conditions
- port state handling
- skipped microframes

Quoted 8 lines

No functional change.

---
  hw/usb/hcd-ehci.h   | 10 ---------
  hw/usb/hcd-ehci.c   | 49 +++++++++++++++++++--------------------------
  hw/usb/trace-events |  9 +++++++++
  3 files changed, 30 insertions(+), 38 deletions(-)
author
#37 lore raw

RE: [PATCH v3 09/17] hw/usb/hcd-ehci: Implement 64-bit QH descriptor addressing

In reply to Philippe Mathieu-Daudé

Hi Philippe,

Quoted 84 lines

Subject: Re: [PATCH v3 09/17] hw/usb/hcd-ehci: Implement 64-bit QH descriptor addressing

Quoted 32 lines · Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register supplies the upper 32 bits which are concatenated with 32-bit link pointer fields to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

The current implementation assumes 32-bit QH descriptor addresses and directly uses link pointer values without applying the CTRLDSSEGMENT upper dword.

Introduce a helper, ehci_get_desc_addr(), to construct full 64-bit descriptor addresses when 64-bit capability is enabled. Update QH traversal paths (async list walk, horizontal QH link, and periodic schedule entry handling) to use the translated 64-bit addresses.

EHCI 64-bit buffer pointer fields are defined in Appendix B as split 32-bit low/high parts located at separate offsets, rather than a single contiguous 64-bit field. Therefore, the buffer pointers cannot be represented as uint64_t bufptr[5] without violating the descriptor layout defined by the specification.

Introduce ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] fields. Use this helper when calculating transfer buffer addresses so that data buffers above 4GB are correctly handled.

Also add bufptr_hi[5] to EHCIqh to support 64-bit buffer pointer fields as defined in Appendix B.

When 64-bit capability is disabled, descriptor addresses remain 32-bit and existing behaviour is unchanged.

Quoted 8 lines

Note: Similar split 64-bit buffer pointer handling is required for qTD, iTD and siTD descriptors, which will be addressed in follow-up changes.

---
  hw/usb/hcd-ehci.h   |  4 ++++
  hw/usb/hcd-ehci.c   | 48

++++++++++++++++++++++++++++++++++-----------

Quoted 30 lines
  hw/usb/trace-events |  2 +-
  3 files changed, 42 insertions(+), 12 deletions(-)
hw/usb/hcd-ehci.h index renamed from hw/usb/hcd-ehci.h +0 -0

fe4c508c32..773f376834 100644

hw/usb/hcd-ehci.h +4 -0
@@ -141,6 +141,9 @@ typedef struct EHCIqtd {
  #define QTD_BUFPTR_SH                 12
  } EHCIqtd;
 
+/* QH overlay: altnext_qtd, token, bufptr[5], bufptr_hi[5] */ #define
+EHCI_QH_OVERLAY_COUNT 12
+
  /*
   * EHCI spec version 1.0 Section 3.6
   */
@@ -194,6 +197,7 @@ typedef struct EHCIqh {
  #define BUFPTR_FRAMETAG_MASK          0x0000001f
  #define BUFPTR_SBYTES_MASK            0x00000fe0
  #define BUFPTR_SBYTES_SH              5
+    uint32_t bufptr_hi[5];
  } EHCIqh;
 
  enum async_state {
hw/usb/hcd-ehci.c index renamed from hw/usb/hcd-ehci.c +0 -0

b9e4251d62..37ae637e08 100644

hw/usb/hcd-ehci.c static const char *addr2str(hwaddr addr) · line 147

Quoted 1 line
hw/usb/hcd-ehci.c +0 -0
@@ -147,6 +147,24 @@ static const char *addr2str(hwaddr addr)
      return nr2str(ehci_mmio_names, ARRAY_SIZE(ehci_mmio_names),

addr);

hw/usb/hcd-ehci.c static const char *addr2str(hwaddr addr) · line 151

Quoted 4 lines
hw/usb/hcd-ehci.c +2 -0
  }
 
+static uint64_t ehci_get_buf_addr(EHCIState *s, uint32_t hi, uint32_t low,
+                                  uint32_t mask)

Rename @mask as @low_mask? Or directly use 64-bit?

Quoted 8 lines

static uint64_t ehci_get_buf_addr(const EHCIState *s,

                                   uint32_t hi, uint32_t lo,
                                   uint64_t mask) {
     uint64_t addr = s->caps_64bit_addr ? deposit64(lo, 0, 32, hi): lo;

     return addr & mask;
}

Thanks for the review and suggestion.

In the current code, the mask (e.g. QTD_BUFPTR_MASK) is defined as a 32-bit value and is intended to apply only to the low 32 bits of the buffer pointer. If we change the function to take a 64-bit mask and apply it to the full 64-bit address (as in the proposed version),
the upper 32 bits (bits 63–32) may be unintentionally cleared when the mask is only 32-bit wide.

For example:

#define QTD_BUFPTR_MASK 0xfffff000

In this case, applying "addr & mask" after combining hi/lo would zero out the high 32 bits, which breaks 64-bit addressing. To preserve the original semantics, the mask should only be applied to the low part before combining with the high 32 bits. So I suggest renaming the parameter to make this explicit. I will change this function as below. What do you think?

static uint64_t ehci_get_buf_addr(EHCIState *s, uint32_t hi, uint32_t lo,

                                  uint32_t lo_mask)
{
    uint64_t addr = lo & lo_mask;

    if (s->caps_64bit_addr) {
        addr = deposit64(addr, 32, 32, hi);
    }

    return addr;
}

Thanks,
Jamin

Quoted 13 lines
Quoted 13 lines
hw/usb/hcd-ehci.c +13 -0
+{
+    uint64_t addr;
+
+    addr = (uint64_t)(low & mask);
+    if (s->caps_64bit_addr) {
+        addr |= (uint64_t)hi << 32;
+    }
+
+    return addr;
+}
+
+static uint64_t ehci_get_desc_addr(EHCIState *s, uint32_t low) {
+    return ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX); }
#38 lore raw

Re: [PATCH v3 09/17] hw/usb/hcd-ehci: Implement 64-bit QH descriptor addressing

Quoted 61 lines · Jamin Lin

Hi Philippe,

Quoted 59 lines

Subject: Re: [PATCH v3 09/17] hw/usb/hcd-ehci: Implement 64-bit QH descriptor addressing

Quoted 32 lines · Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register supplies the upper 32 bits which are concatenated with 32-bit link pointer fields to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

The current implementation assumes 32-bit QH descriptor addresses and directly uses link pointer values without applying the CTRLDSSEGMENT upper dword.

Introduce a helper, ehci_get_desc_addr(), to construct full 64-bit descriptor addresses when 64-bit capability is enabled. Update QH traversal paths (async list walk, horizontal QH link, and periodic schedule entry handling) to use the translated 64-bit addresses.

EHCI 64-bit buffer pointer fields are defined in Appendix B as split 32-bit low/high parts located at separate offsets, rather than a single contiguous 64-bit field. Therefore, the buffer pointers cannot be represented as uint64_t bufptr[5] without violating the descriptor layout defined by the specification.

Introduce ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] fields. Use this helper when calculating transfer buffer addresses so that data buffers above 4GB are correctly handled.

Also add bufptr_hi[5] to EHCIqh to support 64-bit buffer pointer fields as defined in Appendix B.

When 64-bit capability is disabled, descriptor addresses remain 32-bit and existing behaviour is unchanged.

Quoted 8 lines

Note: Similar split 64-bit buffer pointer handling is required for qTD, iTD and siTD descriptors, which will be addressed in follow-up changes.

---
   hw/usb/hcd-ehci.h   |  4 ++++
   hw/usb/hcd-ehci.c   | 48

++++++++++++++++++++++++++++++++++-----------

hw/usb/hcd-ehci.c static const char *addr2str(hwaddr addr) · line 151

Quoted 6 lines
   hw/usb/trace-events |  2 +-
   3 files changed, 42 insertions(+), 12 deletions(-)
hw/usb/hcd-ehci.c +2 -0
+static uint64_t ehci_get_buf_addr(EHCIState *s, uint32_t hi, uint32_t low,
+                                  uint32_t mask)

Rename @mask as @low_mask? Or directly use 64-bit?

static uint64_t ehci_get_buf_addr(const EHCIState *s,

                                    uint32_t hi, uint32_t lo,
                                    uint64_t mask) {
      uint64_t addr = s->caps_64bit_addr ? deposit64(lo, 0, 32, hi): lo;
Quoted 8 lines
Quoted 3 lines
      return addr & mask;
}

Thanks for the review and suggestion.

In the current code, the mask (e.g. QTD_BUFPTR_MASK) is defined as a 32-bit value and is intended to apply only to the low 32 bits of the buffer pointer. If we change the function to take a 64-bit mask and apply it to the full 64-bit address (as in the proposed version),
the upper 32 bits (bits 63–32) may be unintentionally cleared when the mask is only 32-bit wide.

Right.

Quoted 10 lines

For example:

#define QTD_BUFPTR_MASK 0xfffff000

In this case, applying "addr & mask" after combining hi/lo would zero out the high 32 bits, which breaks 64-bit addressing. To preserve the original semantics, the mask should only be applied to the low part before combining with the high 32 bits. So I suggest renaming the parameter to make this explicit. I will change this function as below. What do you think?

static uint64_t ehci_get_buf_addr(

"const"

EHCIState *s, uint32_t hi, uint32_t lo,

Quoted 10 lines
                                   uint32_t lo_mask)
{
     uint64_t addr = lo & lo_mask;

     if (s->caps_64bit_addr) {
         addr = deposit64(addr, 32, 32, hi);
     }

     return addr;
}

Perfect, using "const":

Reviewed-by on v3 10/17
#41 lore raw

Re: [PATCH v3 10/17] hw/usb/hcd-ehci: Implement 64-bit qTD descriptor addressing

In reply to Jamin Lin
Quoted 16 lines · Jamin Lin

EHCI supports 64-bit addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register provides the upper 32 bits that are concatenated with 32-bit link pointer values to form 64-bit control data structure addresses (EHCI 1.0, section 2.3.5 and Appendix B).

qTD link pointers (current_qtd/next_qtd/altnext_qtd and qTD.next) are stored as 32-bit values in the data structures and must be expanded to full 64-bit descriptor addresses when 64-bit mode is enabled. Update the qTD traversal paths to use ehci_get_desc_addr() when following link pointers.

Appendix B also defines high dword fields for qTD buffer pointers. Add bufptr_hi[5] to EHCIqtd and extend qTD fetch and QH overlay handling to load and propagate the high buffer pointer fields.

Quoted 8 lines

When 64-bit capability is disabled, descriptor and buffer addresses remain 32-bit and existing behaviour is unchanged.

---
  hw/usb/hcd-ehci.h |  1 +
  hw/usb/hcd-ehci.c | 19 ++++++++++++-------
  2 files changed, 13 insertions(+), 7 deletions(-)
Reviewed-by on v3 11/17
#43 lore raw

Re: [PATCH v3 11/17] hw/usb/hcd-ehci: Implement 64-bit iTD descriptor addressing

In reply to Jamin Lin
Quoted 17 lines · Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register provides the upper 32 bits that are concatenated with 32-bit link pointer values to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

iTD link pointers are stored as 32-bit values and must be expanded to full 64-bit descriptor addresses when 64-bit mode is enabled. Update the iTD traversal path to use ehci_get_desc_addr() when following link pointers.

Appendix B also defines high dword fields for iTD buffer pointers. Add bufptr_hi[7] to EHCIitd and use ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] fields when processing isochronous transfers. This allows buffers above 4GB to be handled correctly.

Quoted 8 lines

When 64-bit capability is disabled, descriptor and buffer addresses remain 32-bit and existing behaviour is unchanged.

---
  hw/usb/hcd-ehci.h | 1 +
  hw/usb/hcd-ehci.c | 9 ++++++---
  2 files changed, 7 insertions(+), 3 deletions(-)
#49 lore raw

Re: [PATCH v3 13/17] hw/usb/hcd-ehci: Add descriptor address offset property

In reply to Jamin Lin
Quoted 37 lines · Jamin Lin

When 64-bit addressing is supported, the Linux EHCI driver programs the segment register to zero. See ehci_run function:
https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci-hcd.c

The driver comment also notes that descriptor structures allocated from the DMA pool use segment zero semantics.

Descriptor memory is allocated using the DMA API. The platform driver configures a 64-bit DMA mask so memory can be allocated above 4GB. See ehci_platform_probe function:
https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci-platform.c

On AST2700 platforms, system DRAM is mapped above 4GB at 0x400000000. As a result, descriptor addresses constructed directly from the guest EHCI registers do not match the actual system address used by the controller when fetching queue heads (QH) and queue element transfer descriptors (qTD).

Introduce a descriptor-addr-offset property so platforms can provide an address offset applied when constructing descriptor addresses. This allows systems where DRAM resides above 4GB to correctly access EHCI descriptors while keeping the default behavior unchanged for existing machines.

---
  hw/usb/hcd-ehci.h        | 1 +
  hw/usb/hcd-ehci-pci.c    | 2 ++
  hw/usb/hcd-ehci-sysbus.c | 2 ++
  hw/usb/hcd-ehci.c        | 7 ++++++-
  4 files changed, 11 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.h +0 -0
index 6406f536e8..05d8db4a5d 100644

hw/usb/hcd-ehci.h struct EHCIState { · line 267

Quoted 4 lines
hw/usb/hcd-ehci.h +1 -0
@@ -264,6 +264,7 @@ struct EHCIState {
      /* properties */
      uint32_t maxframes;
      bool caps_64bit_addr;
+    uint64_t descriptor_addr_offset;

I note OHCI uses "dma_addr_t dma_offset;" here.

Quoted 8 lines
hw/usb/hcd-ehci.h +0 -0
      /*
       *  EHCI spec version 1.0 Section 2.3
hw/usb/hcd-ehci-pci.c +0 -0
index 2ea8549db9..115d05ede0 100644

hw/usb/hcd-ehci-pci.c static const Property ehci_pci_properties[] = { · line 144

Quoted 6 lines
hw/usb/hcd-ehci-pci.c +2 -0
@@ -139,6 +139,8 @@ static const Property ehci_pci_properties[] = {
      DEFINE_PROP_UINT32("maxframes", EHCIPCIState, ehci.maxframes, 128),
      DEFINE_PROP_BOOL("caps-64bit-addr", EHCIPCIState, ehci.caps_64bit_addr,
                       false),
+    DEFINE_PROP_UINT64("descriptor-addr-offset", EHCIPCIState,
+                       ehci.descriptor_addr_offset, 0),
  };

I note OHCI uses DEFINE_PROP_DMAADDR() here. Single use, maybe better remove it and stick to uint64_t as you use.

Could we have a DEFINE_EHCI_COMMON_PROPERTIES() macro for common entries instead, such DEFINE_SDHCI_COMMON_PROPERTIES()?

Quoted 13 lines
hw/usb/hcd-ehci-pci.c +0 -0
  static const VMStateDescription vmstate_ehci_pci = {
hw/usb/hcd-ehci-sysbus.c +2 -0
index 61215e9f3d..df138fb339 100644
@@ -36,6 +36,8 @@ static const Property ehci_sysbus_properties[] = {
                       false),
      DEFINE_PROP_BOOL("caps-64bit-addr", EHCISysBusState, ehci.caps_64bit_addr,
                       false),
+    DEFINE_PROP_UINT64("descriptor-addr-offset", EHCISysBusState,
+                       ehci.descriptor_addr_offset, 0),
  };
Reviewed-by on v3 14/17
#57 lore raw

Re: [PATCH v3 14/17] hw/arm/aspeed_ast27x0: Enable 64-bit EHCI DMA addressing

In reply to Jamin Lin
Quoted 7 lines · Jamin Lin

AST2700 supports a 64-bit DRAM address space. Therefore, DMA transactions must be capable of accessing 64-bit addresses.

Enable the "caps-64bit-addr" property for the EHCI controllers on AST2700 so that USB DMA operations can correctly handle 64-bit memory addresses.

Quoted 5 lines
---
  hw/arm/aspeed_ast27x0.c | 2 ++
  1 file changed, 2 insertions(+)
#59 lore raw

Re: [PATCH v3 15/17] hw/arm/aspeed_ast27x0: Set EHCI descriptor address offset

In reply to Jamin Lin
Quoted 28 lines · Jamin Lin

On AST2700 platforms, system DRAM is mapped above 4GB with the base address at 0x400000000.

The Linux EHCI driver programs the segment register to zero when 64-bit addressing is supported. As a result, descriptor addresses derived from the EHCI registers do not include the DRAM base address.

Descriptor memory is allocated through the DMA API with a 64-bit DMA mask, which allows descriptors to be placed in DRAM above 4GB. When running on AST2700, this means EHCI queue heads (QH) and queue element transfer descriptors (qTD) reside at addresses starting from 0x400000000.

Set the descriptor-addr-offset property to the DRAM base so the emulated EHCI controller can construct the correct descriptor addresses when accessing system memory.

---
  hw/arm/aspeed_ast27x0.c | 3 +++
  1 file changed, 3 insertions(+)
hw/arm/aspeed_ast27x0.c +0 -0
index 4a1f7cad73..3cdbf78ac1 100644

hw/arm/aspeed_ast27x0.c static void aspeed_soc_ast2700_realize(DeviceState *dev, Error **errp) · line 863

Quoted 6 lines
hw/arm/aspeed_ast27x0.c +3 -0
@@ -858,6 +858,9 @@ static void aspeed_soc_ast2700_realize(DeviceState *dev, Error **errp)
      for (i = 0; i < sc->ehcis_num; i++) {
          object_property_set_bool(OBJECT(&s->ehci[i]), "caps-64bit-addr", true,
                                   &error_abort);
+        object_property_set_int(OBJECT(&s->ehci[i]), "descriptor-addr-offset",
+                                sc->memmap[ASPEED_DEV_SDRAM],
+                                &error_abort);

Squash with previous? Or invert 14 <-> 15 order?

Reviewed-by on v3 16/17
#62 lore raw

Re: [PATCH v3 16/17] tests/functional/arm/test_aspeed_ast2600_sdk: Add USB EHCI test for AST2600 SDK

In reply to Jamin Lin
Quoted 11 lines · Jamin Lin

Add a functional test to verify USB EHCI support on the AST2600 SDK machine by attaching a USB keyboard device and checking its enumeration via lsusb.

This introduces a helper routine that runs lsusb in the guest and validates that the emulated "QEMU USB Keyboard" is detected.

---
  tests/functional/arm/test_aspeed_ast2600_sdk.py | 7 +++++++
  1 file changed, 7 insertions(+)

This should be patch #1 :)

Reviewed-by on v3 17/17
#67 lore raw

Re: [PATCH v3 17/17] tests/functional/aarch64/test_aspeed_ast2700: Add USB EHCI test for AST2700 A1/A2

In reply to Jamin Lin
Quoted 4 lines · Jamin Lin

Add a functional test to verify USB EHCI support on the AST2700 A2/A1 by attaching a USB keyboard device and checking its enumeration via lsusb.

Quoted 8 lines

This introduces a helper routine that runs lsusb in the guest and validates that the emulated "QEMU USB Keyboard" is detected.

---
  tests/functional/aarch64/test_aspeed_ast2700a1.py | 7 +++++++
  tests/functional/aarch64/test_aspeed_ast2700a2.py | 7 +++++++
  2 files changed, 14 insertions(+)
#45 lore raw

Re: [PATCH v3 12/17] hw/usb/hcd-ehci: Implement 64-bit siTD descriptor addressing

In reply to Jamin Lin
Quoted 31 lines · Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register provides the upper 32 bits that are concatenated with 32-bit link pointer values to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

siTD link pointers are stored as 32-bit values and must be expanded to full 64-bit descriptor addresses when 64-bit mode is enabled. Update the siTD traversal path to use ehci_get_desc_addr() when following link pointers.

Appendix B also defines high dword fields for siTD buffer pointers. Add bufptr_hi[] fields to EHCIsitd and use ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] when processing split isochronous transfers. This allows buffers above 4GB to be handled correctly.

When 64-bit capability is disabled, descriptor and buffer addresses remain 32-bit and existing behaviour is unchanged.

---
  hw/usb/hcd-ehci.h | 1 +
  hw/usb/hcd-ehci.c | 2 +-
  2 files changed, 2 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.h +0 -0
index af8c080c60..6406f536e8 100644

hw/usb/hcd-ehci.h typedef struct EHCIsitd { · line 112

Quoted 4 lines
hw/usb/hcd-ehci.h +1 -0
@@ -109,6 +109,7 @@ typedef struct EHCIsitd {
  #define SITD_BUFPTR_TCNT_MASK         0x00000007
 
      uint32_t backptr;                 /* Standard next link pointer */
+    uint32_t bufptr_hi[2];

Not used?

Quoted 16 lines
hw/usb/hcd-ehci.h +0 -0
  } EHCIsitd;
 
  /*
hw/usb/hcd-ehci.c +1 -1
index 3db1c25e4d..e1d5ecbad7 100644
@@ -1791,7 +1791,7 @@ static int ehci_state_fetchsitd(EHCIState *ehci, int async)
          warn_report("Skipping active siTD");
      }
 
-    ehci_set_fetch_addr(ehci, async, sitd.next);
+    ehci_set_fetch_addr(ehci, async, ehci_get_desc_addr(ehci, sitd.next));
      ehci_set_state(ehci, async, EST_FETCHENTRY);
      return 1;
  }
Reviewed-by on v3 1/17
#7 lore raw

Re: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

In reply to Jamin Lin
Quoted 11 lines · Jamin Lin

The EHCIfstn structure, defined according to EHCI spec section 3.7, is currently unused in the EHCI implementation. Remove it to reduce unused code and improve maintainability.

Additionally, drop a block of disabled (#if 0) code in ehci_state_advqueue() that was never implemented. This code path is incomplete and has no functional impact.

No functional change.

Thanks,

C.

