ahci: add rwerror=stop support for ncq

Handle NCQ failures for cases where we want to halt the VM on IO errors.
Upon a VM state change, retry the halted NCQ commands.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 1435767578-32743-7-git-send-email-jsnow@redhat.com

4 files changed, 44 insertions, 2 deletions

diff --git a/hw/ide/ahci.c b/hw/ide/ahci.c
index b0b9b41ed0..d996f37b36 100644
--- a/hw/ide/ahci.c
+++ b/hw/ide/ahci.c
@@ -581,6 +581,7 @@ static void ahci_reset_port(AHCIState *s, int port)
     /* reset ncq queue */
     for (i = 0; i < AHCI_MAX_CMDS; i++) {
         NCQTransferState *ncq_tfs = &s->dev[port].ncq_tfs[i];
+        ncq_tfs->halt = false;
         if (!ncq_tfs->used) {
             continue;
         }
@@ -960,12 +961,23 @@ static void ncq_cb(void *opaque, int ret)
     }
 
     if (ret < 0) {
-        ncq_err(ncq_tfs);
+        bool is_read = ncq_tfs->cmd == READ_FPDMA_QUEUED;
+        BlockErrorAction action = blk_get_error_action(ide_state->blk,
+                                                       is_read, -ret);
+        if (action == BLOCK_ERROR_ACTION_STOP) {
+            ncq_tfs->halt = true;
+            ide_state->bus->error_status = IDE_RETRY_HBA;
+        } else if (action == BLOCK_ERROR_ACTION_REPORT) {
+            ncq_err(ncq_tfs);
+        }
+        blk_error_action(ide_state->blk, action, is_read, -ret);
     } else {
         ide_state->status = READY_STAT | SEEK_STAT;
     }
 
-    ncq_finish(ncq_tfs);
+    if (!ncq_tfs->halt) {
+        ncq_finish(ncq_tfs);
+    }
 }
 
 static int is_ncq(uint8_t ata_cmd)
@@ -988,7 +1000,9 @@ static void execute_ncq_command(NCQTransferState *ncq_tfs)
     AHCIDevice *ad = ncq_tfs->drive;
     IDEState *ide_state = &ad->port.ifs[0];
     int port = ad->port_no;
+
     g_assert(is_ncq(ncq_tfs->cmd));
+    ncq_tfs->halt = false;
 
     switch (ncq_tfs->cmd) {
     case READ_FPDMA_QUEUED:
@@ -1319,6 +1333,23 @@ static void ahci_restart_dma(IDEDMA *dma)
 }
 
 /**
+ * IDE/PIO restarts are handled by the core layer, but NCQ commands
+ * need an extra kick from the AHCI HBA.
+ */
+static void ahci_restart(IDEDMA *dma)
+{
+    AHCIDevice *ad = DO_UPCAST(AHCIDevice, dma, dma);
+    int i;
+
+    for (i = 0; i < AHCI_MAX_CMDS; i++) {
+        NCQTransferState *ncq_tfs = &ad->ncq_tfs[i];
+        if (ncq_tfs->halt) {
+            execute_ncq_command(ncq_tfs);
+        }
+    }
+}
+
+/**
  * Called in DMA R/W chains to read the PRDT, utilizing ahci_populate_sglist.
  * Not currently invoked by PIO R/W chains,
  * which invoke ahci_populate_sglist via ahci_start_transfer.
@@ -1406,6 +1437,7 @@ static void ahci_irq_set(void *opaque, int n, int level)
 
 static const IDEDMAOps ahci_dma_ops = {
     .start_dma = ahci_start_dma,
+    .restart = ahci_restart,
     .restart_dma = ahci_restart_dma,
     .start_transfer = ahci_start_transfer,
     .prepare_buf = ahci_dma_prepare_buf,
diff --git a/hw/ide/ahci.h b/hw/ide/ahci.h
index 33607d7fad..47a3122270 100644
--- a/hw/ide/ahci.h
+++ b/hw/ide/ahci.h
@@ -262,6 +262,7 @@ typedef struct NCQTransferState {
     uint8_t cmd;
     int slot;
     int used;
+    bool halt;
 } NCQTransferState;
 
 struct AHCIDevice {
diff --git a/hw/ide/core.c b/hw/ide/core.c
index be7c350b87..122e955084 100644
--- a/hw/ide/core.c
+++ b/hw/ide/core.c
@@ -2371,6 +2371,13 @@ static void ide_restart_bh(void *opaque)
      * called function can set a new error status. */
     bus->error_status = 0;
 
+    /* The HBA has generically asked to be kicked on retry */
+    if (error_status & IDE_RETRY_HBA) {
+        if (s->bus->dma->ops->restart) {
+            s->bus->dma->ops->restart(s->bus->dma);
+        }
+    }
+
     if (error_status & IDE_RETRY_DMA) {
         if (error_status & IDE_RETRY_TRIM) {
             ide_restart_dma(s, IDE_DMA_TRIM);
diff --git a/hw/ide/internal.h b/hw/ide/internal.h
index 3736e1bbfe..30fdcbc5fa 100644
--- a/hw/ide/internal.h
+++ b/hw/ide/internal.h
@@ -436,6 +436,7 @@ struct IDEDMAOps {
     DMAInt32Func *prepare_buf;
     DMAu32Func *commit_buf;
     DMAIntFunc *rw_buf;
+    DMAVoidFunc *restart;
     DMAVoidFunc *restart_dma;
     DMAStopFunc *set_inactive;
     DMAVoidFunc *cmd_done;
@@ -499,6 +500,7 @@ struct IDEDevice {
 #define IDE_RETRY_READ  0x20
 #define IDE_RETRY_FLUSH 0x40
 #define IDE_RETRY_TRIM 0x80
+#define IDE_RETRY_HBA  0x100
 
 static inline IDEState *idebus_active_if(IDEBus *bus)
 {