All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jan Kiszka <jan.kiszka@web.de>
To: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH] ide: Enable byte&word access to DMA	address register
Date: Tue, 06 May 2008 00:39:08 +0200	[thread overview]
Message-ID: <481F8C8C.60602@web.de> (raw)
In-Reply-To: <20080505212004.GA25810@volta.aurel32.net>

Aurelien Jarno wrote:
> On Mon, May 05, 2008 at 10:20:39AM +0200, Jan Kiszka wrote:
>> According to the specs, also byte- and word-wise access to the busmaster
>> DMA address register is allowed. Patch below fixes the IDE emulation
>> in this regard (avoiding to touch the existing common case of 32-bit
>> access) and makes our guest happy.
>>
>> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
>> ---
>>  hw/ide.c |   25 +++++++++++++++++++++++++
>>  1 file changed, 25 insertions(+)
>>
>> Index: b/hw/ide.c
>> ===================================================================
>> --- a/hw/ide.c
>> +++ b/hw/ide.c
>> @@ -2838,6 +2838,29 @@ static void bmdma_writeb(void *opaque, u
>>      }
>>  }
>>  
>> +static uint32_t bmdma_addr_readb(void *opaque, uint32_t addr)
>> +{
>> +    BMDMAState *bm = opaque;
>> +    uint32_t val;
>> +    val = (bm->addr >> ((addr & 3) * 8)) & 0xff;
>> +#ifdef DEBUG_IDE
>> +    printf("%s: 0x%08x\n", __func__, val);
>> +#endif
>> +    return val;
>> +}
>> +
>> +static void bmdma_addr_writeb(void *opaque, uint32_t addr, uint32_t val)
>> +{
>> +    BMDMAState *bm = opaque;
>> +    int shift = (addr & 3) * 8;
>> +#ifdef DEBUG_IDE
>> +    printf("%s: 0x%08x\n", __func__, val);
>> +#endif
>> +    bm->addr &= ~(0xFF << shift);
>> +    bm->addr |= (val & 0xfc) << shift;
> 
> Are you sure it is correct? If you want to make sure the 2 lowest bits
> are 0, it should be instead:
> 
>        bm->addr |= ((val & 0xFF) << shift) & ~3;

Oh, oh, oh. Of course. I just wonder why my colleague was able to work
with it despite of this bug. Looks like we have been lucky. Find
corrected patch below.

> 
>> +    bm->cur_addr = bm->addr;
>> +}
>> +
>>  static uint32_t bmdma_addr_readl(void *opaque, uint32_t addr)
>>  {
>>      BMDMAState *bm = opaque;
>> @@ -2876,6 +2899,8 @@ static void bmdma_map(PCIDevice *pci_dev
>>          register_ioport_write(addr + 1, 3, 1, bmdma_writeb, bm);
>>          register_ioport_read(addr, 4, 1, bmdma_readb, bm);
>>  
>> +        register_ioport_write(addr + 4, 4, 1, bmdma_addr_writeb, bm);
>> +        register_ioport_read(addr + 4, 4, 1, bmdma_addr_readb, bm);
>>          register_ioport_write(addr + 4, 4, 4, bmdma_addr_writel, bm);
>>          register_ioport_read(addr + 4, 4, 4, bmdma_addr_readl, bm);
>>          addr += 8;
>>
> 
> Otherwise, looks ok. Are word accesses supported? If yes it may be nice
> to implement bmdma_addr_writew and bmdma_addr_readw at the same time.

Yes, word access is supported, the default handler will simply call
twice into our byte handler.

Thanks!
Jan

---
 hw/ide.c |   25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)

Index: b/hw/ide.c
===================================================================
--- a/hw/ide.c
+++ b/hw/ide.c
@@ -2838,6 +2838,29 @@ static void bmdma_writeb(void *opaque, u
     }
 }
 
+static uint32_t bmdma_addr_readb(void *opaque, uint32_t addr)
+{
+    BMDMAState *bm = opaque;
+    uint32_t val;
+    val = (bm->addr >> ((addr & 3) * 8)) & 0xff;
+#ifdef DEBUG_IDE
+    printf("%s: 0x%08x\n", __func__, val);
+#endif
+    return val;
+}
+
+static void bmdma_addr_writeb(void *opaque, uint32_t addr, uint32_t val)
+{
+    BMDMAState *bm = opaque;
+    int shift = (addr & 3) * 8;
+#ifdef DEBUG_IDE
+    printf("%s: 0x%08x\n", __func__, val);
+#endif
+    bm->addr &= ~(0xFF << shift);
+    bm->addr |= ((val & 0xFF) << shift) & ~3;
+    bm->cur_addr = bm->addr;
+}
+
 static uint32_t bmdma_addr_readl(void *opaque, uint32_t addr)
 {
     BMDMAState *bm = opaque;
@@ -2876,6 +2899,8 @@ static void bmdma_map(PCIDevice *pci_dev
         register_ioport_write(addr + 1, 3, 1, bmdma_writeb, bm);
         register_ioport_read(addr, 4, 1, bmdma_readb, bm);
 
+        register_ioport_write(addr + 4, 4, 1, bmdma_addr_writeb, bm);
+        register_ioport_read(addr + 4, 4, 1, bmdma_addr_readb, bm);
         register_ioport_write(addr + 4, 4, 4, bmdma_addr_writel, bm);
         register_ioport_read(addr + 4, 4, 4, bmdma_addr_readl, bm);
         addr += 8;

      reply	other threads:[~2008-05-05 22:39 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-05-05  8:20 [Qemu-devel] [PATCH] ide: Enable byte&word access to DMA address register Jan Kiszka
2008-05-05 21:20 ` Aurelien Jarno
2008-05-05 22:39   ` Jan Kiszka [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=481F8C8C.60602@web.de \
    --to=jan.kiszka@web.de \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.