Quoted 45 lines
---
  hw/usb/hcd-ehci.h |  7 -------
  hw/usb/hcd-ehci.c | 11 -----------
  2 files changed, 18 deletions(-)
hw/usb/hcd-ehci.h +0 -7
index 0ae8c06331..3aec0e047c 100644
@@ -202,13 +202,6 @@ typedef struct EHCIqh {
  #define BUFPTR_SBYTES_SH              5
  } EHCIqh;
 
-/*  EHCI spec version 1.0 Section 3.7
- */
-typedef struct EHCIfstn {
-    uint32_t next;                    /* Standard next link pointer */
-    uint32_t backptr;                 /* Standard next link pointer */
-} EHCIfstn;
-
  enum async_state {
      EHCI_ASYNC_NONE = 0,
      EHCI_ASYNC_INITIALIZED,
hw/usb/hcd-ehci.c +0 -11
index 57f930b099..5ea8461f70 100644
@@ -1756,17 +1756,6 @@ static int ehci_state_fetchsitd(EHCIState *ehci, int async)
  /* Section 4.10.2 - paragraph 3 */
  static int ehci_state_advqueue(EHCIQueue *q)
  {
-#if 0
-    /* TO-DO: 4.10.2 - paragraph 2
-     * if I-bit is set to 1 and QH is not active
-     * go to horizontal QH
-     */
-    if (I-bit set) {
-        ehci_set_state(ehci, async, EST_HORIZONTALQH);
-        goto out;
-    }
-#endif
-
      /*
       * want data and alt-next qTD is valid
       */
#14 lore raw

Re: [PATCH v3 04/17] hw/usb/hcd-ehci.c: Replace fprintf(stderr, ...) with qemu_log_mask(LOG_GUEST_ERROR)

In reply to Jamin Lin
Quoted 31 lines · Jamin Lin

Replace direct fprintf(stderr, ...) calls with qemu_log_mask() using LOG_GUEST_ERROR in the EHCI controller implementation.

Direct writes to stderr are discouraged in QEMU as they bypass the logging framework and cannot be filtered or controlled at runtime. Using qemu_log_mask() ensures that guest-visible errors are properly categorized and can be managed via QEMU logging options.

All affected messages correspond to guest-triggerable error conditions, so LOG_GUEST_ERROR is used consistently.

This change improves integration with QEMU's logging infrastructure and aligns with upstream coding guidelines.

---
  hw/usb/hcd-ehci.c | 43 +++++++++++++++++++++++++------------------
  1 file changed, 25 insertions(+), 18 deletions(-)
hw/usb/hcd-ehci.c +1 -1
@@ -532,7 +533,7 @@ static void ehci_free_packet(EHCIPacket *p)
      }
      if (p->async == EHCI_ASYNC_FINISHED &&
              p->packet.status == USB_RET_SUCCESS) {
-        fprintf(stderr,
+        qemu_log_mask(LOG_GUEST_ERROR,
                  "EHCI: Dropping completed packet from halted %s ep %02X\n",
                  (p->pid == USB_TOKEN_IN) ? "in" : "out",
                  get_field(p->queue->qh.epchar, QH_EPCHAR_EP));

hw/usb/hcd-ehci.c static void ehci_opreg_write(void *ptr, hwaddr addr, · line 1047

Quoted 7 lines
hw/usb/hcd-ehci.c +2 -2
@@ -1042,8 +1043,9 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,
 
          /* not supporting dynamic frame list size at the moment */
          if ((val & USBCMD_FLS) && !(s->usbcmd & USBCMD_FLS)) {
-            fprintf(stderr, "attempt to set frame list size -- value %d\n",
-                    (int)val & USBCMD_FLS);
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "attempt to set frame list size -- value %d\n",

PRId64

Quoted 15 lines
hw/usb/hcd-ehci.c +3 -1
+                          (int)val & USBCMD_FLS);
              val &= ~USBCMD_FLS;
          }
 
@@ -1584,7 +1589,8 @@ static int ehci_state_fetchentry(EHCIState *ehci, int async)
 
      /* section 4.8, only QH in async schedule */
      if (async && (NLPTR_TYPE_GET(entry) != NLPTR_TYPE_QH)) {
-        fprintf(stderr, "non queue head request in async schedule\n");
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "non queue head request in async schedule\n");
          return -1;
      }
 

hw/usb/hcd-ehci.c static int ehci_state_fetchentry(EHCIState *ehci, int async) · line 1615

Quoted 5 lines
hw/usb/hcd-ehci.c +1 -1
@@ -1606,7 +1612,7 @@ static int ehci_state_fetchentry(EHCIState *ehci, int async)
 
      default:
          /* TODO: handle FSTN type */
-        fprintf(stderr, "FETCHENTRY: entry at %X is of type %u "
+        qemu_log_mask(LOG_GUEST_ERROR, "FETCHENTRY: entry at %X is of type %u "

"0x%x"

hw/usb/hcd-ehci.c static void ehci_advance_state(EHCIState *ehci, int async) · line 2121

Quoted 9 lines
hw/usb/hcd-ehci.c +1 -1
                  "which is not supported yet\n", entry, NLPTR_TYPE_GET(entry));
          return -1;
      }
@@ -2112,13 +2118,14 @@ static void ehci_advance_state(EHCIState *ehci, int async)
              break;
 
          default:
-            fprintf(stderr, "Bad state!\n");
+            qemu_log_mask(LOG_GUEST_ERROR, "Bad state!\n");

Just remove for "bad state" entries?

Quoted 12 lines
hw/usb/hcd-ehci.c +2 -1
              g_assert_not_reached();
          }
 
          if (again < 0 || itd_count > 16) {
              /* TODO: notify guest (raise HSE irq?) */
-            fprintf(stderr, "processing error - resetting ehci HC\n");
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "processing error - resetting ehci HC\n");
              ehci_reset(ehci);
              again = 0;
          }

hw/usb/hcd-ehci.c static void ehci_advance_async_state(EHCIState *ehci) · line 2186

Quoted 6 lines
hw/usb/hcd-ehci.c +1 -1
@@ -2175,7 +2182,7 @@ static void ehci_advance_async_state(EHCIState *ehci)
 
      default:
          /* this should only be due to a developer mistake */
-        fprintf(stderr, "ehci: Bad asynchronous state %d. "
+        qemu_log_mask(LOG_GUEST_ERROR, "ehci: Bad asynchronous state %d. "
                  "Resetting to active\n", ehci->astate);

Ditto.

hw/usb/hcd-ehci.c static void ehci_advance_periodic_state(EHCIState *ehci) · line 2237

Quoted 10 lines
hw/usb/hcd-ehci.c +1 -1
          g_assert_not_reached();
      }
@@ -2225,7 +2232,7 @@ static void ehci_advance_periodic_state(EHCIState *ehci)
 
      default:
          /* this should only be due to a developer mistake */
-        fprintf(stderr, "ehci: Bad periodic state %d. "
+        qemu_log_mask(LOG_GUEST_ERROR, "ehci: Bad periodic state %d. "
                  "Resetting to active\n", ehci->pstate);
          g_assert_not_reached();

Ditto.

Quoted 1 line
(quoted diff) +0 -0
      }
Reviewed-by on v3 3/17
#11 lore raw

Re: [PATCH v3 03/17] hw/usb/hcd-ehci.c: Fix coding style issues reported by checkpatch

In reply to Jamin Lin
Quoted 362 lines · Jamin Lin

No functional change.

---
  hw/usb/hcd-ehci.c | 126 +++++++++++++++++++++++++---------------------
  1 file changed, 69 insertions(+), 57 deletions(-)
hw/usb/hcd-ehci.c +66 -55
index 5ea8461f70..9b92384bac 100644
@@ -41,21 +41,23 @@
  #define FRAME_TIMER_NS   (NANOSECONDS_PER_SECOND / FRAME_TIMER_FREQ)
  #define UFRAME_TIMER_NS  (FRAME_TIMER_NS / 8)
 
-#define NB_MAXINTRATE    8        // Max rate at which controller issues ints
-#define BUFF_SIZE        5*4096   // Max bytes to transfer per transaction
-#define MAX_QH           100      // Max allowable queue heads in a chain
+#define NB_MAXINTRATE    8      /* Max rate at which controller issues ints */
+#define BUFF_SIZE        (5 * 4096) /* Max bytes to transfer per transaction */
+#define MAX_QH           100      /* Max allowable queue heads in a chain */
  #define MIN_UFR_PER_TICK 24       /* Min frames to process when catching up */
  #define PERIODIC_ACTIVE  512      /* Micro-frames */
 
-/*  Internal periodic / asynchronous schedule state machine states
+/*
+ * Internal periodic / asynchronous schedule state machine states
   */
  typedef enum {
      EST_INACTIVE = 1000,
      EST_ACTIVE,
      EST_EXECUTING,
      EST_SLEEPING,
-    /*  The following states are internal to the state machine function
-    */
+    /*
+     * The following states are internal to the state machine function
+     */
      EST_WAITLISTHEAD,
      EST_FETCHENTRY,
      EST_FETCHQH,
@@ -71,13 +73,13 @@ typedef enum {
  /* macros for accessing fields within next link pointer entry */
  #define NLPTR_GET(x)             ((x) & 0xffffffe0)
  #define NLPTR_TYPE_GET(x)        (((x) >> 1) & 3)
-#define NLPTR_TBIT(x)            ((x) & 1)  // 1=invalid, 0=valid
+#define NLPTR_TBIT(x)            ((x) & 1)  /* 1=invalid, 0=valid */
 
  /* link pointer types */
-#define NLPTR_TYPE_ITD           0     // isoc xfer descriptor
-#define NLPTR_TYPE_QH            1     // queue head
-#define NLPTR_TYPE_STITD         2     // split xaction, isoc xfer descriptor
-#define NLPTR_TYPE_FSTN          3     // frame span traversal node
+#define NLPTR_TYPE_ITD           0     /* isoc xfer descriptor */
+#define NLPTR_TYPE_QH            1     /* queue head */
+#define NLPTR_TYPE_STITD         2     /* split xaction, isoc xfer descriptor */
+#define NLPTR_TYPE_FSTN          3     /* frame span traversal node */
 
  #define SET_LAST_RUN_CLOCK(s) \
      (s)->last_run_ns = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
@@ -88,10 +90,10 @@ typedef enum {
 
  #define set_field(data, newval, field) do { \
      uint32_t val = *data; \
-    val &= ~ field##_MASK; \
+    val &= ~field##_MASK; \
      val |= ((newval) << field##_SH) & field##_MASK; \
      *data = val; \
-    } while(0)
+    } while (0)
 
  static const char *ehci_state_names[] = {
      [EST_INACTIVE]     = "INACTIVE",
@@ -472,8 +474,10 @@ static bool ehci_verify_pid(EHCIQueue *q, EHCIqtd *qtd)
      }
  }
 
-/* Finish executing and writeback a packet outside of the regular
-   fetchqh -> fetchqtd -> execute -> writeback cycle */
+/*
+ * Finish executing and writeback a packet outside of the regular
+ * fetchqh -> fetchqtd -> execute -> writeback cycle
+ */
  static void ehci_writeback_async_complete_packet(EHCIPacket *p)
  {
      EHCIQueue *q = p->queue;
@@ -733,7 +737,7 @@ static void ehci_detach(USBPort *port)
      ehci_queues_rip_device(s, port->dev, 0);
      ehci_queues_rip_device(s, port->dev, 1);
 
-    *portsc &= ~(PORTSC_CONNECT|PORTSC_PED|PORTSC_SUSPEND);
+    *portsc &= ~(PORTSC_CONNECT | PORTSC_PED | PORTSC_SUSPEND);
      *portsc |= PORTSC_CSC;
 
      ehci_raise_irq(s, USBSTS_PCD);
@@ -858,7 +862,7 @@ void ehci_reset(void *opaque)
       * Do the detach before touching portsc, so that it correctly gets send to
       * us or to our companion based on PORTSC_POWNER before the reset.
       */
-    for(i = 0; i < EHCI_PORTS; i++) {
+    for (i = 0; i < EHCI_PORTS; i++) {
          devs[i] = s->ports[i].dev;
          if (devs[i] && devs[i]->attached) {
              usb_detach(&s->ports[i]);
@@ -877,7 +881,7 @@ void ehci_reset(void *opaque)
      s->astate = EST_INACTIVE;
      s->pstate = EST_INACTIVE;
 
-    for(i = 0; i < EHCI_PORTS; i++) {
+    for (i = 0; i < EHCI_PORTS; i++) {
          if (s->companion_ports[i]) {
              s->portsc[i] = PORTSC_POWNER | PORTSC_PPOWER;
          } else {
@@ -942,8 +946,9 @@ static void handle_port_owner_write(EHCIState *s, int port, uint32_t owner)
      uint32_t *portsc = &s->portsc[port];
      uint32_t orig;
 
-    if (s->companion_ports[port] == NULL)
+    if (s->companion_ports[port] == NULL) {
          return;
+    }
 
      owner = owner & PORTSC_POWNER;
      orig  = *portsc & PORTSC_POWNER;
@@ -988,7 +993,7 @@ static void ehci_port_write(void *ptr, hwaddr addr,
          trace_usb_ehci_port_reset(port, 1);
      }
 
-    if (!(val & PORTSC_PRESET) &&(*portsc & PORTSC_PRESET)) {
+    if (!(val & PORTSC_PRESET) && (*portsc & PORTSC_PRESET)) {
          trace_usb_ehci_port_reset(port, 0);
          if (dev && dev->attached) {
              usb_port_reset(&s->ports[port]);
@@ -1065,8 +1070,10 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,
          break;
 
      case USBSTS:
-        val &= USBSTS_RO_MASK;              // bits 6 through 31 are RO
-        ehci_clear_usbsts(s, val);          // bits 0 through 5 are R/WC
+        /* bits 6 through 31 are RO */
+        val &= USBSTS_RO_MASK;
+        /* bits 0 through 5 are R/WC */
+        ehci_clear_usbsts(s, val);
          val = s->usbsts;
          ehci_update_irq(s);
          break;
@@ -1131,8 +1138,7 @@ static void ehci_flush_qh(EHCIQueue *q)
      put_dwords(q->ehci, addr + 3 * sizeof(uint32_t), qh + 3, dwords - 3);
  }
 
-// 4.10.2
-
+/* 4.10.2 */
  static int ehci_qh_do_overlay(EHCIQueue *q)
  {
      EHCIPacket *p = QTAILQ_FIRST(&q->packets);
@@ -1145,8 +1151,7 @@ static int ehci_qh_do_overlay(EHCIQueue *q)
      assert(p != NULL);
      assert(p->qtdaddr == q->qtdaddr);
 
-    // remember values in fields to preserve in qh after overlay
-
+    /*  remember values in fields to preserve in qh after overlay */
      dtoggle = q->qh.token & QTD_TOKEN_DTOGGLE;
      ping    = q->qh.token & QTD_TOKEN_PING;
 
@@ -1170,7 +1175,7 @@ static int ehci_qh_do_overlay(EHCIQueue *q)
      }
 
      if (!(q->qh.epchar & QH_EPCHAR_DTC)) {
-        // preserve QH DT bit
+        /* preserve QH DT bit */
          q->qh.token &= ~QTD_TOKEN_DTOGGLE;
          q->qh.token |= dtoggle;
      }
@@ -1397,9 +1402,7 @@ static int ehci_execute(EHCIPacket *p, const char *action)
      return 1;
  }
 
-/*  4.7.2
- */
-
+/* 4.7.2 */
  static int ehci_process_itd(EHCIState *ehci,
                              EHCIitd *itd,
                              uint32_t addr)
@@ -1411,13 +1414,13 @@ static int ehci_process_itd(EHCIState *ehci,
 
      ehci->periodic_sched_active = PERIODIC_ACTIVE;
 
-    dir =(itd->bufptr[1] & ITD_BUFPTR_DIRECTION);
+    dir = (itd->bufptr[1] & ITD_BUFPTR_DIRECTION);
      devaddr = get_field(itd->bufptr[0], ITD_BUFPTR_DEVADDR);
      endp = get_field(itd->bufptr[0], ITD_BUFPTR_EP);
      max = get_field(itd->bufptr[1], ITD_BUFPTR_MAXPKT);
      mult = get_field(itd->bufptr[2], ITD_BUFPTR_MULT);
 
-    for(i = 0; i < 8; i++) {
+    for (i = 0; i < 8; i++) {
          if (itd->transact[i] & ITD_XACT_ACTIVE) {
              pg   = get_field(itd->transact[i], ITD_XACT_PGSEL);
              off  = itd->transact[i] & ITD_XACT_OFFSET_MASK;
@@ -1513,8 +1516,9 @@ static int ehci_process_itd(EHCIState *ehci,
  }
 
 
-/*  This state is the entry point for asynchronous schedule
- *  processing.  Entry here constitutes a EHCI start event state (4.8.5)
+/*
+ * This state is the entry point for asynchronous schedule
+ * processing.  Entry here constitutes a EHCI start event state (4.8.5)
   */
  static int ehci_state_waitlisthead(EHCIState *ehci,  int async)
  {
@@ -1531,7 +1535,7 @@ static int ehci_state_waitlisthead(EHCIState *ehci,  int async)
      ehci_queues_rip_unused(ehci, async);
 
      /*  Find the head of the list (4.9.1.1) */
-    for(i = 0; i < MAX_QH; i++) {
+    for (i = 0; i < MAX_QH; i++) {
          if (get_dwords(ehci, NLPTR_GET(entry), (uint32_t *) &qh,
                         sizeof(EHCIqh) >> 2) < 0) {
              return 0;
@@ -1564,8 +1568,9 @@ out:
  }
 
 
-/*  This state is the entry point for periodic schedule processing as
- *  well as being a continuation state for async processing.
+/*
+ * This state is the entry point for periodic schedule processing as
+ * well as being a continuation state for async processing.
   */
  static int ehci_state_fetchentry(EHCIState *ehci, int async)
  {
@@ -1674,7 +1679,7 @@ static EHCIQueue *ehci_state_fetchqh(EHCIState *ehci, int async)
 
  #if EHCI_DEBUG
      if (q->qhaddr != q->qh.next) {
-    DPRINTF("FETCHQH:  QH 0x%08x (h %x halt %x active %x) next 0x%08x\n",
+        DPRINTF("FETCHQH:  QH 0x%08x (h %x halt %x active %x) next 0x%08x\n",
                 q->qhaddr,
                 q->qh.epchar & QH_EPCHAR_H,
                 q->qh.token & QTD_TOKEN_HALT,
@@ -1924,8 +1929,10 @@ static int ehci_state_execute(EHCIQueue *q)
          return -1;
      }
 
-    // TODO verify enough time remains in the uframe as in 4.4.1.1
-    // TODO write back ptr to async list when done or out of time
+    /*
+     * TODO verify enough time remains in the uframe as in 4.4.1.1
+     * TODO write back ptr to async list when done or out of time
+     */
 
      /* 4.10.3, bottom of page 82, go horizontal on transaction counter == 0 */
      if (!q->async && q->transact_ctr == 0) {
@@ -2036,7 +2043,7 @@ static void ehci_advance_state(EHCIState *ehci, int async)
      int again;
 
      do {
-        switch(ehci_get_state(ehci, async)) {
+        switch (ehci_get_state(ehci, async)) {
          case EST_WAITLISTHEAD:
              again = ehci_state_waitlisthead(ehci, async);
              break;
@@ -2115,21 +2122,20 @@ static void ehci_advance_state(EHCIState *ehci, int async)
              ehci_reset(ehci);
              again = 0;
          }
-    }
-    while (again);
+    } while (again);
  }
 
  static void ehci_advance_async_state(EHCIState *ehci)
  {
      const int async = 1;
 
-    switch(ehci_get_state(ehci, async)) {
+    switch (ehci_get_state(ehci, async)) {
      case EST_INACTIVE:
          if (!ehci_async_enabled(ehci)) {
              break;
          }
          ehci_set_state(ehci, async, EST_ACTIVE);
-        // No break, fall through to ACTIVE
+        /* No break, fall through to ACTIVE */
 
      case EST_ACTIVE:
          if (!ehci_async_enabled(ehci)) {
@@ -2153,7 +2159,8 @@ static void ehci_advance_async_state(EHCIState *ehci)
          ehci_set_state(ehci, async, EST_WAITLISTHEAD);
          ehci_advance_state(ehci, async);
 
-        /* If the doorbell is set, the guest wants to make a change to the
+        /*
+         * If the doorbell is set, the guest wants to make a change to the
           * schedule. The host controller needs to release cached data.
           * (section 4.8.2)
           */
@@ -2180,13 +2187,13 @@ static void ehci_advance_periodic_state(EHCIState *ehci)
      uint32_t list;
      const int async = 0;
 
-    // 4.6
+    /* 4.6 */
 
-    switch(ehci_get_state(ehci, async)) {
+    switch (ehci_get_state(ehci, async)) {
      case EST_INACTIVE:
          if (!(ehci->frindex & 7) && ehci_periodic_enabled(ehci)) {
              ehci_set_state(ehci, async, EST_ACTIVE);
-            // No break, fall through to ACTIVE
+            /* No break, fall through to ACTIVE */
          } else
              break;
 
@@ -2210,7 +2217,7 @@ static void ehci_advance_periodic_state(EHCIState *ehci)
 
          DPRINTF("PERIODIC state adv fr=%d.  [%08X] -> %08X\n",
                  ehci->frindex / 8, list, entry);
-        ehci_set_fetch_addr(ehci, async,entry);
+        ehci_set_fetch_addr(ehci, async, entry);
          ehci_set_state(ehci, async, EST_FETCHENTRY);
          ehci_advance_state(ehci, async);
          ehci_queues_rip_unused(ehci, async);
@@ -2235,7 +2242,8 @@ static void ehci_update_frindex(EHCIState *ehci, int uframes)
          ehci_raise_irq(ehci, USBSTS_FLR);
      }
 
-    /* How many times will frindex roll over 0x4000 with this frame count?
+    /*
+     * How many times will frindex roll over 0x4000 with this frame count?
       * usbsts_frindex is decremented by 0x4000 on rollover until it reaches 0
       */
      int rollovers = (ehci->frindex + uframes) / 0x4000;
@@ -2315,8 +2323,9 @@ static void ehci_work_bh(void *opaque)
          ehci->async_stepdown++;
      }
 
-    /*  Async is not inside loop since it executes everything it can once
-     *  called
+    /*
+     * Async is not inside loop since it executes everything it can once
+     * called
       */
      if (ehci_async_enabled(ehci) || ehci->astate != EST_INACTIVE) {
          need_timer++;
@@ -2334,15 +2343,18 @@ static void ehci_work_bh(void *opaque)
      }
 
      if (need_timer) {
-        /* If we've raised int, we speed up the timer, so that we quickly
-         * notice any new packets queued up in response */
+        /*
+         * If we've raised int, we speed up the timer, so that we quickly
+         * notice any new packets queued up in response
+         */
          if (ehci->int_req_by_async && (ehci->usbsts & USBSTS_INT)) {
              expire_time = t_now +
                  NANOSECONDS_PER_SECOND / (FRAME_TIMER_FREQ * 4);

hw/usb/hcd-ehci.c static void ehci_work_bh(void *opaque) · line 2360

Quoted 10 lines
hw/usb/hcd-ehci.c +3 -2
              ehci->int_req_by_async = false;
          } else {
-            expire_time = t_now + (NANOSECONDS_PER_SECOND
-                               * (ehci->async_stepdown+1) / FRAME_TIMER_FREQ);
+            expire_time = t_now
+                + (NANOSECONDS_PER_SECOND * (ehci->async_stepdown + 1) /
+                   FRAME_TIMER_FREQ);
          }
          timer_mod(ehci->frame_timer, expire_time);
      }

Thanks,

C.

#33 lore raw

Re: [PATCH v3 08/17] hw/usb/hcd-ehci: Reject CTRLDSSEGMENT writes without 64-bit capability

In reply to Jamin Lin
Quoted 22 lines · Jamin Lin

The EHCI CTRLDSSEGMENT register provides the upper 32 bits [63:32] used to form 64-bit addresses for EHCI control data structures. Per EHCI 1.0 spec section 2.3.5, when the HCCPARAMS 64-bit Addressing Capability bit is zero, CTRLDSSEGMENT is not used: software cannot write it and reads must return zero.

Add a capability check in the operational register write handler and reject guest writes to CTRLDSSEGMENT when 64-bit addressing is not enabled.

---
  hw/usb/hcd-ehci.c | 9 ++++++++-
  1 file changed, 8 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.c +0 -0
index 6409457186..b9e4251d62 100644
@@ -1108,7 +1108,14 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,
                "      is enabled and HC is enabled\n");
          }

hw/usb/hcd-ehci.c static void ehci_opreg_write(void *ptr, hwaddr addr, · line 1118

Quoted 10 lines
hw/usb/hcd-ehci.c +8 -1
          break;
-
+    case CTRLDSSEGMENT:
+        if (!s->caps_64bit_addr) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "ehci: write to CTRLDSSEGMENT while "
+                          "      64-bit addressing capability is disabled\n");
+            return;
+        }
+        break;

I'd squash with previous.

#68 lore raw
In reply to Jamin Lin

Hi Jamin,

Quoted 14 lines · Jamin Lin

v1
1. Fix checkpatch coding style issues
2. Implement 64-bit addressing for QH/qTD/iTD/siTD descriptors
3. Add descriptor address offset property
4. Enable 64-bit EHCI DMA addressing on AST2700
5. Configure descriptor address offset for AST2700

v2
1. Remove unused EHCIfstn structure and dead code
2. Replace fprintf(stderr, ...) with qemu_log_mask(LOG_GUEST_ERROR)
3. Replace DPRINTF debug logs with trace events
4. Add functional tests for USB EHCI on AST2600 and AST2700 A1/A2
5. Fix review issue

Quoted 3 lines

v3:
1. Add Migration version test function
2. Add EHCI 64-bit buffer pointer fields description in commit log

I left trivial comments on v3 and expect v4 to be the last one :)

Regards,

Phil.

#51 lore raw

Re: [PATCH v3 13/17] hw/usb/hcd-ehci: Add descriptor address offset property

In reply to Jamin Lin
Quoted 75 lines · Jamin Lin

When 64-bit addressing is supported, the Linux EHCI driver programs the segment register to zero. See ehci_run function:
https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci-hcd.c

The driver comment also notes that descriptor structures allocated from the DMA pool use segment zero semantics.

Descriptor memory is allocated using the DMA API. The platform driver configures a 64-bit DMA mask so memory can be allocated above 4GB. See ehci_platform_probe function:
https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci-platform.c

On AST2700 platforms, system DRAM is mapped above 4GB at 0x400000000. As a result, descriptor addresses constructed directly from the guest EHCI registers do not match the actual system address used by the controller when fetching queue heads (QH) and queue element transfer descriptors (qTD).

Introduce a descriptor-addr-offset property so platforms can provide an address offset applied when constructing descriptor addresses. This allows systems where DRAM resides above 4GB to correctly access EHCI descriptors while keeping the default behavior unchanged for existing machines.

---
  hw/usb/hcd-ehci.h        | 1 +
  hw/usb/hcd-ehci-pci.c    | 2 ++
  hw/usb/hcd-ehci-sysbus.c | 2 ++
  hw/usb/hcd-ehci.c        | 7 ++++++-
  4 files changed, 11 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.h +1 -0
index 6406f536e8..05d8db4a5d 100644
@@ -264,6 +264,7 @@ struct EHCIState {
      /* properties */
      uint32_t maxframes;
      bool caps_64bit_addr;
+    uint64_t descriptor_addr_offset;
 
      /*
       *  EHCI spec version 1.0 Section 2.3
hw/usb/hcd-ehci-pci.c +2 -0
index 2ea8549db9..115d05ede0 100644
@@ -139,6 +139,8 @@ static const Property ehci_pci_properties[] = {
      DEFINE_PROP_UINT32("maxframes", EHCIPCIState, ehci.maxframes, 128),
      DEFINE_PROP_BOOL("caps-64bit-addr", EHCIPCIState, ehci.caps_64bit_addr,
                       false),
+    DEFINE_PROP_UINT64("descriptor-addr-offset", EHCIPCIState,
+                       ehci.descriptor_addr_offset, 0),
  };
 
  static const VMStateDescription vmstate_ehci_pci = {
hw/usb/hcd-ehci-sysbus.c +2 -0
index 61215e9f3d..df138fb339 100644
@@ -36,6 +36,8 @@ static const Property ehci_sysbus_properties[] = {
                       false),
      DEFINE_PROP_BOOL("caps-64bit-addr", EHCISysBusState, ehci.caps_64bit_addr,
                       false),
+    DEFINE_PROP_UINT64("descriptor-addr-offset", EHCISysBusState,
+                       ehci.descriptor_addr_offset, 0),
  };
 
  static void usb_ehci_sysbus_realize(DeviceState *dev, Error **errp)
hw/usb/hcd-ehci.c +0 -0
index e1d5ecbad7..ea9d5846bc 100644

hw/usb/hcd-ehci.c static uint64_t ehci_get_buf_addr(EHCIState *s, uint32_t hi, uint32_t low, · line 168

Quoted 8 lines
hw/usb/hcd-ehci.c +4 -1
@@ -162,7 +162,12 @@ static uint64_t ehci_get_buf_addr(EHCIState *s, uint32_t hi, uint32_t low,
 
  static uint64_t ehci_get_desc_addr(EHCIState *s, uint32_t low)
  {
-    return ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX);
+    uint64_t addr;
+
+    addr = ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX);
+    addr += s->descriptor_addr_offset;

Instead, can we force a default value for ctrldssegment when caps_64bit_addr is set ?

I am still trying to digest :

https://lore.kernel.org/qemu-devel/TYPPR06MB8206C7461E004C4FF07B8E92FC242@TYPPR06MB8206.apcprd06.prod.outlook.com/

We have time let's take a moment to think about the workaround.

Thanks,

C.

Quoted 4 lines
hw/usb/hcd-ehci.c +1 -0
+    return addr;
  }
 
  static void ehci_trace_usbsts(uint32_t mask, int state)
#20 lore raw

Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

In reply to Jamin Lin

+ Peter, Fabiano,

Quoted 32 lines · Jamin Lin

Change internal EHCI descriptor addresses from uint32_t to uint64_t.

The following fields are updated:
- EHCIPacket::qtdaddr
- EHCIQueue::{qhaddr, qtdaddr}
- EHCIState::{a_fetch_addr, p_fetch_addr}

Update get_dwords() and put_dwords() to take 64-bit addresses and propagate the type change through the descriptor traversal paths.

Adjust NLPTR_GET() to operate on 64-bit values:

     #define NLPTR_GET(x) ((x) & ~0x1fULL)

so that link pointer masking works correctly when descriptor addresses exceed 32-bit space. The previous mask (0xffffffe0) implicitly truncated addresses to 32 bits.

This patch does not change the on-wire descriptor layout yet. It only removes the internal 32-bit address limit and prepares for later patches that will add full 64-bit QH/qTD/iTD/siTD support.

Update the EHCI trace-events prototypes for QH, qTD, iTD, and siTD to use uint64_t for the address argument and print it with PRIx64. This ensures full 64-bit addresses are shown in trace output and improves debugging of queue heads and transfer descriptors.

Migration compatibility:

The fetch address fields in EHCIState are extended from 32-bit to 64-bit, requiring a VMState version bump (v2 -> v3).

Quoted 6 lines

Backward compatibility is preserved by keeping the legacy 32-bit fields (a_fetch_addr_pre_v3, p_fetch_addr_pre_v3) for pre-v3 migration streams via VMSTATE_UINT32_TEST(), and introducing new 64-bit fields for v3+.

In post_load, pre-v3 state is promoted to the 64-bit fields.

Since ehci is supported by downstream, I'd prefer that the migration maintainers also verify the implementation.

Thanks,

C.

Quoted 315 lines

No functional change.

---
  hw/usb/hcd-ehci.h   | 12 +++++----
  hw/usb/hcd-ehci.c   | 64 +++++++++++++++++++++++++++++----------------
  hw/usb/trace-events | 24 ++++++++---------
  3 files changed, 60 insertions(+), 40 deletions(-)
hw/usb/hcd-ehci.h +7 -5
index 4234591cb4..24eea9d23b 100644
@@ -208,7 +208,7 @@ struct EHCIPacket {
      QTAILQ_ENTRY(EHCIPacket) next;
 
      EHCIqtd qtd;           /* copy of current QTD (being worked on) */
-    uint32_t qtdaddr;      /* address QTD read from                 */
+    uint64_t qtdaddr;      /* address QTD read from                 */
 
      USBPacket packet;
      QEMUSGList sgl;
@@ -229,8 +229,8 @@ struct EHCIQueue {
       * when guest removes an entry (doorbell, handshake sequence)
       */
      EHCIqh qh;             /* copy of current QH (being worked on) */
-    uint32_t qhaddr;       /* address QH read from                 */
-    uint32_t qtdaddr;      /* address QTD read from                */
+    uint64_t qhaddr;       /* address QH read from                 */
+    uint64_t qtdaddr;      /* address QTD read from                */
      int last_pid;          /* pid of last packet executed          */
      USBDevice *dev;
      QTAILQ_HEAD(, EHCIPacket) packets;
@@ -294,8 +294,10 @@ struct EHCIState {
      EHCIQueueHead pqueues;
 
      /* which address to look at next */
-    uint32_t a_fetch_addr;
-    uint32_t p_fetch_addr;
+    uint32_t a_fetch_addr_pre_v3;
+    uint32_t p_fetch_addr_pre_v3;
+    uint64_t a_fetch_addr;
+    uint64_t p_fetch_addr;
 
      USBPacket ipacket;
      QEMUSGList isgl;
hw/usb/hcd-ehci.c +41 -23
index d0ea4949f9..35173d27b4 100644
@@ -72,7 +72,7 @@ typedef enum {
  } EHCI_STATES;
 
  /* macros for accessing fields within next link pointer entry */
-#define NLPTR_GET(x)             ((x) & 0xffffffe0)
+#define NLPTR_GET(x)             ((x) & ~0x1fULL)
  #define NLPTR_TYPE_GET(x)        (((x) >> 1) & 3)
  #define NLPTR_TBIT(x)            ((x) & 1)  /* 1=invalid, 0=valid */
 
@@ -287,7 +287,7 @@ static int ehci_get_state(EHCIState *s, int async)
      return async ? s->astate : s->pstate;
  }
 
-static void ehci_set_fetch_addr(EHCIState *s, int async, uint32_t addr)
+static void ehci_set_fetch_addr(EHCIState *s, int async, uint64_t addr)
  {
      if (async) {
          s->a_fetch_addr = addr;
@@ -296,7 +296,7 @@ static void ehci_set_fetch_addr(EHCIState *s, int async, uint32_t addr)
      }
  }
 
-static int ehci_get_fetch_addr(EHCIState *s, int async)
+static uint64_t ehci_get_fetch_addr(EHCIState *s, int async)
  {
      return async ? s->a_fetch_addr : s->p_fetch_addr;
  }
@@ -373,7 +373,7 @@ static inline bool ehci_periodic_enabled(EHCIState *s)
  }
 
  /* Get an array of dwords from main memory */
-static inline int get_dwords(EHCIState *ehci, uint32_t addr,
+static inline int get_dwords(EHCIState *ehci, uint64_t addr,
                               uint32_t *buf, int num)
  {
      int i;
@@ -395,7 +395,7 @@ static inline int get_dwords(EHCIState *ehci, uint32_t addr,
  }
 
  /* Put an array of dwords in to main memory */
-static inline int put_dwords(EHCIState *ehci, uint32_t addr,
+static inline int put_dwords(EHCIState *ehci, uint64_t addr,
                               uint32_t *buf, int num)
  {
      int i;
@@ -549,7 +549,7 @@ static void ehci_free_packet(EHCIPacket *p)
 
  /* queue management */
 
-static EHCIQueue *ehci_alloc_queue(EHCIState *ehci, uint32_t addr, int async)
+static EHCIQueue *ehci_alloc_queue(EHCIState *ehci, uint64_t addr, int async)
  {
      EHCIQueueHead *head = async ? &ehci->aqueues : &ehci->pqueues;
      EHCIQueue *q;
@@ -622,7 +622,7 @@ static void ehci_free_queue(EHCIQueue *q, const char *warn)
      g_free(q);
  }
 
-static EHCIQueue *ehci_find_queue_by_qh(EHCIState *ehci, uint32_t addr,
+static EHCIQueue *ehci_find_queue_by_qh(EHCIState *ehci, uint64_t addr,
                                          int async)
  {
      EHCIQueueHead *head = async ? &ehci->aqueues : &ehci->pqueues;
@@ -1135,7 +1135,7 @@ static void ehci_flush_qh(EHCIQueue *q)
  {
      uint32_t *qh = (uint32_t *) &q->qh;
      uint32_t dwords = sizeof(EHCIqh) >> 2;
-    uint32_t addr = NLPTR_GET(q->qhaddr);
+    uint64_t addr = NLPTR_GET(q->qhaddr);
 
      put_dwords(q->ehci, addr + 3 * sizeof(uint32_t), qh + 3, dwords - 3);
  }
@@ -1406,12 +1406,13 @@ static int ehci_execute(EHCIPacket *p, const char *action)
  /* 4.7.2 */
  static int ehci_process_itd(EHCIState *ehci,
                              EHCIitd *itd,
-                            uint32_t addr)
+                            uint64_t addr)
  {
      USBDevice *dev;
      USBEndpoint *ep;
      uint32_t i, len, pid, dir, devaddr, endp;
-    uint32_t pg, off, ptr1, ptr2, max, mult;
+    uint32_t pg, off, max, mult;
+    uint64_t ptr1, ptr2;
 
      ehci->periodic_sched_active = PERIODIC_ACTIVE;
 
@@ -1528,7 +1529,7 @@ static int ehci_state_waitlisthead(EHCIState *ehci,  int async)
      EHCIqh qh;
      int i = 0;
      int again = 0;
-    uint32_t entry = ehci->asynclistaddr;
+    uint64_t entry = ehci->asynclistaddr;
 
      /* set reclamation flag at start event (4.8.6) */
      if (async) {
@@ -1578,7 +1579,7 @@ out:
  static int ehci_state_fetchentry(EHCIState *ehci, int async)
  {
      int again = 0;
-    uint32_t entry = ehci_get_fetch_addr(ehci, async);
+    uint64_t entry = ehci_get_fetch_addr(ehci, async);
 
      if (NLPTR_TBIT(entry)) {
          ehci_set_state(ehci, async, EST_ACTIVE);
@@ -1610,8 +1611,10 @@ static int ehci_state_fetchentry(EHCIState *ehci, int async)
 
      default:
          /* TODO: handle FSTN type */
-        qemu_log_mask(LOG_GUEST_ERROR, "FETCHENTRY: entry at %X is of type %u "
-                "which is not supported yet\n", entry, NLPTR_TYPE_GET(entry));
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "FETCHENTRY: entry at %" PRIx64 "is of type %" PRIu64
+                      "which is not supported yet\n",
+                      entry, NLPTR_TYPE_GET(entry));
          return -1;
      }
 
@@ -1621,7 +1624,7 @@ out:
 
  static EHCIQueue *ehci_state_fetchqh(EHCIState *ehci, int async)
  {
-    uint32_t entry;
+    uint64_t entry;
      EHCIQueue *q;
      EHCIqh qh;
 
@@ -1710,7 +1713,7 @@ out:
 
  static int ehci_state_fetchitd(EHCIState *ehci, int async)
  {
-    uint32_t entry;
+    uint64_t entry;
      EHCIitd itd;
 
      assert(!async);
@@ -1736,7 +1739,7 @@ static int ehci_state_fetchitd(EHCIState *ehci, int async)
 
  static int ehci_state_fetchsitd(EHCIState *ehci, int async)
  {
-    uint32_t entry;
+    uint64_t entry;
      EHCIsitd sitd;
 
      assert(!async);
@@ -1794,7 +1797,7 @@ static int ehci_state_fetchqtd(EHCIQueue *q)
      EHCIqtd qtd;
      EHCIPacket *p;
      int again = 1;
-    uint32_t addr;
+    uint64_t addr;
 
      addr = NLPTR_GET(q->qtdaddr);
      if (get_dwords(q->ehci, addr +  8, &qtd.token,   1) < 0) {
@@ -1877,7 +1880,7 @@ static int ehci_fill_queue(EHCIPacket *p)
      USBEndpoint *ep = p->packet.ep;
      EHCIQueue *q = p->queue;
      EHCIqtd qtd = p->qtd;
-    uint32_t qtdaddr;
+    uint64_t qtdaddr;
 
      for (;;) {
          if (NLPTR_TBIT(qtd.next) != 0) {
@@ -2000,7 +2003,8 @@ static int ehci_state_executing(EHCIQueue *q)
  static int ehci_state_writeback(EHCIQueue *q)
  {
      EHCIPacket *p = QTAILQ_FIRST(&q->packets);
-    uint32_t *qtd, addr;
+    uint32_t *qtd;
+    uint64_t addr;
      int again = 0;
 
      /*  Write back the QTD from the QH area */
@@ -2441,6 +2445,11 @@ static int usb_ehci_post_load(void *opaque, int version_id)
          }
      }
 
+    if (version_id < 3) {
+        s->a_fetch_addr = s->a_fetch_addr_pre_v3;
+        s->p_fetch_addr = s->p_fetch_addr_pre_v3;
+    }
+
      return 0;
  }
 
@@ -2470,9 +2479,14 @@ static void usb_ehci_vm_state_change(void *opaque, bool running, RunState state)
      }
  }
 
+static bool ehci_core_is_before_version_3(void *opaque, int version_id)
+{
+    return version_id < 3;
+}
+
  const VMStateDescription vmstate_ehci = {
      .name        = "ehci-core",
-    .version_id  = 2,
+    .version_id  = 3,
      .minimum_version_id  = 1,
      .pre_save    = usb_ehci_pre_save,
      .post_load   = usb_ehci_post_load,
@@ -2501,8 +2515,12 @@ const VMStateDescription vmstate_ehci = {
          /* schedule state */
          VMSTATE_UINT32(astate, EHCIState),
          VMSTATE_UINT32(pstate, EHCIState),
-        VMSTATE_UINT32(a_fetch_addr, EHCIState),
-        VMSTATE_UINT32(p_fetch_addr, EHCIState),
+        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),
          VMSTATE_END_OF_LIST()
      }
  };
hw/usb/trace-events +12 -12
index 0d4318dcf1..8c90688bb3 100644
@@ -86,15 +86,15 @@ usb_ehci_portsc_write(uint32_t addr, uint32_t port, uint32_t val) "wr mmio 0x%04
  usb_ehci_portsc_change(uint32_t addr, uint32_t port, uint32_t new, uint32_t old) "ch mmio 0x%04x [port %d] = 0x%x (old: 0x%x)"
  usb_ehci_usbsts(const char *sts, int state) "usbsts %s %d"
  usb_ehci_state(const char *schedule, const char *state) "%s schedule %s"
-usb_ehci_qh_ptrs(void *q, uint32_t addr, uint32_t nxt, uint32_t c_qtd, uint32_t n_qtd, uint32_t a_qtd) "q %p - QH @ 0x%08x: next 0x%08x qtds 0x%08x,0x%08x,0x%08x"
-usb_ehci_qh_fields(uint32_t addr, int rl, int mplen, int eps, int ep, int devaddr) "QH @ 0x%08x - rl %d, mplen %d, eps %d, ep %d, dev %d"
-usb_ehci_qh_bits(uint32_t addr, int c, int h, int dtc, int i) "QH @ 0x%08x - c %d, h %d, dtc %d, i %d"
+usb_ehci_qh_ptrs(void *q, uint64_t addr, uint32_t nxt, uint32_t c_qtd, uint32_t n_qtd, uint32_t a_qtd) "q %p - QH @ 0x%" PRIx64 ": next 0x%08x qtds 0x%08x,0x%08x,0x%08x"
+usb_ehci_qh_fields(uint64_t addr, int rl, int mplen, int eps, int ep, int devaddr) "QH @ 0x%" PRIx64 " - rl %d, mplen %d, eps %d, ep %d, dev %d"
+usb_ehci_qh_bits(uint64_t addr, int c, int h, int dtc, int i) "QH @ 0x%" PRIx64 " - c %d, h %d, dtc %d, i %d"
  usb_ehci_qh_tbytes(uint32_t tbytes) "updating tbytes to %d"
-usb_ehci_qtd_ptrs(void *q, uint32_t addr, uint32_t nxt, uint32_t altnext) "q %p - QTD @ 0x%08x: next 0x%08x altnext 0x%08x"
-usb_ehci_qtd_fields(uint32_t addr, int tbytes, int cpage, int cerr, int pid) "QTD @ 0x%08x - tbytes %d, cpage %d, cerr %d, pid %d"
-usb_ehci_qtd_bits(uint32_t addr, int ioc, int active, int halt, int babble, int xacterr) "QTD @ 0x%08x - ioc %d, active %d, halt %d, babble %d, xacterr %d"
-usb_ehci_itd(uint32_t addr, uint32_t nxt, uint32_t mplen, uint32_t mult, uint32_t ep, uint32_t devaddr) "ITD @ 0x%08x: next 0x%08x - mplen %d, mult %d, ep %d, dev %d"
-usb_ehci_sitd(uint32_t addr, uint32_t nxt, uint32_t active) "ITD @ 0x%08x: next 0x%08x - active %d"
+usb_ehci_qtd_ptrs(void *q, uint64_t addr, uint32_t nxt, uint32_t altnext) "q %p - QTD @ 0x%" PRIx64 ": next 0x%08x altnext 0x%08x"
+usb_ehci_qtd_fields(uint64_t addr, int tbytes, int cpage, int cerr, int pid) "QTD @ 0x%" PRIx64 " - tbytes %d, cpage %d, cerr %d, pid %d"
+usb_ehci_qtd_bits(uint64_t addr, int ioc, int active, int halt, int babble, int xacterr) "QTD @ 0x%" PRIx64 " - ioc %d, active %d, halt %d, babble %d, xacterr %d"
+usb_ehci_itd(uint64_t addr, uint32_t nxt, uint32_t mplen, uint32_t mult, uint32_t ep, uint32_t devaddr) "ITD @ 0x%" PRIx64 ": next 0x%08x - mplen %d, mult %d, ep %d, dev %d"
+usb_ehci_sitd(uint64_t addr, uint32_t nxt, uint32_t active) "SITD @ 0x%" PRIx64 ": next 0x%08x - active %d"
  usb_ehci_port_attach(uint32_t port, const char *owner, const char *device) "attach port #%d, owner %s, device %s"
  usb_ehci_port_detach(uint32_t port, const char *owner) "detach port #%d, owner %s"
  usb_ehci_port_reset(uint32_t port, int enable) "reset port #%d - %d"
@@ -104,15 +104,15 @@ usb_ehci_port_resume(uint32_t port) "port #%d"
  usb_ehci_port_disable(uint32_t port) "port #%d"
  usb_ehci_queue_action(void *q, const char *action) "q %p: %s"
  usb_ehci_packet_action(void *q, void *p, const char *action) "q %p p %p: %s"
-usb_ehci_packet_submit(uint32_t qhaddr, uint32_t next, uint32_t qtdaddr, int pid, size_t len, int endp, int status, int actual_length) "qh=0x%x, next=0x%x, qtd=0x%x, pid=0x%x, len=%zd, endp=0x%x, status=%d, actual_length=%d"
+usb_ehci_packet_submit(uint64_t qhaddr, uint32_t next, uint64_t qtdaddr, int pid, size_t len, int endp, int status, int actual_length) "qh=0x%" PRIx64 ", next=0x%x, qtd=0x%" PRIx64 ", pid=0x%x, len=%zd, endp=0x%x, status=%d, actual_length=%d"
  usb_ehci_irq(uint32_t level, uint32_t frindex, uint32_t sts, uint32_t mask) "level %d, frindex 0x%04x, sts 0x%x, mask 0x%x"
  usb_ehci_guest_bug(const char *reason) "%s"
  usb_ehci_doorbell_ring(void) ""
  usb_ehci_doorbell_ack(void) ""
  usb_ehci_dma_error(void) ""
-usb_ehci_execute_complete(uint32_t qhaddr, uint32_t next, uint32_t qtdaddr, int status, int actual_length) "qhaddr=0x%x, next=0x%x, qtdaddr=0x%x, status=%d, actual_length=%d"
-usb_ehci_fetchqh_reclaim_done(uint32_t qhaddr) "QH 0x%08x H-bit set, reclamation status reset - done processing"
-usb_ehci_fetchqh_dbg(uint32_t qhaddr, uint32_t h, uint32_t halt, uint32_t active, uint32_t next) "QH 0x%08x (h 0x%x halt 0x%x active 0x%x) next 0x%08x"
+usb_ehci_execute_complete(uint64_t qhaddr, uint32_t next, uint64_t qtdaddr, int status, int actual_length) "qhaddr=0x%" PRIx64 ", next=0x%x, qtdaddr=0x%" PRIx64 ", status=%d, actual_length=%d"
+usb_ehci_fetchqh_reclaim_done(uint64_t qhaddr) "QH 0x%" PRIx64 " H-bit set, reclamation status reset - done processing"
+usb_ehci_fetchqh_dbg(uint64_t qhaddr, uint32_t h, uint32_t halt, uint32_t active, uint32_t next) "QH 0x%" PRIx64 " (h 0x%x halt 0x%x active 0x%x) next 0x%08x"
  usb_ehci_periodic_state_advance(uint32_t frame, uint32_t list, uint32_t entry) "frame=%d, list=0x%x, entry=0x%x"
  usb_ehci_skipped_uframes(uint64_t skipped_uframes) "skipped %" PRIu64 " uframes"
  usb_ehci_log(const char *msg) "%s"
author
#39 lore raw

RE: [PATCH v3 09/17] hw/usb/hcd-ehci: Implement 64-bit QH descriptor addressing

In reply to Philippe Mathieu-Daudé

Hi Philippe,

Quoted 87 lines

Subject: Re: [PATCH v3 09/17] hw/usb/hcd-ehci: Implement 64-bit QH descriptor addressing

Quoted 62 lines · Jamin Lin

Hi Philippe,

Quoted 60 lines

Subject: Re: [PATCH v3 09/17] hw/usb/hcd-ehci: Implement 64-bit QH descriptor addressing

Quoted 32 lines · Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register supplies the upper 32 bits which are concatenated with 32-bit link pointer fields to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

The current implementation assumes 32-bit QH descriptor addresses and directly uses link pointer values without applying the CTRLDSSEGMENT upper dword.

Introduce a helper, ehci_get_desc_addr(), to construct full 64-bit descriptor addresses when 64-bit capability is enabled. Update QH traversal paths (async list walk, horizontal QH link, and periodic schedule entry handling) to use the translated 64-bit addresses.

EHCI 64-bit buffer pointer fields are defined in Appendix B as split 32-bit low/high parts located at separate offsets, rather than a single contiguous 64-bit field. Therefore, the buffer pointers cannot be represented as uint64_t bufptr[5] without violating the descriptor layout defined by the specification.

Introduce ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] fields. Use this helper when calculating transfer buffer addresses so that data buffers above 4GB are correctly handled.

Also add bufptr_hi[5] to EHCIqh to support 64-bit buffer pointer fields as defined in Appendix B.

When 64-bit capability is disabled, descriptor addresses remain 32-bit and existing behaviour is unchanged.

Quoted 8 lines

Note: Similar split 64-bit buffer pointer handling is required for qTD, iTD and siTD descriptors, which will be addressed in follow-up changes.

---
   hw/usb/hcd-ehci.h   |  4 ++++
   hw/usb/hcd-ehci.c   | 48

++++++++++++++++++++++++++++++++++-----------

hw/usb/hcd-ehci.c static const char *addr2str(hwaddr addr) · line 151

Quoted 6 lines
   hw/usb/trace-events |  2 +-
   3 files changed, 42 insertions(+), 12 deletions(-)
hw/usb/hcd-ehci.c +2 -0
+static uint64_t ehci_get_buf_addr(EHCIState *s, uint32_t hi, uint32_t low,
+                                  uint32_t mask)

Rename @mask as @low_mask? Or directly use 64-bit?

static uint64_t ehci_get_buf_addr(const EHCIState *s,

                                    uint32_t hi, uint32_t lo,
                                    uint64_t mask) {
      uint64_t addr = s->caps_64bit_addr ? deposit64(lo, 0, 32, hi):

lo;

Quoted 6 lines
Quoted 3 lines
      return addr & mask;
}

Thanks for the review and suggestion.

In the current code, the mask (e.g. QTD_BUFPTR_MASK) is defined as a

32-bit value and is intended to apply only to the low 32 bits of the buffer pointer.

Quoted 2 lines

If we change the function to take a 64-bit mask and apply it to the full 64-bit address (as in the proposed version), the upper 32 bits (bits 63–32)

may be unintentionally cleared when the mask is only 32-bit wide.

Right.

Quoted 6 lines

For example:

#define QTD_BUFPTR_MASK 0xfffff000

In this case, applying "addr & mask" after combining hi/lo would zero out the

high 32 bits, which breaks 64-bit addressing.

Quoted 9 lines
Quoted 1 line

To preserve the original semantics, the mask should only be applied to the

low part before combining with the high 32 bits.

Quoted 1 line

So I suggest renaming the parameter to make this explicit. I will change this

function as below. What do you think?

Quoted 2 lines

static uint64_t ehci_get_buf_addr(

"const"

Thanks for the suggestion.
Will add.
Jamin

Quoted 13 lines

EHCIState *s, uint32_t hi, uint32_t lo,

Quoted 9 lines
                                   uint32_t lo_mask) {
     uint64_t addr = lo & lo_mask;

     if (s->caps_64bit_addr) {
         addr = deposit64(addr, 32, 32, hi);
     }

     return addr;
}

Perfect, using "const":

#52 lore raw

Re: [PATCH v3 13/17] hw/usb/hcd-ehci: Add descriptor address offset property

In reply to Cédric Le Goater
Quoted 52 lines · Cédric Le Goater
Quoted 38 lines · Jamin Lin

When 64-bit addressing is supported, the Linux EHCI driver programs the segment register to zero. See ehci_run function:
https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci-hcd.c

The driver comment also notes that descriptor structures allocated from the DMA pool use segment zero semantics.

Descriptor memory is allocated using the DMA API. The platform driver configures a 64-bit DMA mask so memory can be allocated above 4GB. See ehci_platform_probe function:
https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci- platform.c

On AST2700 platforms, system DRAM is mapped above 4GB at 0x400000000. As a result, descriptor addresses constructed directly from the guest EHCI registers do not match the actual system address used by the controller when fetching queue heads (QH) and queue element transfer descriptors (qTD).

Introduce a descriptor-addr-offset property so platforms can provide an address offset applied when constructing descriptor addresses. This allows systems where DRAM resides above 4GB to correctly access EHCI descriptors while keeping the default behavior unchanged for existing machines.

---
  hw/usb/hcd-ehci.h        | 1 +
  hw/usb/hcd-ehci-pci.c    | 2 ++
  hw/usb/hcd-ehci-sysbus.c | 2 ++
  hw/usb/hcd-ehci.c        | 7 ++++++-

  4 files changed, 11 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.c +0 -0
index e1d5ecbad7..ea9d5846bc 100644

hw/usb/hcd-ehci.c static uint64_t ehci_get_buf_addr(EHCIState *s, · line 168

Quoted 8 lines
hw/usb/hcd-ehci.c +4 -1
@@ -162,7 +162,12 @@ static uint64_t ehci_get_buf_addr(EHCIState *s, 
 uint32_t hi, uint32_t low,
   static uint64_t ehci_get_desc_addr(EHCIState *s, uint32_t low)
   {
-    return ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX);
+    uint64_t addr;
+
+    addr = ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX);
+    addr += s->descriptor_addr_offset;

Instead, can we force a default value for ctrldssegment when caps_64bit_addr is set ?

Quoted 4 lines

I am still trying to digest :

  https://lore.kernel.org/qemu-devel/
TYPPR06MB8206C7461E004C4FF07B8E92FC242@TYPPR06MB8206.apcprd06.prod.outlook.com/

Same here, but since it doesn't break and I don't have much more time to dedicate to this I ended thinking "good enough for now, could be improved later", so no objection on my side.

Quoted 2 lines

We have time let's take a moment to think about the workaround.

Then keep me in the loop!

author
#63 lore raw

RE: [PATCH v3 16/17] tests/functional/arm/test_aspeed_ast2600_sdk: Add USB EHCI test for AST2600 SDK

In reply to Philippe Mathieu-Daudé

Hi Philippe

Quoted 11 lines

Subject: Re: [PATCH v3 16/17] tests/functional/arm/test_aspeed_ast2600_sdk:
Add USB EHCI test for AST2600 SDK

Quoted 7 lines · Jamin Lin

Add a functional test to verify USB EHCI support on the AST2600 SDK machine by attaching a USB keyboard device and checking its enumeration via lsusb.

This introduces a helper routine that runs lsusb in the guest and validates that the emulated "QEMU USB Keyboard" is detected.

Quoted 7 lines
Quoted 4 lines
---
  tests/functional/arm/test_aspeed_ast2600_sdk.py | 7 +++++++
  1 file changed, 7 insertions(+)

This should be patch #1 :)

This patch series updates the USB EHCI model to support 64-bit.

I added a functional test (AST2600) to ensure that these changes do not impact the existing 32-bit support. For this reason, I placed this patch in the current order.
If you still think the patch order should be adjusted, please let me know.

Thanks,
Jamin

Quoted 1 line
author
#46 lore raw

RE: [PATCH v3 12/17] hw/usb/hcd-ehci: Implement 64-bit siTD descriptor addressing

In reply to Philippe Mathieu-Daudé

Hi Philippe,

Quoted 37 lines

Subject: Re: [PATCH v3 12/17] hw/usb/hcd-ehci: Implement 64-bit siTD descriptor addressing

Quoted 33 lines · Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register provides the upper 32 bits that are concatenated with 32-bit link pointer values to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

siTD link pointers are stored as 32-bit values and must be expanded to full 64-bit descriptor addresses when 64-bit mode is enabled. Update the siTD traversal path to use ehci_get_desc_addr() when following link pointers.

Appendix B also defines high dword fields for siTD buffer pointers. Add bufptr_hi[] fields to EHCIsitd and use ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] when processing split isochronous transfers. This allows buffers above 4GB to be handled correctly.

When 64-bit capability is disabled, descriptor and buffer addresses remain 32-bit and existing behaviour is unchanged.

---
  hw/usb/hcd-ehci.h | 1 +
  hw/usb/hcd-ehci.c | 2 +-
  2 files changed, 2 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.h index renamed from hw/usb/hcd-ehci.h +0 -0

af8c080c60..6406f536e8 100644

hw/usb/hcd-ehci.h +0 -0
@@ -109,6 +109,7 @@ typedef struct EHCIsitd {
  #define SITD_BUFPTR_TCNT_MASK         0x00000007
 
Quoted 6 lines

hw/usb/hcd-ehci.h typedef struct EHCIsitd { · line 111

Quoted 1 line
hw/usb/hcd-ehci.h +0 -0
      uint32_t backptr;                 /* Standard next link pointer

*/

hw/usb/hcd-ehci.h typedef struct EHCIsitd { · line 112

Quoted 1 line
hw/usb/hcd-ehci.h +1 -0
+    uint32_t bufptr_hi[2];

Not used?

Thanks for the review.
Yes, bufptr_hi is not used for siTD. I will remove it.

Thanks,
Jamin

Quoted 18 lines

hw/usb/hcd-ehci.c static int ehci_state_fetchsitd(EHCIState *ehci, int

Quoted 8 lines
hw/usb/hcd-ehci.h +0 -0
  } EHCIsitd;
 
  /*
hw/usb/hcd-ehci.c index renamed from hw/usb/hcd-ehci.c +0 -0

3db1c25e4d..e1d5ecbad7 100644

hw/usb/hcd-ehci.c +0 -0
@@ -1791,7 +1791,7 @@ static int ehci_state_fetchsitd(EHCIState *ehci, int

async)

Quoted 9 lines
hw/usb/hcd-ehci.c +2 -1
          warn_report("Skipping active siTD");
      }
 
-    ehci_set_fetch_addr(ehci, async, sitd.next);
+    ehci_set_fetch_addr(ehci, async, ehci_get_desc_addr(ehci,
+ sitd.next));
      ehci_set_state(ehci, async, EST_FETCHENTRY);
      return 1;
  }
author
#34 lore raw

RE: [PATCH v3 08/17] hw/usb/hcd-ehci: Reject CTRLDSSEGMENT writes without 64-bit capability

In reply to Philippe Mathieu-Daudé

Hi Philippe,

Quoted 34 lines

Subject: Re: [PATCH v3 08/17] hw/usb/hcd-ehci: Reject CTRLDSSEGMENT writes without 64-bit capability

hw/usb/hcd-ehci.c static void ehci_opreg_write(void *ptr, hwaddr

Quoted 20 lines · Jamin Lin

The EHCI CTRLDSSEGMENT register provides the upper 32 bits [63:32] used to form 64-bit addresses for EHCI control data structures. Per EHCI 1.0 spec section 2.3.5, when the HCCPARAMS 64-bit Addressing Capability bit is zero, CTRLDSSEGMENT is not used: software cannot write it and reads must return zero.

Add a capability check in the operational register write handler and reject guest writes to CTRLDSSEGMENT when 64-bit addressing is not enabled.

---
  hw/usb/hcd-ehci.c | 9 ++++++++-
  1 file changed, 8 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.c index renamed from hw/usb/hcd-ehci.c +0 -0

6409457186..b9e4251d62 100644

hw/usb/hcd-ehci.c +0 -0
@@ -1108,7 +1108,14 @@ static void ehci_opreg_write(void *ptr, hwaddr

addr,

hw/usb/hcd-ehci.c static void ehci_opreg_write(void *ptr, hwaddr addr,

Quoted 9 lines
hw/usb/hcd-ehci.c +5 -1
                "      is enabled and HC is enabled\n");
          }
          break;
-
+    case CTRLDSSEGMENT:
+        if (!s->caps_64bit_addr) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "ehci: write to CTRLDSSEGMENT while "
+                          "      64-bit addressing capability is
Quoted 6 lines

disabled\n");

hw/usb/hcd-ehci.c static void ehci_opreg_write(void *ptr, hwaddr addr, · line 1118

Quoted 3 lines
hw/usb/hcd-ehci.c +3 -0
+            return;
+        }
+        break;

I'd squash with previous.

Thanks for the review and suggestion.
Will do.
Jamin

author
#15 lore raw

RE: [PATCH v3 04/17] hw/usb/hcd-ehci.c: Replace fprintf(stderr, ...) with qemu_log_mask(LOG_GUEST_ERROR)

In reply to Philippe Mathieu-Daudé

Hi Philippe,

Quoted 46 lines

Subject: Re: [PATCH v3 04/17] hw/usb/hcd-ehci.c: Replace fprintf(stderr, ...) with qemu_log_mask(LOG_GUEST_ERROR)

Quoted 21 lines · Jamin Lin

Replace direct fprintf(stderr, ...) calls with qemu_log_mask() using LOG_GUEST_ERROR in the EHCI controller implementation.

Direct writes to stderr are discouraged in QEMU as they bypass the logging framework and cannot be filtered or controlled at runtime. Using qemu_log_mask() ensures that guest-visible errors are properly categorized and can be managed via QEMU logging options.

All affected messages correspond to guest-triggerable error conditions, so LOG_GUEST_ERROR is used consistently.

This change improves integration with QEMU's logging infrastructure and aligns with upstream coding guidelines.

---
  hw/usb/hcd-ehci.c | 43 +++++++++++++++++++++++++------------------
  1 file changed, 25 insertions(+), 18 deletions(-)

hw/usb/hcd-ehci.c static void ehci_free_packet(EHCIPacket *p) · line 537

Quoted 6 lines
hw/usb/hcd-ehci.c +1 -1
@@ -532,7 +533,7 @@ static void ehci_free_packet(EHCIPacket *p)
      }
      if (p->async == EHCI_ASYNC_FINISHED &&
              p->packet.status == USB_RET_SUCCESS) {
-        fprintf(stderr,
+        qemu_log_mask(LOG_GUEST_ERROR,
                  "EHCI: Dropping completed packet from halted %s ep

%02X\n",

hw/usb/hcd-ehci.c static void ehci_free_packet(EHCIPacket *p) · line 539

Quoted 2 lines
hw/usb/hcd-ehci.c +0 -0
                  (p->pid == USB_TOKEN_IN) ? "in" : "out",
                  get_field(p->queue->qh.epchar, QH_EPCHAR_EP));

@@

hw/usb/hcd-ehci.c static void ehci_opreg_write(void *ptr, hwaddr addr,

Quoted 10 lines

-1042,8 +1043,9 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,

          /* not supporting dynamic frame list size at the moment */
          if ((val & USBCMD_FLS) && !(s->usbcmd & USBCMD_FLS)) {
hw/usb/hcd-ehci.c +3 -2
-            fprintf(stderr, "attempt to set frame list size -- value %d\n",
-                    (int)val & USBCMD_FLS);
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "attempt to set frame list size -- value
+ %d\n",
Quoted 1 line

PRId64

Will update

Quoted 19 lines
Quoted 19 lines
hw/usb/hcd-ehci.c +3 -1
+                          (int)val & USBCMD_FLS);
              val &= ~USBCMD_FLS;
          }
@@ -1584,7 +1589,8 @@ static int ehci_state_fetchentry(EHCIState
 *ehci, int async)
 
      /* section 4.8, only QH in async schedule */
      if (async && (NLPTR_TYPE_GET(entry) != NLPTR_TYPE_QH)) {
-        fprintf(stderr, "non queue head request in async schedule\n");
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "non queue head request in async schedule\n");
          return -1;
      }
@@ -1606,7 +1612,7 @@ static int ehci_state_fetchentry(EHCIState
 *ehci, int async)
 
Quoted 7 lines

hw/usb/hcd-ehci.c static int ehci_state_fetchentry(EHCIState *ehci, int async)

Quoted 4 lines
hw/usb/hcd-ehci.c +1 -1
      default:
          /* TODO: handle FSTN type */
-        fprintf(stderr, "FETCHENTRY: entry at %X is of type %u "
+        qemu_log_mask(LOG_GUEST_ERROR, "FETCHENTRY: entry at %X

is of type %u "

"0x%x"

Will update

Quoted 7 lines

Review static int ehci_state_fetchentry(EHCIState · line 1617

Quoted 2 lines
(quoted diff) +0 -0
                  "which is not supported yet\n", entry,

NLPTR_TYPE_GET(entry));

Review static void ehci_advance_state(EHCIState *ehci,

Quoted 3 lines
(quoted diff) +0 -0
          return -1;
      }
(quoted diff) +0 -0
@@ -2112,13 +2118,14 @@ static void ehci_advance_state(EHCIState *ehci,

int async)

Quoted 8 lines

hw/usb/hcd-ehci.c static void ehci_advance_state(EHCIState *ehci, int async) · line 2121

Quoted 5 lines
hw/usb/hcd-ehci.c +1 -1
              break;
 
          default:
-            fprintf(stderr, "Bad state!\n");
+            qemu_log_mask(LOG_GUEST_ERROR, "Bad state!\n");

Just remove for "bad state" entries?

Will remove it.

Quoted 14 lines
Quoted 14 lines
hw/usb/hcd-ehci.c +2 -1
              g_assert_not_reached();
          }
 
          if (again < 0 || itd_count > 16) {
              /* TODO: notify guest (raise HSE irq?) */
-            fprintf(stderr, "processing error - resetting ehci HC\n");
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "processing error - resetting ehci HC\n");
              ehci_reset(ehci);
              again = 0;
          }
@@ -2175,7 +2182,7 @@ static void ehci_advance_async_state(EHCIState
 *ehci)
 
Quoted 9 lines

hw/usb/hcd-ehci.c static void ehci_advance_async_state(EHCIState *ehci)

Quoted 4 lines
hw/usb/hcd-ehci.c +1 -1
      default:
          /* this should only be due to a developer mistake */
-        fprintf(stderr, "ehci: Bad asynchronous state %d. "
+        qemu_log_mask(LOG_GUEST_ERROR, "ehci: Bad asynchronous

state %d. "

hw/usb/hcd-ehci.c static void ehci_advance_async_state(EHCIState *ehci) · line 2186

Quoted 1 line
hw/usb/hcd-ehci.c +0 -0
                  "Resetting to active\n", ehci->astate);

Ditto.

Will remove it.

Quoted 10 lines

hw/usb/hcd-ehci.c static void · line 2236

Quoted 9 lines
(quoted diff) +0 -0
          g_assert_not_reached();
      }
hw/usb/hcd-ehci.c +1 -1
@@ -2225,7 +2232,7 @@ static void
 ehci_advance_periodic_state(EHCIState *ehci)
 
      default:
          /* this should only be due to a developer mistake */
-        fprintf(stderr, "ehci: Bad periodic state %d. "
+        qemu_log_mask(LOG_GUEST_ERROR, "ehci: Bad periodic state

%d. "

Quoted 5 lines

hw/usb/hcd-ehci.c static void ehci_advance_periodic_state(EHCIState *ehci) · line 2237

Quoted 2 lines
hw/usb/hcd-ehci.c +0 -0
                  "Resetting to active\n", ehci->pstate);
          g_assert_not_reached();

Ditto.

Will remove it.

Thanks for the review and suggestion
Jamin

Quoted 1 line
Quoted 1 line
      }
#47 lore raw

Re: [PATCH v3 12/17] hw/usb/hcd-ehci: Implement 64-bit siTD descriptor addressing

In reply to Jamin Lin
Quoted 41 lines · Jamin Lin

Hi Philippe,

Quoted 39 lines

Subject: Re: [PATCH v3 12/17] hw/usb/hcd-ehci: Implement 64-bit siTD descriptor addressing

Quoted 31 lines · Jamin Lin

EHCI supports 64-bit control data structure addressing when the 64-bit Addressing Capability bit in HCCPARAMS is set. In that mode, the CTRLDSSEGMENT register provides the upper 32 bits that are concatenated with 32-bit link pointer values to form full 64-bit descriptor addresses (EHCI 1.0, section 2.3.5 and Appendix B).

siTD link pointers are stored as 32-bit values and must be expanded to full 64-bit descriptor addresses when 64-bit mode is enabled. Update the siTD traversal path to use ehci_get_desc_addr() when following link pointers.

Appendix B also defines high dword fields for siTD buffer pointers. Add bufptr_hi[] fields to EHCIsitd and use ehci_get_buf_addr() to construct full 64-bit buffer addresses from bufptr[] and bufptr_hi[] when processing split isochronous transfers. This allows buffers above 4GB to be handled correctly.

When 64-bit capability is disabled, descriptor and buffer addresses remain 32-bit and existing behaviour is unchanged.

---
   hw/usb/hcd-ehci.h | 1 +
   hw/usb/hcd-ehci.c | 2 +-
   2 files changed, 2 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.h index renamed from hw/usb/hcd-ehci.h +0 -0

af8c080c60..6406f536e8 100644

hw/usb/hcd-ehci.h typedef struct EHCIsitd { · line 111

Quoted 3 lines
hw/usb/hcd-ehci.h +0 -0
@@ -109,6 +109,7 @@ typedef struct EHCIsitd {
   #define SITD_BUFPTR_TCNT_MASK         0x00000007
 
       uint32_t backptr;                 /* Standard next link pointer

*/

Quoted 7 lines
Quoted 4 lines

hw/usb/hcd-ehci.h typedef struct EHCIsitd { · line 112

Quoted 1 line
hw/usb/hcd-ehci.h +1 -0
+    uint32_t bufptr_hi[2];

Not used?

Thanks for the review.
Yes, bufptr_hi is not used for siTD. I will remove it.

With it removed:

#64 lore raw

Re: [PATCH v3 16/17] tests/functional/arm/test_aspeed_ast2600_sdk: Add USB EHCI test for AST2600 SDK

In reply to Jamin Lin
Quoted 18 lines · Jamin Lin

Hi Philippe

Quoted 16 lines

Subject: Re: [PATCH v3 16/17] tests/functional/arm/test_aspeed_ast2600_sdk:
Add USB EHCI test for AST2600 SDK

Quoted 11 lines · Jamin Lin

Add a functional test to verify USB EHCI support on the AST2600 SDK machine by attaching a USB keyboard device and checking its enumeration via lsusb.

This introduces a helper routine that runs lsusb in the guest and validates that the emulated "QEMU USB Keyboard" is detected.

---
   tests/functional/arm/test_aspeed_ast2600_sdk.py | 7 +++++++
   1 file changed, 7 insertions(+)
Quoted 6 lines
Quoted 2 lines

This should be patch #1 :)

This patch series updates the USB EHCI model to support 64-bit.

I added a functional test (AST2600) to ensure that these changes do not impact the existing 32-bit support.

This is also my understanding, first add the AST2600 test to ensure we do not break anything while implementing 64-bit desc addressing.

Quoted 7 lines

For this reason, I placed this patch in the current order.
If you still think the patch order should be adjusted, please let me know.

Thanks,
Jamin

Quoted 1 line
author
#65 lore raw

RE: [PATCH v3 16/17] tests/functional/arm/test_aspeed_ast2600_sdk: Add USB EHCI test for AST2600 SDK

Hi Philippe

Quoted 26 lines

Subject: Re: [PATCH v3 16/17] tests/functional/arm/test_aspeed_ast2600_sdk:
Add USB EHCI test for AST2600 SDK

Quoted 3 lines · Jamin Lin

Hi Philippe

Quoted 1 line

Subject: Re: [PATCH v3 16/17]

tests/functional/arm/test_aspeed_ast2600_sdk:

Quoted 18 lines
Quoted 17 lines

Add USB EHCI test for AST2600 SDK

Quoted 11 lines · Jamin Lin

Add a functional test to verify USB EHCI support on the AST2600 SDK machine by attaching a USB keyboard device and checking its enumeration via lsusb.

This introduces a helper routine that runs lsusb in the guest and validates that the emulated "QEMU USB Keyboard" is detected.

---
   tests/functional/arm/test_aspeed_ast2600_sdk.py | 7 +++++++
   1 file changed, 7 insertions(+)

This should be patch #1 :)

Quoted 8 lines
Quoted 3 lines

This patch series updates the USB EHCI model to support 64-bit.

I added a functional test (AST2600) to ensure that these changes do not

impact the existing 32-bit support.

This is also my understanding, first add the AST2600 test to ensure we do not break anything while implementing 64-bit desc addressing.

Thanks for the suggestion.
I will move this patch before the 64-bit descriptor addressing changes.

Quoted 7 lines
Quoted 7 lines

For this reason, I placed this patch in the current order.
If you still think the patch order should be adjusted, please let me know.

Thanks,
Jamin

Quoted 1 line
#5 lore raw

RE: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

In reply to Jamin Lin
Quoted 94 lines · Jamin Lin

Hi all,

Quoted 64 lines

Subject: RE: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

Hi all,

Quoted 28 lines

Subject: Re: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

Quoted 24 lines · Jamin Lin

The EHCIfstn structure, defined according to EHCI spec section 3.7, is currently unused in the EHCI implementation. Remove it to reduce unused code and improve maintainability.

Additionally, drop a block of disabled (#if 0) code in ehci_state_advqueue() that was never implemented. This code path is incomplete and has no functional impact.

No functional change.

---
hw/usb/hcd-ehci.h | 7 -------
hw/usb/hcd-ehci.c | 11 -----------

2 files changed, 18 deletions(-)
hw/usb/hcd-ehci.h index renamed from hw/usb/hcd-ehci.h +0 -0

0ae8c06331..3aec0e047c 100644

hw/usb/hcd-ehci.h +0 -0
@@ -202,13 +202,6 @@ typedef struct EHCIqh {
 #define BUFPTR_SBYTES_SH              5
 } EHCIqh;
 
Quoted 4 lines

hw/usb/hcd-ehci.h typedef struct EHCIqh { · line 208

Quoted 4 lines
hw/usb/hcd-ehci.h +0 -4
-/*  EHCI spec version 1.0 Section 3.7
- */
-typedef struct EHCIfstn {
-    uint32_t next;                    /* Standard next link pointer

*/

Quoted 22 lines
Quoted 14 lines
hw/usb/hcd-ehci.h +0 -3
-    uint32_t backptr;                 /* Standard next link pointer */
-} EHCIfstn;
-
 enum async_state {
    EHCI_ASYNC_NONE = 0,
    EHCI_ASYNC_INITIALIZED,
hw/usb/hcd-ehci.c index renamed from hw/usb/hcd-ehci.c +0 -0

57f930b099..5ea8461f70 100644

hw/usb/hcd-ehci.c +0 -0
@@ -1756,17 +1756,6 @@ static int ehci_state_fetchsitd(EHCIState
 *ehci, int async)
 /* Section 4.10.2 - paragraph 3 */

static int ehci_state_advqueue(EHCIQueue *q) { -#if 0

hw/usb/hcd-ehci.c

Quoted 4 lines
hw/usb/hcd-ehci.c +0 -4
-    /* TO-DO: 4.10.2 - paragraph 2
-     * if I-bit is set to 1 and QH is not active
-     * go to horizontal QH
-     */

Is it worth to keep the comment in case it's still valid to remind we have something missing here?

Quoted 3 lines

Regards,
BALATON Zoltan

Quoted 6 lines

If we decide to keep these comments as reminders for unimplemented parts, please let me know. I will resend the patch series and drop this patch.

Thanks,
Jamin

Another approach:
We can remove the #if 0 dead code and keep only the comments as a reminder that this part is not yet implemented.

Ex:

hw/usb/hcd-ehci.c +0 -5
@@ -1756,16 +1756,10 @@ static int ehci_state_fetchsitd(EHCIState *ehci, int async)
 /* Section 4.10.2 - paragraph 3 */
 static int ehci_state_advqueue(EHCIQueue *q)
 {
-#if 0
    /* TO-DO: 4.10.2 - paragraph 2
     * if I-bit is set to 1 and QH is not active
     * go to horizontal QH
     */
-    if (I-bit set) {
-        ehci_set_state(ehci, async, EST_HORIZONTALQH);
-        goto out;
-    }
 
 Thanks,
Quoted 1 line

Jamin

That's what I meant to leave the comment but OK the remove ineffective code.

Regards,
BALATON Zoltan

#21 lore raw

Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

In reply to Cédric Le Goater
Quoted 40 lines · Cédric Le Goater

+ Peter, Fabiano,

Quoted 37 lines · Jamin Lin

Change internal EHCI descriptor addresses from uint32_t to uint64_t.

The following fields are updated:
- EHCIPacket::qtdaddr
- EHCIQueue::{qhaddr, qtdaddr}
- EHCIState::{a_fetch_addr, p_fetch_addr}

Update get_dwords() and put_dwords() to take 64-bit addresses and propagate the type change through the descriptor traversal paths.

Adjust NLPTR_GET() to operate on 64-bit values:

     #define NLPTR_GET(x) ((x) & ~0x1fULL)

so that link pointer masking works correctly when descriptor addresses exceed 32-bit space. The previous mask (0xffffffe0) implicitly truncated addresses to 32 bits.

This patch does not change the on-wire descriptor layout yet. It only removes the internal 32-bit address limit and prepares for later patches that will add full 64-bit QH/qTD/iTD/siTD support.

Update the EHCI trace-events prototypes for QH, qTD, iTD, and siTD to use uint64_t for the address argument and print it with PRIx64. This ensures full 64-bit addresses are shown in trace output and improves debugging of queue heads and transfer descriptors.

Migration compatibility:

The fetch address fields in EHCIState are extended from 32-bit to 64-bit, requiring a VMState version bump (v2 -> v3).

Backward compatibility is preserved by keeping the legacy 32-bit fields (a_fetch_addr_pre_v3, p_fetch_addr_pre_v3) for pre-v3 migration streams via VMSTATE_UINT32_TEST(), and introducing new 64-bit fields for v3+.

Quoted 4 lines
Quoted 1 line

In post_load, pre-v3 state is promoted to the 64-bit fields.

Since ehci is supported by downstream, I'd prefer that the migration maintainers also verify the implementation.

Thanks, I'll only look at the migration bits.

[...]

Quoted 29 lines
Quoted 29 lines
hw/usb/hcd-ehci.c +11 -1
@@ -2441,6 +2445,11 @@ static int usb_ehci_post_load(void *opaque, int version_id)
          }
      }
+    if (version_id < 3) {
+        s->a_fetch_addr = s->a_fetch_addr_pre_v3;
+        s->p_fetch_addr = s->p_fetch_addr_pre_v3;
+    }
+
      return 0;
  }
@@ -2470,9 +2479,14 @@ static void usb_ehci_vm_state_change(void *opaque, bool running, RunState state)
      }
  }
+static bool ehci_core_is_before_version_3(void *opaque, int version_id)
+{
+    return version_id < 3;
+}
+
  const VMStateDescription vmstate_ehci = {
      .name        = "ehci-core",
-    .version_id  = 2,
+    .version_id  = 3,
      .minimum_version_id  = 1,
      .pre_save    = usb_ehci_pre_save,
      .post_load   = usb_ehci_post_load,
@@ -2501,8 +2515,12 @@ const VMStateDescription vmstate_ehci = {
          /* schedule state */
          VMSTATE_UINT32(astate, EHCIState),
          VMSTATE_UINT32(pstate, EHCIState),
Quoted 8 lines

hw/usb/hcd-ehci.c const VMStateDescription vmstate_ehci = { · line 2523

Quoted 8 lines
hw/usb/hcd-ehci.c +6 -2
-        VMSTATE_UINT32(a_fetch_addr, EHCIState),
-        VMSTATE_UINT32(p_fetch_addr, EHCIState),
+        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),

TL;DR: I feel like we still need machine type compat properties.

Details:

When a v2 stream arrives, two _TEST()s will do the loading, then post_load() extend it to 64bits, looks fine.

When a v3 stream arrives, two _TEST()s got skipped then latter two take effect. post_load() skips. Looks fine.

When migrating to another QEMU, due to the fact saving vmstates always take vmsd's version declared (3), I don't see how it can migrate back to a v2 stream; it didn't know about v3.

Jiamin, have you tested migrating from a new QEMU binary back to another old one? For upstream and serious devices, we need to guarantee bi-directional migrations, back and forth.

Thanks,

Quoted 3 lines
Quoted 3 lines
hw/usb/hcd-ehci.c +0 -0
          VMSTATE_END_OF_LIST()
      }
  };
Signature · 1 line
Peter Xu
author
#6 lore raw

RE: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

In reply to BALATON Zoltan

Hi BALATON

Quoted 99 lines

Subject: RE: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

Quoted 41 lines · Jamin Lin

Hi all,

Quoted 39 lines

Subject: RE: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

Hi all,

Quoted 28 lines

Subject: Re: [PATCH v3 01/17] hw/usb/hcd-ehci: Remove unused EHCIfstn structure and dead code

Quoted 24 lines · Jamin Lin

The EHCIfstn structure, defined according to EHCI spec section 3.7, is currently unused in the EHCI implementation. Remove it to reduce unused code and improve maintainability.

Additionally, drop a block of disabled (#if 0) code in ehci_state_advqueue() that was never implemented. This code path is incomplete and has no functional impact.

No functional change.

---
hw/usb/hcd-ehci.h | 7 -------
hw/usb/hcd-ehci.c | 11 -----------

2 files changed, 18 deletions(-)
hw/usb/hcd-ehci.h index renamed from hw/usb/hcd-ehci.h +0 -0

0ae8c06331..3aec0e047c 100644

hw/usb/hcd-ehci.h +0 -0
@@ -202,13 +202,6 @@ typedef struct EHCIqh {
 #define BUFPTR_SBYTES_SH              5
 } EHCIqh;
 
Quoted 4 lines

hw/usb/hcd-ehci.h typedef struct EHCIqh { · line 208

Quoted 4 lines
hw/usb/hcd-ehci.h +0 -4
-/*  EHCI spec version 1.0 Section 3.7
- */
-typedef struct EHCIfstn {
-    uint32_t next;                    /* Standard next link pointer

*/

Quoted 1 line
Quoted 1 line
hw/usb/hcd-ehci.h +0 -1
-    uint32_t backptr;                 /* Standard next link pointer

*/

Quoted 28 lines
Quoted 28 lines
Quoted 21 lines
Quoted 13 lines
hw/usb/hcd-ehci.h +0 -2
-} EHCIfstn;
-
 enum async_state {
    EHCI_ASYNC_NONE = 0,
    EHCI_ASYNC_INITIALIZED,
hw/usb/hcd-ehci.c index renamed from hw/usb/hcd-ehci.c +0 -0

57f930b099..5ea8461f70 100644

hw/usb/hcd-ehci.c +0 -0
@@ -1756,17 +1756,6 @@ static int ehci_state_fetchsitd(EHCIState
 *ehci, int async)
 /* Section 4.10.2 - paragraph 3 */

static int ehci_state_advqueue(EHCIQueue *q) { -#if 0

hw/usb/hcd-ehci.c

Quoted 4 lines
hw/usb/hcd-ehci.c +0 -4
-    /* TO-DO: 4.10.2 - paragraph 2
-     * if I-bit is set to 1 and QH is not active
-     * go to horizontal QH
-     */

Is it worth to keep the comment in case it's still valid to remind we have something missing here?

Quoted 3 lines

Regards,
BALATON Zoltan

If we decide to keep these comments as reminders for unimplemented parts, please let me know. I will resend the patch series and drop this patch.

Quoted 6 lines
Quoted 3 lines

Thanks,
Jamin

Another approach:
We can remove the #if 0 dead code and keep only the comments as a

reminder that this part is not yet implemented.

Quoted 17 lines

Ex:

hw/usb/hcd-ehci.c +0 -0
@@ -1756,16 +1756,10 @@ static int ehci_state_fetchsitd(EHCIState
 *ehci, int async)
 /* Section 4.10.2 - paragraph 3 */

static int ehci_state_advqueue(EHCIQueue *q) { -#if 0
/* TO-DO: 4.10.2 - paragraph 2
* if I-bit is set to 1 and QH is not active
* go to horizontal QH
*/

hw/usb/hcd-ehci.c +0 -4
-    if (I-bit set) {
-        ehci_set_state(ehci, async, EST_HORIZONTALQH);
-        goto out;
-    }
Quoted 4 lines
Quoted 2 lines

Thanks,
Jamin

That's what I meant to leave the comment but OK the remove ineffective code.

Thanks for the suggestion.
Will do.
Jamin

Quoted 3 lines

Regards,
BALATON Zoltan

author
#53 lore raw

RE: [PATCH v3 13/17] hw/usb/hcd-ehci: Add descriptor address offset property

In reply to Philippe Mathieu-Daudé

Hi Philippe, Cédric

Quoted 18 lines
Quoted 7 lines
Quoted 5 lines

-    return ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX); +    uint64_t addr;
+
+    addr = ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX); +    addr += s->descriptor_addr_offset;

Quoted 7 lines

Instead, can we force a default value for ctrldssegment when caps_64bit_addr is set ?

I am still trying to digest :

  https://lore.kernel.org/qemu-devel/

[email protected]. prod.out

Quoted 1 line

look.com/

Quoted 8 lines

Same here, but since it doesn't break and I don't have much more time to dedicate to this I ended thinking "good enough for now, could be improved later", so no objection on my side.

Quoted 2 lines

We have time let's take a moment to think about the workaround.

Then keep me in the loop!

Thanks for the review and the suggestion.

I tried to address this by introducing a new property:
aspeed-ast2700-workaround.

When this property is enabled:

1. We force bit 2 of the CTRLDSSEGMENT register to 1 during initialization (when caps_64bit_addr is set).
2. Additionally, if firmware writes to the CTRLDSSEGMENT register, we also ensure that bit 2 remains set.

This keeps the behavior contained to AST2700-specific usage and avoids impacting other platforms. Let me know if you think this approach makes sense, or if you'd prefer a different direction.

Jamin

hw/usb/hcd-ehci.h +1 -0
index af8c080c60..0a58dca668 100644
@@ -263,6 +263,7 @@ struct EHCIState {
     /* properties */
     uint32_t maxframes;
     bool caps_64bit_addr;
+    bool aspeed_ast2700_workaround;
 
     /*
      *  EHCI spec version 1.0 Section 2.3
hw/arm/aspeed_ast27x0.c +3 -0
index 4a1f7cad73..83114bae90 100644
@@ -858,6 +858,9 @@ static void aspeed_soc_ast2700_realize(DeviceState *dev, Error **errp)
     for (i = 0; i < sc->ehcis_num; i++) {
         object_property_set_bool(OBJECT(&s->ehci[i]), "caps-64bit-addr", true,
                                  &error_abort);
+        object_property_set_bool(OBJECT(&s->ehci[i]),
+                                 "aspeed-ast2700-workaround", true,
+                                 &error_abort);
         if (!sysbus_realize(SYS_BUS_DEVICE(&s->ehci[i]), errp)) {
             return;
         }
hw/usb/hcd-ehci-sysbus.c +3 -0
index 61215e9f3d..01f72a5a1e 100644
@@ -36,6 +36,9 @@ static const Property ehci_sysbus_properties[] = {
                      false),
     DEFINE_PROP_BOOL("caps-64bit-addr", EHCISysBusState, ehci.caps_64bit_addr,
                      false),
+    DEFINE_PROP_BOOL("aspeed-ast2700-workaround", EHCISysBusState,
+                     ehci.aspeed_ast2700_workaround,
+                     false),
 };
 
 static void usb_ehci_sysbus_realize(DeviceState *dev, Error **errp)
hw/usb/hcd-ehci.c +6 -0
index 7f90fd119d..403d7e84a7 100644
@@ -1133,6 +1133,9 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,
                           "      64-bit addressing capability is disabled\n");
             return;
         }
+        if (s->aspeed_ast2700_workaround) {
+            val |= BIT(2);
+        }
         break;
     case ASYNCLISTADDR:
         if (ehci_async_enabled(s)) {
@@ -2582,6 +2585,9 @@ void usb_ehci_realize(EHCIState *s, DeviceState *dev, Error **errp)
     }
     if (s->caps_64bit_addr) {
         s->caps[0x08] |= BIT(0);
+        if (s->aspeed_ast2700_workaround) {
+            s->ctrldssegment |= BIT(2);
+        }
     }
author
#22 lore raw

RE: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

In reply to Peter Xu

Hi Peter, Philippe, Cédric

Quoted 86 lines

Subject: Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Quoted 40 lines · Cédric Le Goater

+ Peter, Fabiano,

Quoted 37 lines · Jamin Lin

Change internal EHCI descriptor addresses from uint32_t to uint64_t.

The following fields are updated:
- EHCIPacket::qtdaddr
- EHCIQueue::{qhaddr, qtdaddr}
- EHCIState::{a_fetch_addr, p_fetch_addr}

Update get_dwords() and put_dwords() to take 64-bit addresses and propagate the type change through the descriptor traversal paths.

Adjust NLPTR_GET() to operate on 64-bit values:

     #define NLPTR_GET(x) ((x) & ~0x1fULL)

so that link pointer masking works correctly when descriptor addresses exceed 32-bit space. The previous mask (0xffffffe0) implicitly truncated addresses to 32 bits.

This patch does not change the on-wire descriptor layout yet. It only removes the internal 32-bit address limit and prepares for later patches that will add full 64-bit QH/qTD/iTD/siTD support.

Update the EHCI trace-events prototypes for QH, qTD, iTD, and siTD to use uint64_t for the address argument and print it with PRIx64. This ensures full 64-bit addresses are shown in trace output and improves debugging of queue heads and transfer descriptors.

Migration compatibility:

The fetch address fields in EHCIState are extended from 32-bit to 64-bit, requiring a VMState version bump (v2 -> v3).

Backward compatibility is preserved by keeping the legacy 32-bit fields (a_fetch_addr_pre_v3, p_fetch_addr_pre_v3) for pre-v3 migration streams via VMSTATE_UINT32_TEST(), and introducing new 64-bit fields for v3+.

Quoted 4 lines
Quoted 1 line

In post_load, pre-v3 state is promoted to the 64-bit fields.

Since ehci is supported by downstream, I'd prefer that the migration maintainers also verify the implementation.

Thanks, I'll only look at the migration bits.

[...]

Quoted 1 line

Review static int usb_ehci_post_load(void *opaque, int

Quoted 1 line
(quoted diff) +0 -0
@@ -2441,6 +2445,11 @@ static int usb_ehci_post_load(void *opaque, int

version_id)

Quoted 10 lines

hw/usb/hcd-ehci.c static void usb_ehci_vm_state_change(void

Quoted 10 lines
hw/usb/hcd-ehci.c +5 -0
          }
      }
+    if (version_id < 3) {
+        s->a_fetch_addr = s->a_fetch_addr_pre_v3;
+        s->p_fetch_addr = s->p_fetch_addr_pre_v3;
+    }
+
      return 0;
  }
@@ -2470,9 +2479,14 @@ static void usb_ehci_vm_state_change(void

*opaque, bool running, RunState state)

Quoted 20 lines
Quoted 20 lines
hw/usb/hcd-ehci.c +6 -3
      }
  }
+static bool ehci_core_is_before_version_3(void *opaque, int
+version_id) {
+    return version_id < 3;
+}
+
  const VMStateDescription vmstate_ehci = {
      .name        = "ehci-core",
-    .version_id  = 2,
+    .version_id  = 3,
      .minimum_version_id  = 1,
      .pre_save    = usb_ehci_pre_save,
      .post_load   = usb_ehci_post_load,
@@ -2501,8 +2515,12 @@ const VMStateDescription vmstate_ehci = {
          /* schedule state */
          VMSTATE_UINT32(astate, EHCIState),
          VMSTATE_UINT32(pstate, EHCIState),
-        VMSTATE_UINT32(a_fetch_addr, EHCIState),
-        VMSTATE_UINT32(p_fetch_addr, EHCIState),
Quoted 9 lines
Quoted 6 lines

hw/usb/hcd-ehci.c const VMStateDescription vmstate_ehci = { · line 2523

Quoted 6 lines
hw/usb/hcd-ehci.c +6 -0
+        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),

TL;DR: I feel like we still need machine type compat properties.

I don't understand this: Could you please describe it in more detail?

Quoted 8 lines

Details:

When a v2 stream arrives, two _TEST()s will do the loading, then post_load() extend it to 64bits, looks fine.

When a v3 stream arrives, two _TEST()s got skipped then latter two take effect. post_load() skips. Looks fine.

Quoted 8 lines

When migrating to another QEMU, due to the fact saving vmstates always take vmsd's version declared (3), I don't see how it can migrate back to a v2 stream; it didn't know about v3.

Jiamin, have you tested migrating from a new QEMU binary back to another old one? For upstream and serious devices, we need to guarantee bi-directional migrations, back and forth.

Sorry, I am not familiar with the design of QEMU migration, and I don't know how to test it. I added these hook functions based on Philippe's suggestion and review.
Could you tell me how to test migration to ensure both version 3 and version 2 are workable?

Thanks for the help, review, and suggestions.

Quoted 8 lines

Thanks,

Quoted 3 lines

hw/usb/hcd-ehci.c const VMStateDescription vmstate_ehci = { · line 2526

Quoted 3 lines
hw/usb/hcd-ehci.c +0 -0
          VMSTATE_END_OF_LIST()
      }
  };

--
Peter Xu

#54 lore raw

Re: [PATCH v3 13/17] hw/usb/hcd-ehci: Add descriptor address offset property

In reply to Jamin Lin
Quoted 27 lines · Jamin Lin

Hi Philippe, Cédric

Quoted 25 lines
Quoted 7 lines
Quoted 5 lines

-    return ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX); +    uint64_t addr;
+
+    addr = ehci_get_buf_addr(s, s->ctrldssegment, low, UINT32_MAX); +    addr += s->descriptor_addr_offset;

Quoted 7 lines

Instead, can we force a default value for ctrldssegment when caps_64bit_addr is set ?

I am still trying to digest :

  https://lore.kernel.org/qemu-devel/

[email protected]. prod.out

Quoted 1 line

look.com/

Same here, but since it doesn't break and I don't have much more time to dedicate to this I ended thinking "good enough for now, could be improved later", so no objection on my side.

Quoted 2 lines

We have time let's take a moment to think about the workaround.

Quoted 7 lines
Quoted 1 line

Then keep me in the loop!

Thanks for the review and the suggestion.

I tried to address this by introducing a new property:
aspeed-ast2700-workaround.

This is not specific to AST2700.

I think a forced default value for CTRLDSSEGMENT would be more appropriate. Perhaps "ctrldssegment-default", as a uint32_t, which defaults to 0x0 and when the ast2700 SoC is realized, set to BIT(2).

Quoted 56 lines

When this property is enabled:

1. We force bit 2 of the CTRLDSSEGMENT register to 1 during initialization (when caps_64bit_addr is set).
2. Additionally, if firmware writes to the CTRLDSSEGMENT register, we also ensure that bit 2 remains set.

This keeps the behavior contained to AST2700-specific usage and avoids impacting other platforms. Let me know if you think this approach makes sense, or if you'd prefer a different direction.

Jamin

hw/usb/hcd-ehci.h +1 -0
index af8c080c60..0a58dca668 100644
@@ -263,6 +263,7 @@ struct EHCIState {
      /* properties */
      uint32_t maxframes;
      bool caps_64bit_addr;
+    bool aspeed_ast2700_workaround;
 
      /*
       *  EHCI spec version 1.0 Section 2.3
hw/arm/aspeed_ast27x0.c +3 -0
index 4a1f7cad73..83114bae90 100644
@@ -858,6 +858,9 @@ static void aspeed_soc_ast2700_realize(DeviceState *dev, Error **errp)
      for (i = 0; i < sc->ehcis_num; i++) {
          object_property_set_bool(OBJECT(&s->ehci[i]), "caps-64bit-addr", true,
                                   &error_abort);
+        object_property_set_bool(OBJECT(&s->ehci[i]),
+                                 "aspeed-ast2700-workaround", true,
+                                 &error_abort);
          if (!sysbus_realize(SYS_BUS_DEVICE(&s->ehci[i]), errp)) {
              return;
          }
hw/usb/hcd-ehci-sysbus.c +3 -0
index 61215e9f3d..01f72a5a1e 100644
@@ -36,6 +36,9 @@ static const Property ehci_sysbus_properties[] = {
                       false),
      DEFINE_PROP_BOOL("caps-64bit-addr", EHCISysBusState, ehci.caps_64bit_addr,
                       false),
+    DEFINE_PROP_BOOL("aspeed-ast2700-workaround", EHCISysBusState,
+                     ehci.aspeed_ast2700_workaround,
+                     false),
  };
 
  static void usb_ehci_sysbus_realize(DeviceState *dev, Error **errp)
hw/usb/hcd-ehci.c +0 -0
index 7f90fd119d..403d7e84a7 100644

hw/usb/hcd-ehci.c static void ehci_opreg_write(void *ptr, hwaddr addr, · line 1138

Quoted 6 lines
hw/usb/hcd-ehci.c +3 -0
@@ -1133,6 +1133,9 @@ static void ehci_opreg_write(void *ptr, hwaddr addr,
                            "      64-bit addressing capability is disabled\n");
              return;
          }
+        if (s->aspeed_ast2700_workaround) {
+            val |= BIT(2);
+        }

With the proposal above, this would be changed to :

	val |= s->ctrldssegment_default;

hw/usb/hcd-ehci.c void usb_ehci_realize(EHCIState *s, DeviceState *dev, Error **errp) · line 2589

Quoted 9 lines
hw/usb/hcd-ehci.c +2 -0
          break;
      case ASYNCLISTADDR:
          if (ehci_async_enabled(s)) {
@@ -2582,6 +2585,9 @@ void usb_ehci_realize(EHCIState *s, DeviceState *dev, Error **errp)
      }
      if (s->caps_64bit_addr) {
          s->caps[0x08] |= BIT(0);
+        if (s->aspeed_ast2700_workaround) {
+            s->ctrldssegment |= BIT(2);

Shouldn't these assignments be in the reset handlers ?

Thanks,

C.

Quoted 2 lines
hw/usb/hcd-ehci.c +1 -0
+        }
      }
#23 lore raw

Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

In reply to Jamin Lin
Quoted 1 line · Jamin Lin

Hi Peter, Philippe, Cédric

Hi, Jamin,

[...]

Quoted 16 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
hw/usb/hcd-ehci.c +0 -2
@@ -2501,8 +2515,12 @@ const VMStateDescription vmstate_ehci = {
          /* schedule state */
          VMSTATE_UINT32(astate, EHCIState),
          VMSTATE_UINT32(pstate, EHCIState),
-        VMSTATE_UINT32(a_fetch_addr, EHCIState),
-        VMSTATE_UINT32(p_fetch_addr, EHCIState),
Quoted 9 lines
Quoted 6 lines

hw/usb/hcd-ehci.c const VMStateDescription vmstate_ehci = { · line 2523

Quoted 6 lines
hw/usb/hcd-ehci.c +6 -0
+        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),

TL;DR: I feel like we still need machine type compat properties.

Quoted 1 line

I don't understand this: Could you please describe it in more detail?

See hw/core/machine.c and entries added into hw_compat_*[] arrays. IIUC, if this is an important device and if we want to guarantee bidirectional migration (I'll explain later), then we should stick with machine compat properties.

Quoted 13 lines
Quoted 13 lines

Details:

When a v2 stream arrives, two _TEST()s will do the loading, then post_load() extend it to 64bits, looks fine.

When a v3 stream arrives, two _TEST()s got skipped then latter two take effect. post_load() skips. Looks fine.

When migrating to another QEMU, due to the fact saving vmstates always take vmsd's version declared (3), I don't see how it can migrate back to a v2 stream; it didn't know about v3.

Quoted 8 lines
Quoted 4 lines

Jiamin, have you tested migrating from a new QEMU binary back to another old one? For upstream and serious devices, we need to guarantee bi-directional migrations, back and forth.

Sorry, I am not familiar with the design of QEMU migration, and I don't know how to test it. I added these hook functions based on Philippe's suggestion and review.
Could you tell me how to test migration to ensure both version 3 and version 2 are workable?

Since this patch already modified VMSDs, I believe migration test should have been carried out at the very minimum.. Logically if vmsd versioning is used, we should also have tested forward migrations, because that's only for that.

What I'm talking about is I think backward migration will fail.

Forward migration describes the case when a VM hosted by an old QEMU binary (e.g. QEMU v10.2.0) is migrated to a new QEMU (e.g., after this patch applied on top of master).

Backward migration describes the case when a VM hosted by a new QEMU binary (e.g., after this patch applied on top of master) is migrated back to an old QEMU binary (e.g. QEMU v10.2.0).

For different devices, we should allow different level of support on migration. Any device that may be important to either upstream QEMU or downstreams should better allow bi-directional migration. In this case we need to use machine compat properties I mentioned. I'm not yet sure where EHCI controller falls e.g. v.s. XHCI.. when not sure, we can also just stick with machine compat properties to be on the safe side.

For more info, you can also refer to:

https://www.qemu.org/docs/master/devel/migration/compatibility.html

Thanks,

Signature · 1 line
Peter Xu
#28 lore raw

Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

In reply to Jamin Lin
Quoted 112 lines · Jamin Lin

Hi Peter, Philippe, Cédric

Quoted 86 lines

Subject: Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Quoted 40 lines · Cédric Le Goater

+ Peter, Fabiano,

Quoted 37 lines · Jamin Lin

Change internal EHCI descriptor addresses from uint32_t to uint64_t.

The following fields are updated:
- EHCIPacket::qtdaddr
- EHCIQueue::{qhaddr, qtdaddr}
- EHCIState::{a_fetch_addr, p_fetch_addr}

Update get_dwords() and put_dwords() to take 64-bit addresses and propagate the type change through the descriptor traversal paths.

Adjust NLPTR_GET() to operate on 64-bit values:

      #define NLPTR_GET(x) ((x) & ~0x1fULL)

so that link pointer masking works correctly when descriptor addresses exceed 32-bit space. The previous mask (0xffffffe0) implicitly truncated addresses to 32 bits.

This patch does not change the on-wire descriptor layout yet. It only removes the internal 32-bit address limit and prepares for later patches that will add full 64-bit QH/qTD/iTD/siTD support.

Update the EHCI trace-events prototypes for QH, qTD, iTD, and siTD to use uint64_t for the address argument and print it with PRIx64. This ensures full 64-bit addresses are shown in trace output and improves debugging of queue heads and transfer descriptors.

Migration compatibility:

The fetch address fields in EHCIState are extended from 32-bit to 64-bit, requiring a VMState version bump (v2 -> v3).

Backward compatibility is preserved by keeping the legacy 32-bit fields (a_fetch_addr_pre_v3, p_fetch_addr_pre_v3) for pre-v3 migration streams via VMSTATE_UINT32_TEST(), and introducing new 64-bit fields for v3+.

Quoted 4 lines
Quoted 1 line

In post_load, pre-v3 state is promoted to the 64-bit fields.

Since ehci is supported by downstream, I'd prefer that the migration maintainers also verify the implementation.

Thanks, I'll only look at the migration bits.

[...]

Quoted 1 line

Review static int usb_ehci_post_load(void *opaque, int

Quoted 1 line
(quoted diff) +0 -0
@@ -2441,6 +2445,11 @@ static int usb_ehci_post_load(void *opaque, int

version_id)

Quoted 10 lines

hw/usb/hcd-ehci.c static void usb_ehci_vm_state_change(void

Quoted 10 lines
hw/usb/hcd-ehci.c +5 -0
           }
       }
+    if (version_id < 3) {
+        s->a_fetch_addr = s->a_fetch_addr_pre_v3;
+        s->p_fetch_addr = s->p_fetch_addr_pre_v3;
+    }
+
       return 0;
   }
@@ -2470,9 +2479,14 @@ static void usb_ehci_vm_state_change(void

*opaque, bool running, RunState state)

Quoted 20 lines
Quoted 20 lines
hw/usb/hcd-ehci.c +6 -3
       }
   }
+static bool ehci_core_is_before_version_3(void *opaque, int
+version_id) {
+    return version_id < 3;
+}
+
   const VMStateDescription vmstate_ehci = {
       .name        = "ehci-core",
-    .version_id  = 2,
+    .version_id  = 3,
       .minimum_version_id  = 1,
       .pre_save    = usb_ehci_pre_save,
       .post_load   = usb_ehci_post_load,
@@ -2501,8 +2515,12 @@ const VMStateDescription vmstate_ehci = {
           /* schedule state */
           VMSTATE_UINT32(astate, EHCIState),
           VMSTATE_UINT32(pstate, EHCIState),
-        VMSTATE_UINT32(a_fetch_addr, EHCIState),
-        VMSTATE_UINT32(p_fetch_addr, EHCIState),
Quoted 9 lines
Quoted 6 lines

hw/usb/hcd-ehci.c const VMStateDescription vmstate_ehci = { · line 2523

Quoted 6 lines
hw/usb/hcd-ehci.c +6 -0
+        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),

TL;DR: I feel like we still need machine type compat properties.

I don't understand this: Could you please describe it in more detail?

Quoted 12 lines

Details:

When a v2 stream arrives, two _TEST()s will do the loading, then post_load() extend it to 64bits, looks fine.

When a v3 stream arrives, two _TEST()s got skipped then latter two take effect. post_load() skips. Looks fine.

When migrating to another QEMU, due to the fact saving vmstates always take vmsd's version declared (3), I don't see how it can migrate back to a v2 stream; it didn't know about v3.

Quoted 8 lines
Quoted 4 lines

Jiamin, have you tested migrating from a new QEMU binary back to another old one? For upstream and serious devices, we need to guarantee bi-directional migrations, back and forth.

Sorry, I am not familiar with the design of QEMU migration, and I don't know how to test it. I added these hook functions based on Philippe's suggestion and review.
Could you tell me how to test migration to ensure both version 3 and version 2 are workable?

Migrating a running guest back and forth between QEMU [1] and [2] should be possible :

[1] path/to/qemu-10.2/install/bin/qemu-system-x86_64 -M pc-q35-10.2 \
    ... \
    -device usb-ehci \
    -device usb-kbd
[2] path/to/myqemu-with-patches/install/bin/qemu-system-x86_64 -M pc-q35-10.2 \
    ... \
    -device usb-ehci \
    -device usb-kbd

This is required to guarantee forward and backward compatibility of QEMU pc-q35-10.2 machine instances.

C.

author
#50 lore raw

RE: [PATCH v3 13/17] hw/usb/hcd-ehci: Add descriptor address offset property

In reply to Philippe Mathieu-Daudé

Hi Philippe,

Quoted 62 lines

Subject: Re: [PATCH v3 13/17] hw/usb/hcd-ehci: Add descriptor address offset property

Quoted 8 lines · Jamin Lin

When 64-bit addressing is supported, the Linux EHCI driver programs the segment register to zero. See ehci_run function:
https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci-hc
d.c

The driver comment also notes that descriptor structures allocated from the DMA pool use segment zero semantics.

Quoted 7 lines

Descriptor memory is allocated using the DMA API. The platform driver configures a 64-bit DMA mask so memory can be allocated above 4GB. See ehci_platform_probe function:
https://github.com/torvalds/linux/blob/master/drivers/usb/host/ehci-pl
atform.c

On AST2700 platforms, system DRAM is mapped above 4GB at

0x400000000.

Quoted 24 lines

As a result, descriptor addresses constructed directly from the guest EHCI registers do not match the actual system address used by the controller when fetching queue heads (QH) and queue element transfer descriptors (qTD).

Introduce a descriptor-addr-offset property so platforms can provide an address offset applied when constructing descriptor addresses. This allows systems where DRAM resides above 4GB to correctly access EHCI descriptors while keeping the default behavior unchanged for existing machines.

---
  hw/usb/hcd-ehci.h        | 1 +
  hw/usb/hcd-ehci-pci.c    | 2 ++
  hw/usb/hcd-ehci-sysbus.c | 2 ++
  hw/usb/hcd-ehci.c        | 7 ++++++-
  4 files changed, 11 insertions(+), 1 deletion(-)
hw/usb/hcd-ehci.h index renamed from hw/usb/hcd-ehci.h +0 -0

6406f536e8..05d8db4a5d 100644

hw/usb/hcd-ehci.h struct EHCIState { · line 267

Quoted 4 lines
hw/usb/hcd-ehci.h +1 -0
@@ -264,6 +264,7 @@ struct EHCIState {
      /* properties */
      uint32_t maxframes;
      bool caps_64bit_addr;
+    uint64_t descriptor_addr_offset;

I note OHCI uses "dma_addr_t dma_offset;" here.

hw/usb/hcd-ehci-pci.c static const Property ehci_pci_properties[] = { · line 139

Quoted 9 lines
hw/usb/hcd-ehci.h +0 -0
      /*
       *  EHCI spec version 1.0 Section 2.3 diff --git

a/hw/usb/hcd-ehci-pci.c b/hw/usb/hcd-ehci-pci.c index 2ea8549db9..115d05ede0 100644

hw/usb/hcd-ehci-pci.c +0 -0
@@ -139,6 +139,8 @@ static const Property ehci_pci_properties[] = {
      DEFINE_PROP_UINT32("maxframes", EHCIPCIState, ehci.maxframes,

128),

hw/usb/hcd-ehci-pci.c static const Property ehci_pci_properties[] = { · line 140

Quoted 1 line
hw/usb/hcd-ehci-pci.c +0 -0
      DEFINE_PROP_BOOL("caps-64bit-addr", EHCIPCIState,
Quoted 9 lines

ehci.caps_64bit_addr,

hw/usb/hcd-ehci-pci.c static const Property ehci_pci_properties[] = { · line 144

Quoted 4 lines
hw/usb/hcd-ehci-pci.c +2 -0
                       false),
+    DEFINE_PROP_UINT64("descriptor-addr-offset", EHCIPCIState,
+                       ehci.descriptor_addr_offset, 0),
  };

I note OHCI uses DEFINE_PROP_DMAADDR() here. Single use, maybe better remove it and stick to uint64_t as you use.

Thanks for the suggestion and review.

Based on Cédric’s feedback, I will update the ctrldssegment-default property to use a uint32_t type.

Quoted 4 lines

Could we have a DEFINE_EHCI_COMMON_PROPERTIES() macro for common entries instead, such DEFINE_SDHCI_COMMON_PROPERTIES()?

Will do

Thanks,
Jamin

Quoted 15 lines

hw/usb/hcd-ehci-sysbus.c static const Property ehci_sysbus_properties[] = { · line 37

Quoted 9 lines
hw/usb/hcd-ehci-pci.c +0 -0
  static const VMStateDescription vmstate_ehci_pci = { diff --git

a/hw/usb/hcd-ehci-sysbus.c b/hw/usb/hcd-ehci-sysbus.c index 61215e9f3d..df138fb339 100644

hw/usb/hcd-ehci-sysbus.c +0 -0
@@ -36,6 +36,8 @@ static const Property ehci_sysbus_properties[] = {
                       false),
      DEFINE_PROP_BOOL("caps-64bit-addr", EHCISysBusState,

ehci.caps_64bit_addr,

Quoted 4 lines
hw/usb/hcd-ehci-sysbus.c +2 -0
                       false),
+    DEFINE_PROP_UINT64("descriptor-addr-offset", EHCISysBusState,
+                       ehci.descriptor_addr_offset, 0),
  };
author
#60 lore raw

RE: [PATCH v3 15/17] hw/arm/aspeed_ast27x0: Set EHCI descriptor address offset

In reply to Philippe Mathieu-Daudé

Hi Philippe

Quoted 35 lines

Subject: Re: [PATCH v3 15/17] hw/arm/aspeed_ast27x0: Set EHCI descriptor address offset

hw/arm/aspeed_ast27x0.c static void aspeed_soc_ast2700_realize(DeviceState

Quoted 27 lines · Jamin Lin

On AST2700 platforms, system DRAM is mapped above 4GB with the base address at 0x400000000.

The Linux EHCI driver programs the segment register to zero when 64-bit addressing is supported. As a result, descriptor addresses derived from the EHCI registers do not include the DRAM base address.

Descriptor memory is allocated through the DMA API with a 64-bit DMA mask, which allows descriptors to be placed in DRAM above 4GB. When running on AST2700, this means EHCI queue heads (QH) and queue element transfer descriptors (qTD) reside at addresses starting from 0x400000000.

Set the descriptor-addr-offset property to the DRAM base so the emulated EHCI controller can construct the correct descriptor addresses when accessing system memory.

---
  hw/arm/aspeed_ast27x0.c | 3 +++
  1 file changed, 3 insertions(+)
hw/arm/aspeed_ast27x0.c index renamed from hw/arm/aspeed_ast27x0.c +0 -0

4a1f7cad73..3cdbf78ac1 100644

hw/arm/aspeed_ast27x0.c +0 -0
@@ -858,6 +858,9 @@ static void aspeed_soc_ast2700_realize(DeviceState

*dev, Error **errp)

hw/arm/aspeed_ast27x0.c static void aspeed_soc_ast2700_realize(DeviceState · line 859

Quoted 2 lines
hw/arm/aspeed_ast27x0.c +0 -0
      for (i = 0; i < sc->ehcis_num; i++) {
          object_property_set_bool(OBJECT(&s->ehci[i]),

"caps-64bit-addr", true,

Quoted 8 lines

hw/arm/aspeed_ast27x0.c static void aspeed_soc_ast2700_realize(DeviceState · line 861

Quoted 2 lines
hw/arm/aspeed_ast27x0.c +1 -0
                                   &error_abort);
+        object_property_set_int(OBJECT(&s->ehci[i]),

"descriptor-addr-offset",

hw/arm/aspeed_ast27x0.c static void aspeed_soc_ast2700_realize(DeviceState · line 862

Quoted 1 line
hw/arm/aspeed_ast27x0.c +1 -0
+

sc->memmap[ASPEED_DEV_SDRAM],

hw/arm/aspeed_ast27x0.c static void aspeed_soc_ast2700_realize(DeviceState *dev, Error **errp) · line 863

Quoted 1 line
hw/arm/aspeed_ast27x0.c +1 -0
+                                &error_abort);

Squash with previous? Or invert 14 <-> 15 order?

Thanks for the suggestion and review.
Will invert patch order.
Jamin

author
#55 lore raw

RE: [PATCH v3 13/17] hw/usb/hcd-ehci: Add descriptor address offset property

In reply to Cédric Le Goater

Hi Cédric,

Quoted 12 lines

Subject: Re: [PATCH v3 13/17] hw/usb/hcd-ehci: Add descriptor address offset property

Quoted 3 lines

I tried to address this by introducing a new property:
aspeed-ast2700-workaround.

This is not specific to AST2700.

I think a forced default value for CTRLDSSEGMENT would be more appropriate. Perhaps "ctrldssegment-default", as a uint32_t, which defaults to 0x0 and when the ast2700 SoC is realized, set to BIT(2).

Thanks for the suggestion and review.
Will do

Quoted 10 lines

hw/usb/hcd-ehci.c static void ehci_opreg_write(void *ptr, hwaddr addr, · line 1138

Quoted 5 lines
hw/usb/hcd-ehci.c +3 -0
+        if (s->aspeed_ast2700_workaround) {
+            val |= BIT(2);
+        }

With the proposal above, this would be changed to :

	val |= s->ctrldssegment_default;

Will do

Quoted 10 lines

hw/usb/hcd-ehci.c void usb_ehci_realize(EHCIState *s, DeviceState *dev, Error **errp) · line 2589

Quoted 9 lines
          break;
      case ASYNCLISTADDR:
          if (ehci_async_enabled(s)) { @@ -2582,6 +2585,9 @@ void
usb_ehci_realize(EHCIState *s, DeviceState *dev, Error **errp)
      }
      if (s->caps_64bit_addr) {
          s->caps[0x08] |= BIT(0);
hw/usb/hcd-ehci.c +2 -0
+        if (s->aspeed_ast2700_workaround) {
+            s->ctrldssegment |= BIT(2);
Quoted 2 lines

Shouldn't these assignments be in the reset handlers ?

Yes, with above solution, we do not need it in the reset handlers. Will remove it.
Appreciate your kindly support and help.

Jamin

Quoted 7 lines

Thanks,

C.

Quoted 2 lines

+ }

      }
author
#29 lore raw

RE: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

In reply to Cédric Le Goater

Hi Cédric

Quoted 138 lines

Subject: Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Quoted 99 lines · Jamin Lin

Hi Peter, Philippe, Cédric

Quoted 87 lines

Subject: Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Quoted 40 lines · Cédric Le Goater

+ Peter, Fabiano,

Quoted 37 lines · Jamin Lin

Change internal EHCI descriptor addresses from uint32_t to uint64_t.

The following fields are updated:
- EHCIPacket::qtdaddr
- EHCIQueue::{qhaddr, qtdaddr}
- EHCIState::{a_fetch_addr, p_fetch_addr}

Update get_dwords() and put_dwords() to take 64-bit addresses and propagate the type change through the descriptor traversal paths.

Adjust NLPTR_GET() to operate on 64-bit values:

      #define NLPTR_GET(x) ((x) & ~0x1fULL)

so that link pointer masking works correctly when descriptor addresses exceed 32-bit space. The previous mask (0xffffffe0) implicitly truncated addresses to 32 bits.

This patch does not change the on-wire descriptor layout yet. It only removes the internal 32-bit address limit and prepares for later patches that will add full 64-bit QH/qTD/iTD/siTD support.

Update the EHCI trace-events prototypes for QH, qTD, iTD, and siTD to use uint64_t for the address argument and print it with PRIx64. This ensures full 64-bit addresses are shown in trace output and improves debugging of queue heads and transfer descriptors.

Migration compatibility:

The fetch address fields in EHCIState are extended from 32-bit to 64-bit, requiring a VMState version bump (v2 -> v3).

Backward compatibility is preserved by keeping the legacy 32-bit fields (a_fetch_addr_pre_v3, p_fetch_addr_pre_v3) for pre-v3 migration streams via VMSTATE_UINT32_TEST(), and introducing new 64-bit fields for v3+.

Quoted 4 lines
Quoted 1 line

In post_load, pre-v3 state is promoted to the 64-bit fields.

Since ehci is supported by downstream, I'd prefer that the migration maintainers also verify the implementation.

Thanks, I'll only look at the migration bits.

[...]

Quoted 2 lines
Quoted 2 lines
(quoted diff) +0 -0
@@ -2441,6 +2445,11 @@ static int usb_ehci_post_load(void *opaque,

int

version_id)

Quoted 10 lines

hw/usb/hcd-ehci.c static void usb_ehci_vm_state_change(void

Quoted 10 lines
           }
       }
hw/usb/hcd-ehci.c +5 -0
+    if (version_id < 3) {
+        s->a_fetch_addr = s->a_fetch_addr_pre_v3;
+        s->p_fetch_addr = s->p_fetch_addr_pre_v3;
+    }
+
       return 0;
   }
@@ -2470,9 +2479,14 @@ static void usb_ehci_vm_state_change(void

*opaque, bool running, RunState state)

Quoted 20 lines
Quoted 20 lines
hw/usb/hcd-ehci.c +6 -3
       }
   }
+static bool ehci_core_is_before_version_3(void *opaque, int
+version_id) {
+    return version_id < 3;
+}
+
   const VMStateDescription vmstate_ehci = {
       .name        = "ehci-core",
-    .version_id  = 2,
+    .version_id  = 3,
       .minimum_version_id  = 1,
       .pre_save    = usb_ehci_pre_save,
       .post_load   = usb_ehci_post_load,
@@ -2501,8 +2515,12 @@ const VMStateDescription vmstate_ehci = {
           /* schedule state */
           VMSTATE_UINT32(astate, EHCIState),
           VMSTATE_UINT32(pstate, EHCIState),
-        VMSTATE_UINT32(a_fetch_addr, EHCIState),
-        VMSTATE_UINT32(p_fetch_addr, EHCIState),
Quoted 9 lines
Quoted 6 lines

hw/usb/hcd-ehci.c const VMStateDescription vmstate_ehci = { · line 2523

Quoted 6 lines
hw/usb/hcd-ehci.c +6 -0
+        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),

TL;DR: I feel like we still need machine type compat properties.

Quoted 8 lines

I don't understand this: Could you please describe it in more detail?

Quoted 6 lines

Details:

When a v2 stream arrives, two _TEST()s will do the loading, then post_load() extend it to 64bits, looks fine.

When a v3 stream arrives, two _TEST()s got skipped then latter two take

effect.

Quoted 6 lines
Quoted 6 lines

post_load() skips. Looks fine.

When migrating to another QEMU, due to the fact saving vmstates always take vmsd's version declared (3), I don't see how it can migrate back to a v2 stream; it didn't know about v3.

Quoted 6 lines
Quoted 4 lines

Jiamin, have you tested migrating from a new QEMU binary back to another old one? For upstream and serious devices, we need to guarantee bi-directional migrations, back and forth.

Sorry, I am not familiar with the design of QEMU migration, and I don't know

how to test it.

Quoted 2 lines

I added these hook functions based on Philippe's suggestion and review. Could you tell me how to test migration to ensure both version 3 and version

2 are workable?

Migrating a running guest back and forth between QEMU [1] and [2] should be possible :

[1] path/to/qemu-10.2/install/bin/qemu-system-x86_64 -M pc-q35-10.2 \
     ... \
     -device usb-ehci \
     -device usb-kbd
Quoted 10 lines

[2] path/to/myqemu-with-patches/install/bin/qemu-system-x86_64 -M

pc-q35-10.2 \
     ... \
     -device usb-ehci \
     -device usb-kbd

This is required to guarantee forward and backward compatibility of QEMU pc-q35-10.2 machine instances.

Thanks for the migration test steps.
I have tested both forward and backward migration for EHCI and observed failures in both directions. I will revise my changes in a follow-up patch to address these issues and ensure that the migration tests pass.

Thanks,
Jamin

Quoted 1 line

C.

author
#24 lore raw

RE: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

In reply to Peter Xu

Hi Peter, Cédric and Philippe

Quoted 79 lines

Subject: Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Quoted 1 line · Jamin Lin

Hi Peter, Philippe, Cédric

Hi, Jamin,

[...]

Quoted 16 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
(quoted diff) +0 -2
@@ -2501,8 +2515,12 @@ const VMStateDescription vmstate_ehci = {
          /* schedule state */
          VMSTATE_UINT32(astate, EHCIState),
          VMSTATE_UINT32(pstate, EHCIState),
-        VMSTATE_UINT32(a_fetch_addr, EHCIState),
-        VMSTATE_UINT32(p_fetch_addr, EHCIState),
Quoted 9 lines
Quoted 6 lines

Review const VMStateDescription vmstate_ehci = { · line 2523

Quoted 6 lines
(quoted diff) +6 -0
+        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),

TL;DR: I feel like we still need machine type compat properties.

Quoted 1 line

I don't understand this: Could you please describe it in more detail?

See hw/core/machine.c and entries added into hw_compat_*[] arrays. IIUC, if this is an important device and if we want to guarantee bidirectional migration (I'll explain later), then we should stick with machine compat properties.

Quoted 7 lines
Quoted 7 lines

Details:

When a v2 stream arrives, two _TEST()s will do the loading, then post_load() extend it to 64bits, looks fine.

When a v3 stream arrives, two _TEST()s got skipped then latter two take

effect.

Quoted 6 lines
Quoted 6 lines

post_load() skips. Looks fine.

When migrating to another QEMU, due to the fact saving vmstates always take vmsd's version declared (3), I don't see how it can migrate back to a v2 stream; it didn't know about v3.

Quoted 6 lines
Quoted 4 lines

Jiamin, have you tested migrating from a new QEMU binary back to another old one? For upstream and serious devices, we need to guarantee bi-directional migrations, back and forth.

Sorry, I am not familiar with the design of QEMU migration, and I don't know

how to test it.

Quoted 2 lines

I added these hook functions based on Philippe's suggestion and review. Could you tell me how to test migration to ensure both version 3 and version

2 are workable?

Since this patch already modified VMSDs, I believe migration test should have been carried out at the very minimum.. Logically if vmsd versioning is used, we should also have tested forward migrations, because that's only for that.

What I'm talking about is I think backward migration will fail.

Forward migration describes the case when a VM hosted by an old QEMU binary (e.g. QEMU v10.2.0) is migrated to a new QEMU (e.g., after this patch applied on top of master).

Backward migration describes the case when a VM hosted by a new QEMU binary (e.g., after this patch applied on top of master) is migrated back to an old QEMU binary (e.g. QEMU v10.2.0).

For different devices, we should allow different level of support on migration. Any device that may be important to either upstream QEMU or downstreams should better allow bi-directional migration. In this case we need to use machine compat properties I mentioned. I'm not yet sure where EHCI controller falls e.g. v.s. XHCI.. when not sure, we can also just stick with machine compat properties to be on the safe side.

Quoted 8 lines

For more info, you can also refer to:

https://www.qemu.org/docs/master/devel/migration/compatibility.html

Thanks,

--
Peter Xu

Thanks for your support and the detailed explanation.
Following Cédric’s comments, I now understand how to properly test migration. With the changes below, both forward and backward migration tests pass. Could you please review these changes, or would you prefer that I resend them as v4?

Thanks again for your guidance and support.

Jamin

hw/usb/hcd-ehci.h +4 -1
index 310183bbc4..bd165c793c 100644
@@ -264,6 +264,7 @@ struct EHCIState {
     uint32_t maxframes;
     bool caps_64bit_addr;
     uint32_t ctrldssegment_default;
+    bool migrate_fetch_addr_64bit;
 
     /*
      *  EHCI spec version 1.0 Section 2.3
@@ -321,7 +322,9 @@ struct EHCIState {
     DEFINE_PROP_UINT32("maxframes", _state, ehci.maxframes, 128), \
     DEFINE_PROP_BOOL("caps-64bit-addr", _state, ehci.caps_64bit_addr, false), \
     DEFINE_PROP_UINT32("ctrldssegment-default", _state, \
-                       ehci.ctrldssegment_default, 0)
+                       ehci.ctrldssegment_default, 0), \
+    DEFINE_PROP_BOOL("x-migrate-fetch-addr-64bit", _state, \
+                     ehci.migrate_fetch_addr_64bit, true)
 
 extern const VMStateDescription vmstate_ehci;
 
hw/core/machine.c +2 -0
index 0aa77a57e9..e77400c7ab 100644
@@ -41,6 +41,8 @@
 GlobalProperty hw_compat_10_2[] = {
     { "scsi-block", "migrate-pr", "off" },
     { "isa-cirrus-vga", "global-vmstate", "true" },
+    { "sysbus-ehci-usb", "x-migrate-fetch-addr-64bit", "off" },
+    { "pci-ehci-usb", "x-migrate-fetch-addr-64bit", "off" },
 };
 const size_t hw_compat_10_2_len = G_N_ELEMENTS(hw_compat_10_2);
 
hw/usb/hcd-ehci.c +30 -14

74 lines not on this page · raw message

#25 lore raw

Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Quoted 85 lines · Jamin Lin

Hi Peter, Cédric and Philippe

Quoted 83 lines

Subject: Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Quoted 1 line · Jamin Lin

Hi Peter, Philippe, Cédric

Hi, Jamin,

[...]

Quoted 16 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
hw/usb/hcd-ehci.c +0 -2
@@ -2501,8 +2515,12 @@ const VMStateDescription vmstate_ehci = {
          /* schedule state */
          VMSTATE_UINT32(astate, EHCIState),
          VMSTATE_UINT32(pstate, EHCIState),
-        VMSTATE_UINT32(a_fetch_addr, EHCIState),
-        VMSTATE_UINT32(p_fetch_addr, EHCIState),
Quoted 9 lines
Quoted 6 lines

Review const VMStateDescription vmstate_ehci = { · line 2523

Quoted 6 lines
(quoted diff) +6 -0
+        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3, EHCIState,
+                            ehci_core_is_before_version_3),
+        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),

TL;DR: I feel like we still need machine type compat properties.

Quoted 1 line

I don't understand this: Could you please describe it in more detail?

See hw/core/machine.c and entries added into hw_compat_*[] arrays. IIUC, if this is an important device and if we want to guarantee bidirectional migration (I'll explain later), then we should stick with machine compat properties.

Quoted 7 lines
Quoted 7 lines

Details:

When a v2 stream arrives, two _TEST()s will do the loading, then post_load() extend it to 64bits, looks fine.

When a v3 stream arrives, two _TEST()s got skipped then latter two take

effect.

Quoted 6 lines
Quoted 6 lines

post_load() skips. Looks fine.

When migrating to another QEMU, due to the fact saving vmstates always take vmsd's version declared (3), I don't see how it can migrate back to a v2 stream; it didn't know about v3.

Quoted 6 lines
Quoted 4 lines

Jiamin, have you tested migrating from a new QEMU binary back to another old one? For upstream and serious devices, we need to guarantee bi-directional migrations, back and forth.

Sorry, I am not familiar with the design of QEMU migration, and I don't know

how to test it.

Quoted 2 lines

I added these hook functions based on Philippe's suggestion and review. Could you tell me how to test migration to ensure both version 3 and version

2 are workable?

Since this patch already modified VMSDs, I believe migration test should have been carried out at the very minimum.. Logically if vmsd versioning is used, we should also have tested forward migrations, because that's only for that.

What I'm talking about is I think backward migration will fail.

Forward migration describes the case when a VM hosted by an old QEMU binary (e.g. QEMU v10.2.0) is migrated to a new QEMU (e.g., after this patch applied on top of master).

Backward migration describes the case when a VM hosted by a new QEMU binary (e.g., after this patch applied on top of master) is migrated back to an old QEMU binary (e.g. QEMU v10.2.0).

For different devices, we should allow different level of support on migration. Any device that may be important to either upstream QEMU or downstreams should better allow bi-directional migration. In this case we need to use machine compat properties I mentioned. I'm not yet sure where EHCI controller falls e.g. v.s. XHCI.. when not sure, we can also just stick with machine compat properties to be on the safe side.

For more info, you can also refer to:

https://www.qemu.org/docs/master/devel/migration/compatibility.html

Quoted 7 lines
Quoted 4 lines

Thanks,

--
Peter Xu

Thanks for your support and the detailed explanation.
Following Cédric’s comments, I now understand how to properly test migration. With the changes below, both forward and backward migration tests pass. Could you please review these changes, or would you prefer that I resend them as v4?

Per what I read on Phil's reply on the cover letter, he prefers you to resend anyway. Better do it. Comments inline.

Quoted 18 lines

Thanks again for your guidance and support.

Jamin

hw/usb/hcd-ehci.h +1 -0
index 310183bbc4..bd165c793c 100644
@@ -264,6 +264,7 @@ struct EHCIState {
     uint32_t maxframes;
     bool caps_64bit_addr;
     uint32_t ctrldssegment_default;
+    bool migrate_fetch_addr_64bit;
 
     /*
      *  EHCI spec version 1.0 Section 2.3

hw/usb/hcd-ehci.h struct EHCIState { · line 327

Quoted 7 lines
hw/usb/hcd-ehci.h +3 -1
@@ -321,7 +322,9 @@ struct EHCIState {
     DEFINE_PROP_UINT32("maxframes", _state, ehci.maxframes, 128), \
     DEFINE_PROP_BOOL("caps-64bit-addr", _state, ehci.caps_64bit_addr, false), \
     DEFINE_PROP_UINT32("ctrldssegment-default", _state, \
-                       ehci.ctrldssegment_default, 0)
+                       ehci.ctrldssegment_default, 0), \
+    DEFINE_PROP_BOOL("x-migrate-fetch-addr-64bit", _state, \
+                     ehci.migrate_fetch_addr_64bit, true)

The diff looks a bit weird; this chunk is not present in hw/usb/hcd-ehci.h in master branch, but maybe this series touched it.. in that case it's fine.

hw/core/machine.c line 45

Quoted 13 lines
hw/usb/hcd-ehci.h +0 -0
 extern const VMStateDescription vmstate_ehci;
 
hw/core/machine.c +2 -0
index 0aa77a57e9..e77400c7ab 100644
@@ -41,6 +41,8 @@
 GlobalProperty hw_compat_10_2[] = {
     { "scsi-block", "migrate-pr", "off" },
     { "isa-cirrus-vga", "global-vmstate", "true" },
+    { "sysbus-ehci-usb", "x-migrate-fetch-addr-64bit", "off" },
+    { "pci-ehci-usb", "x-migrate-fetch-addr-64bit", "off" },

I'm still looking at master branch, vmstate_ehci is referenced in both "ehci-sysbus" (vmstate_ehci_sysbus) and "ehci" (vmstate_ehci_pci). Here you used different names, I'm not sure if it's relevant to your other patches, please double check.

Also if there're two drivers involved, IIUC we should need to add two bool properties.. I only saw one above. Again, not sure if it's relevant to what your prior patch did, but please double check.

hw/usb/hcd-ehci.c static int usb_ehci_pre_save(void *opaque) · line 2473

Quoted 13 lines
hw/core/machine.c +0 -0
 };
 const size_t hw_compat_10_2_len = G_N_ELEMENTS(hw_compat_10_2);
 
hw/usb/hcd-ehci.c +2 -0
index 983b1731f9..87528ddc02 100644
@@ -2469,6 +2469,9 @@ static int usb_ehci_pre_save(void *opaque)
     ehci->last_run_ns -= (ehci->frindex - new_frindex) * UFRAME_TIMER_NS;
     ehci->frindex = new_frindex;
 
+    ehci->a_fetch_addr_pre_v3 = ehci->a_fetch_addr;
+    ehci->p_fetch_addr_pre_v3 = ehci->p_fetch_addr;

This looks fine if you always prepare the 32bit versions of the addr, but it can also be put under a "if (!ehci_fetch_addr_64_needed(...))". Not a big deal.

Quoted 5 lines
hw/usb/hcd-ehci.c +1 -0
+
     return 0;
 }
 

hw/usb/hcd-ehci.c static int usb_ehci_post_load(void *opaque, int version_id) · line 2496

Quoted 9 lines
hw/usb/hcd-ehci.c +2 -4
@@ -2489,10 +2492,8 @@ static int usb_ehci_post_load(void *opaque, int version_id)
         }
     }
 
-    if (version_id < 3) {
-        s->a_fetch_addr = s->a_fetch_addr_pre_v3;
-        s->p_fetch_addr = s->p_fetch_addr_pre_v3;
-    }
+    s->a_fetch_addr = s->a_fetch_addr_pre_v3;
+    s->p_fetch_addr = s->p_fetch_addr_pre_v3;

This one looks risky, it will also run on new machine types, I think a "if (!ehci_fetch_addr_64_needed(...))" is required here (unlike the pre_save), otherwise I suspect you may overwrite the 64bits with 32bits after migration. It may explode when high 32bits have something nonzero?

Quoted 5 lines
hw/usb/hcd-ehci.c +0 -0
     return 0;
 }
@@ -2523,14 +2524,29 @@ static void usb_ehci_vm_state_change(void *opaque, bool running, RunState state)
     }

hw/usb/hcd-ehci.c static void usb_ehci_vm_state_change(void *opaque, bool running, RunState state) · line 2532

Quoted 10 lines
hw/usb/hcd-ehci.c +4 -2
 }
 
-static bool ehci_core_is_before_version_3(void *opaque, int version_id)
+static bool ehci_fetch_addr_64_needed(void *opaque)
 {
-    return version_id < 3;
+    EHCIState *s = opaque;
+
+    return s->migrate_fetch_addr_64bit;
 }

Yes this is the traditional way.

Quoted 28 lines
hw/usb/hcd-ehci.c +14 -4
+static const VMStateDescription vmstate_ehci_fetch_addr_64 = {
+    .name = "ehci-core/fetch-addr64",
+    .version_id = 1,
+    .minimum_version_id = 1,
+    .needed = ehci_fetch_addr_64_needed,
+    .fields = (const VMStateField[]) {
+        /* mmio registers */
+        VMSTATE_UINT64(a_fetch_addr, EHCIState),
+        VMSTATE_UINT64(p_fetch_addr, EHCIState),
+        VMSTATE_END_OF_LIST()
+    }
+};
+
 const VMStateDescription vmstate_ehci = {
     .name        = "ehci-core",
-    .version_id  = 3,
+    .version_id  = 2,
     .minimum_version_id  = 1,
     .pre_save    = usb_ehci_pre_save,
     .post_load   = usb_ehci_post_load,
@@ -2559,14 +2575,14 @@ const VMStateDescription vmstate_ehci = {
         /* schedule state */
         VMSTATE_UINT32(astate, EHCIState),
         VMSTATE_UINT32(pstate, EHCIState),
-        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3, EHCIState,
-                            ehci_core_is_before_version_3),
-        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3, EHCIState,

hw/usb/hcd-ehci.c const VMStateDescription vmstate_ehci = { · line 2583

Quoted 10 lines
hw/usb/hcd-ehci.c +5 -4
-                            ehci_core_is_before_version_3),
-        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
-        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT32(a_fetch_addr_pre_v3, EHCIState),
+        VMSTATE_UINT32(p_fetch_addr_pre_v3, EHCIState),
         VMSTATE_END_OF_LIST()
-    }
+    },
+    .subsections = (const VMStateDescription * const []) {
+        &vmstate_ehci_fetch_addr_64,

IMHO you don't need to make it a subsection, you can use VMSTATE_UINT64_TEST().

Similarly you can also use VMSTATE_UINT32_TEST() to only migrate 32bit versions in old machine types.

With those you can keep them together and add a comment, may look better than subsections where you split them apart, it'll be harder to follow in the future.

Thanks,

Quoted 5 lines
hw/usb/hcd-ehci.c +2 -0
+        NULL,
+    },
 };
 
 void usb_ehci_realize(EHCIState *s, DeviceState *dev, Error **errp)
Signature · 1 line
Peter Xu
author
#26 lore raw

RE: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Hi Peter,

Quoted 106 lines

Subject: Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Quoted 5 lines · Jamin Lin

Hi Peter, Cédric and Philippe

Quoted 3 lines

Subject: Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Quoted 8 lines
Quoted 8 lines · Jamin Lin
Quoted 1 line

Hi Peter, Philippe, Cédric

Hi, Jamin,

[...]

Quoted 1 line
Quoted 1 line
Quoted 1 line

Review const VMStateDescription vmstate_ehci

Quoted 1 line
(quoted diff) +0 -0
@@ -2501,8 +2515,12 @@ const VMStateDescription vmstate_ehci

= {

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

hw/usb/hcd-ehci.c const VMStateDescription vmstate_ehci = {

Quoted 6 lines
hw/usb/hcd-ehci.c +1 -2
          /* schedule state */
          VMSTATE_UINT32(astate, EHCIState),
          VMSTATE_UINT32(pstate, EHCIState),
-        VMSTATE_UINT32(a_fetch_addr, EHCIState),
-        VMSTATE_UINT32(p_fetch_addr, EHCIState),
+        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3,

EHCIState,

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

Review const VMStateDescription vmstate_ehci · line 2519

Quoted 1 line
(quoted diff) +1 -0
+

ehci_core_is_before_version_3),

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

Review const VMStateDescription vmstate_ehci · line 2520

Quoted 1 line
(quoted diff) +1 -0
+        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3,

EHCIState,

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

Review const VMStateDescription vmstate_ehci · line 2521

Quoted 1 line
(quoted diff) +1 -0
+

ehci_core_is_before_version_3),

Quoted 36 lines
Quoted 36 lines
Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

Review const VMStateDescription vmstate_ehci · line 2523

Quoted 2 lines
(quoted diff) +2 -0
+        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),

TL;DR: I feel like we still need machine type compat properties.

I don't understand this: Could you please describe it in more detail?

See hw/core/machine.c and entries added into hw_compat_*[] arrays. IIUC, if this is an important device and if we want to guarantee bidirectional migration (I'll explain later), then we should stick with machine compat properties.

Quoted 8 lines
Quoted 8 lines

Details:

When a v2 stream arrives, two _TEST()s will do the loading, then post_load() extend it to 64bits, looks fine.

When a v3 stream arrives, two _TEST()s got skipped then latter two take

effect.

Quoted 6 lines
Quoted 6 lines

post_load() skips. Looks fine.

When migrating to another QEMU, due to the fact saving vmstates always take vmsd's version declared (3), I don't see how it can migrate back to a v2 stream; it didn't know about v3.

Quoted 7 lines
Quoted 4 lines

Jiamin, have you tested migrating from a new QEMU binary back to another old one? For upstream and serious devices, we need to guarantee bi-directional migrations, back and forth.

Sorry, I am not familiar with the design of QEMU migration, and I don't know

how to test it.

Quoted 8 lines
Quoted 8 lines
Quoted 3 lines

I added these hook functions based on Philippe's suggestion and review. Could you tell me how to test migration to ensure both version 3 and version

2 are workable?

Since this patch already modified VMSDs, I believe migration test should have been carried out at the very minimum.. Logically if vmsd versioning is used, we should also have tested forward migrations,

because that's only for that.

Quoted 7 lines
Quoted 7 lines

What I'm talking about is I think backward migration will fail.

Forward migration describes the case when a VM hosted by an old QEMU binary (e.g. QEMU v10.2.0) is migrated to a new QEMU (e.g., after this patch applied on top of master).

Quoted 5 lines
Quoted 5 lines

Backward migration describes the case when a VM hosted by a new QEMU binary (e.g., after this patch applied on top of master) is migrated back to an old QEMU binary (e.g. QEMU v10.2.0).

For different devices, we should allow different level of support on

migration.

Quoted 5 lines
Quoted 5 lines

Any device that may be important to either upstream QEMU or downstreams should better allow bi-directional migration. In this case we need to use machine compat properties I mentioned. I'm not yet sure where EHCI controller falls e.g. v.s. XHCI.. when not sure, we can also just stick with machine compat properties to be on the safe

side.

Quoted 11 lines
Quoted 9 lines

For more info, you can also refer to:

https://www.qemu.org/docs/master/devel/migration/compatibility.html

Thanks,

--
Peter Xu

Thanks for your support and the detailed explanation. Following Cédric’s comments, I now understand how to properly test

Quoted 8 lines

migration. With the changes below, both forward and backward migration tests pass.

Quoted 1 line

Could you please review these changes, or would you prefer that I resend

them as v4?

Per what I read on Phil's reply on the cover letter, he prefers you to resend anyway. Better do it. Comments inline.

Thanks for the review and suggestion.
Will resend v4 soon and loop you.

Quoted 33 lines
Quoted 17 lines

Thanks again for your guidance and support.

Jamin

hw/usb/hcd-ehci.h index renamed from hw/usb/hcd-ehci.h +0 -0

310183bbc4..bd165c793c 100644

hw/usb/hcd-ehci.h +1 -0
@@ -264,6 +264,7 @@ struct EHCIState {
     uint32_t maxframes;
     bool caps_64bit_addr;
     uint32_t ctrldssegment_default;
+    bool migrate_fetch_addr_64bit;
 
     /*
      *  EHCI spec version 1.0 Section 2.3 @@ -321,7 +322,9 @@ struct
Quoted 3 lines
EHCIState {
     DEFINE_PROP_UINT32("maxframes", _state, ehci.maxframes, 128), \
     DEFINE_PROP_BOOL("caps-64bit-addr", _state, ehci.caps_64bit_addr,

false), \

hw/usb/hcd-ehci.h struct EHCIState { · line 327

Quoted 5 lines
     DEFINE_PROP_UINT32("ctrldssegment-default", _state, \
hw/usb/hcd-ehci.h +3 -1
-                       ehci.ctrldssegment_default, 0)
+                       ehci.ctrldssegment_default, 0), \
+    DEFINE_PROP_BOOL("x-migrate-fetch-addr-64bit", _state, \
+                     ehci.migrate_fetch_addr_64bit, true)

The diff looks a bit weird; this chunk is not present in hw/usb/hcd-ehci.h in master branch, but maybe this series touched it.. in that case it's fine.

Quoted 3 lines

extern const VMStateDescription vmstate_ehci;

Quoted 16 lines

hw/core/machine.c line 45

Quoted 10 lines
hw/core/machine.c index renamed from hw/core/machine.c +0 -0

0aa77a57e9..e77400c7ab 100644

hw/core/machine.c +2 -0
@@ -41,6 +41,8 @@
 GlobalProperty hw_compat_10_2[] = {
     { "scsi-block", "migrate-pr", "off" },
     { "isa-cirrus-vga", "global-vmstate", "true" },
+    { "sysbus-ehci-usb", "x-migrate-fetch-addr-64bit", "off" },
+    { "pci-ehci-usb", "x-migrate-fetch-addr-64bit", "off" },

I'm still looking at master branch, vmstate_ehci is referenced in both "ehci-sysbus" (vmstate_ehci_sysbus) and "ehci" (vmstate_ehci_pci). Here you used different names, I'm not sure if it's relevant to your other patches, please double check.

It seems that we need to use the device type name here. For the PCI EHCI controller, looking at hcd-ehci-pci.c:

The VMStateDescription name is "ehci":

static const VMStateDescription vmstate_ehci_pci = {
.name = "ehci",
.version_id = 2,
.minimum_version_id = 1,
.fields = (const VMStateField[]) {
VMSTATE_PCI_DEVICE(pcidev, EHCIPCIState), VMSTATE_STRUCT(ehci, EHCIPCIState, 2, vmstate_ehci, EHCIState), VMSTATE_END_OF_LIST()
}
};

However, the actual device type name is defined via TYPE_PCI_EHCI, which resolves to "pci-ehci-usb" and is used when registering the device:

static void ehci_pci_register_types(void)

{
    TypeInfo ehci_type_info = {
        .parent        = TYPE_PCI_EHCI,
        .class_init    = ehci_data_class_init,
    };
    int i;

    type_register_static(&ehci_pci_type_info);

    for (i = 0; i < ARRAY_SIZE(ehci_pci_info); i++) {
        ehci_type_info.name = ehci_pci_info[i].name;
        ehci_type_info.class_data = ehci_pci_info + i;
        type_register_static(&ehci_type_info);
    }
}

So in machine compat properties, it seems we should use the device type name ("pci-ehci-usb"), rather than the VMStateDescription name ("ehci").

I also tried changing it to "ehci", but migration failed in that case, so it looks like "pci-ehci-usb" is the correct one to use here.

Could you please confirm if this is the expected usage?

Quoted 14 lines

Also if there're two drivers involved, IIUC we should need to add two bool properties.. I only saw one above. Again, not sure if it's relevant to what your prior patch did, but please double check.

hw/usb/hcd-ehci.c static int usb_ehci_pre_save(void *opaque) · line 2469

Quoted 9 lines
hw/core/machine.c +0 -0
 };
 const size_t hw_compat_10_2_len = G_N_ELEMENTS(hw_compat_10_2);
 
hw/usb/hcd-ehci.c index renamed from hw/usb/hcd-ehci.c +0 -0

983b1731f9..87528ddc02 100644

hw/usb/hcd-ehci.c +0 -0
@@ -2469,6 +2469,9 @@ static int usb_ehci_pre_save(void *opaque)
     ehci->last_run_ns -= (ehci->frindex - new_frindex) *

UFRAME_TIMER_NS;

Quoted 8 lines

hw/usb/hcd-ehci.c static int usb_ehci_pre_save(void *opaque) · line 2473

Quoted 4 lines
hw/usb/hcd-ehci.c +2 -0
     ehci->frindex = new_frindex;
 
+    ehci->a_fetch_addr_pre_v3 = ehci->a_fetch_addr;
+    ehci->p_fetch_addr_pre_v3 = ehci->p_fetch_addr;

This looks fine if you always prepare the 32bit versions of the addr, but it can also be put under a "if (!ehci_fetch_addr_64_needed(...))". Not a big deal.

Will add

Quoted 16 lines

hw/usb/hcd-ehci.c static int usb_ehci_post_load(void *opaque, int

Quoted 5 lines
hw/usb/hcd-ehci.c +1 -0
+
     return 0;
 }
 
@@ -2489,10 +2492,8 @@ static int usb_ehci_post_load(void *opaque, int

version_id)

hw/usb/hcd-ehci.c static int usb_ehci_post_load(void *opaque, int version_id) · line 2496

Quoted 9 lines
hw/usb/hcd-ehci.c +2 -4
         }
     }
 
-    if (version_id < 3) {
-        s->a_fetch_addr = s->a_fetch_addr_pre_v3;
-        s->p_fetch_addr = s->p_fetch_addr_pre_v3;
-    }
+    s->a_fetch_addr = s->a_fetch_addr_pre_v3;
+    s->p_fetch_addr = s->p_fetch_addr_pre_v3;
Quoted 5 lines

This one looks risky, it will also run on new machine types, I think a "if (!ehci_fetch_addr_64_needed(...))" is required here (unlike the pre_save), otherwise I suspect you may overwrite the 64bits with 32bits after migration. It may explode when high 32bits have something nonzero?

Will add

Quoted 21 lines

hw/usb/hcd-ehci.c static void usb_ehci_vm_state_change(void

Quoted 4 lines
hw/usb/hcd-ehci.c +0 -0
     return 0;
 }
@@ -2523,14 +2524,29 @@ static void usb_ehci_vm_state_change(void

*opaque, bool running, RunState state)

hw/usb/hcd-ehci.c static void usb_ehci_vm_state_change(void *opaque, bool running, RunState state) · line 2532

Quoted 12 lines
hw/usb/hcd-ehci.c +4 -2
     }
 }
 
-static bool ehci_core_is_before_version_3(void *opaque, int
 version_id)
+static bool ehci_fetch_addr_64_needed(void *opaque)
 {
-    return version_id < 3;
+    EHCIState *s = opaque;
+
+    return s->migrate_fetch_addr_64bit;
 }

Yes this is the traditional way.

Quoted 1 line
Quoted 13 lines

hw/usb/hcd-ehci.c static void usb_ehci_vm_state_change(void *opaque, bool running, RunState state) · line 2546

Quoted 13 lines
hw/usb/hcd-ehci.c +13 -0
+static const VMStateDescription vmstate_ehci_fetch_addr_64 = {
+    .name = "ehci-core/fetch-addr64",
+    .version_id = 1,
+    .minimum_version_id = 1,
+    .needed = ehci_fetch_addr_64_needed,
+    .fields = (const VMStateField[]) {
+        /* mmio registers */
+        VMSTATE_UINT64(a_fetch_addr, EHCIState),
+        VMSTATE_UINT64(p_fetch_addr, EHCIState),
+        VMSTATE_END_OF_LIST()
+    }
+};
+

Will remove

Quoted 19 lines
Quoted 19 lines

const VMStateDescription vmstate_ehci = {
.name = "ehci-core",

hw/usb/hcd-ehci.c +3 -7
-    .version_id  = 3,
+    .version_id  = 2,
     .minimum_version_id  = 1,
     .pre_save    = usb_ehci_pre_save,
     .post_load   = usb_ehci_post_load,
@@ -2559,14 +2575,14 @@ const VMStateDescription vmstate_ehci = {
         /* schedule state */
         VMSTATE_UINT32(astate, EHCIState),
         VMSTATE_UINT32(pstate, EHCIState),
-        VMSTATE_UINT32_TEST(a_fetch_addr_pre_v3, EHCIState,
-                            ehci_core_is_before_version_3),
-        VMSTATE_UINT32_TEST(p_fetch_addr_pre_v3, EHCIState,
-                            ehci_core_is_before_version_3),
-        VMSTATE_UINT64_V(a_fetch_addr, EHCIState, 3),
-        VMSTATE_UINT64_V(p_fetch_addr, EHCIState, 3),
+        VMSTATE_UINT32(a_fetch_addr_pre_v3, EHCIState),
+        VMSTATE_UINT32(p_fetch_addr_pre_v3, EHCIState),
Quoted 9 lines

hw/usb/hcd-ehci.c const VMStateDescription vmstate_ehci = { · line 2583

Quoted 5 lines
hw/usb/hcd-ehci.c +3 -1
         VMSTATE_END_OF_LIST()
-    }
+    },
+    .subsections = (const VMStateDescription * const []) {
+        &vmstate_ehci_fetch_addr_64,

IMHO you don't need to make it a subsection, you can use VMSTATE_UINT64_TEST().

Will add

Quoted 3 lines

Similarly you can also use VMSTATE_UINT32_TEST() to only migrate 32bit versions in old machine types.

Will add

Quoted 6 lines

With those you can keep them together and add a comment, may look better than subsections where you split them apart, it'll be harder to follow in the future.

Thanks,

Quoted 8 lines

hw/usb/hcd-ehci.c const VMStateDescription vmstate_ehci = { · line 2588

Quoted 5 lines
hw/usb/hcd-ehci.c +2 -0
+        NULL,
+    },
 };
 
 void usb_ehci_realize(EHCIState *s, DeviceState *dev, Error **errp)

--
Peter Xu

Thanks,
Jamin

#27 lore raw

Re: [PATCH v3 06/17] hw/usb/hcd-ehci: Change descriptor addresses to 64-bit

Quoted 1 line · Jamin Lin

It seems that we need to use the device type name here.

Heh, yeah, you should be right.

Signature · 1 line
Peter Xu