From: Aurelien Jarno <aurelien@aurel32.net>
To: Shin-ichiro KAWASAKI <kawasaki@juno.dti.ne.jp>
Cc: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH] SH4: Serial controller improvement and sleep op bug fix
Date: Sun, 14 Sep 2008 18:56:48 +0200 [thread overview]
Message-ID: <20080914165648.GE22422@volta.aurel32.net> (raw)
In-Reply-To: <48C3E662.1010204@juno.dti.ne.jp>
On Sun, Sep 07, 2008 at 11:34:10PM +0900, Shin-ichiro KAWASAKI wrote:
> Hi, all.
Hi!
> This patch adds some improvements for SH4 system emulation.
> It enables serial controller to receive characters, so the
> shell can receive command inputs. Now we can input 'ls'
> and see it works, at last!
>
> It does:
> - adds receive character feature to SH4 SCIF.
> SH4-SCI feature implementation work is left.
> - fixed a bug on 'sleep' instruction, which have caused halt of idle task.
> As i386 'hlt' instruction does, it should save PC before sleep.
Next time could you please send two different patches, so that one patch
doesn't block the other.
> I checked the patch with the kernel at the following URL,
> which I mentioned before.
>
> > > http://www.assembla.com/file/qemu-sh4/3_zImage
> > > The command line is like this.
> > > % ./qemu/sh4-softmmu/qemu-system-sh4 -M r2d -serial vc -serial stdio -m 1024M -kernel zImage
>
> Comments on it and commit to the trunk will be appreciated.
Please find my comments below.
> Regards,
> Shin-ichiro KAWASAKI
>
>
> Index: target-sh4/helper.h
> ===================================================================
> --- target-sh4/helper.h (revision 5132)
> +++ target-sh4/helper.h (working copy)
> @@ -6,7 +6,7 @@
> DEF_HELPER(void, helper_raise_illegal_instruction, (void))
> DEF_HELPER(void, helper_raise_slot_illegal_instruction, (void))
> DEF_HELPER(void, helper_debug, (void))
> -DEF_HELPER(void, helper_sleep, (void))
> +DEF_HELPER(void, helper_sleep, (int32_t))
> DEF_HELPER(void, helper_trapa, (uint32_t))
>
> DEF_HELPER(uint32_t, helper_addv, (uint32_t, uint32_t))
> Index: target-sh4/op_helper.c
> ===================================================================
> --- target-sh4/op_helper.c (revision 5132)
> +++ target-sh4/op_helper.c (working copy)
> @@ -94,10 +94,11 @@
> cpu_loop_exit();
> }
>
> -void helper_sleep(void)
> +void helper_sleep(int32_t next_pc)
Given then env->pc is unsigned, I think it should be uint32_t instead of
int32_t here. Same in helper.h
> {
> env->halted = 1;
> env->exception_index = EXCP_HLT;
> + env-> pc = next_pc;
> cpu_loop_exit();
> }
>
> Index: target-sh4/translate.c
> ===================================================================
> --- target-sh4/translate.c (revision 5132)
> +++ target-sh4/translate.c (working copy)
> @@ -446,7 +446,7 @@
> return;
> case 0x001b: /* sleep */
> if (ctx->memidx) {
> - tcg_gen_helper_0_0(helper_sleep);
> + tcg_gen_helper_0_1(helper_sleep, tcg_const_i32(ctx->pc + 2));
> } else {
> tcg_gen_helper_0_0(helper_raise_illegal_instruction);
> ctx->bstate = BS_EXCP;
> Index: hw/sh_serial.c
> ===================================================================
> --- hw/sh_serial.c (revision 5132)
> +++ hw/sh_serial.c (working copy)
> @@ -53,6 +53,7 @@
> int freq;
> int feat;
> int flags;
> + int rtrg;
>
> CharDriverState *chr;
>
> @@ -80,6 +81,7 @@
> s->brr = val;
> return;
> case 0x08: /* SCR */
> + /* TODO : For SH7751, SCIF mask should be 0xfb. */
> s->scr = val & ((s->feat & SH_SERIAL_FEAT_SCIF) ? 0xfa : 0xff);
> if (!(val & (1 << 5)))
> s->flags |= SH_SERIAL_FLAG_TEND;
> @@ -89,6 +91,9 @@
> else if (!(val & (1 << 7)) && s->txi->asserted)
> sh_intc_toggle_source(s->txi, 0, -1);
> }
> + if (!(val & (1 << 6)) && s->rxi->asserted) {
> + sh_intc_toggle_source(s->rxi, 0, -1);
> + }
> return;
> case 0x0c: /* FTDR / TDR */
> if (s->chr) {
> @@ -117,12 +122,37 @@
> s->flags &= ~SH_SERIAL_FLAG_RDF;
> if (!(val & (1 << 0)))
> s->flags &= ~SH_SERIAL_FLAG_DR;
> +
> + if (!(val & (1 << 1)) || !(val & (1 << 0))) {
> + if (s->rxi && s->rxi->asserted) {
> + sh_intc_toggle_source(s->rxi, 0, -1);
> + }
> + }
> return;
> case 0x18: /* FCR */
> s->fcr = val;
> + switch ((val >> 6) & 3) {
> + case 0:
> + s->rtrg = 1;
> + break;
> + case 1:
> + s->rtrg = 4;
> + break;
> + case 2:
> + s->rtrg = 8;
> + break;
> + case 3:
> + s->rtrg = 14;
> + break;
> + }
> + if (val & (1 << 1)) {
> + s->rx_cnt = 0;
> + s->sr &= ~(1 << 1);
> + }
> +
> return;
> case 0x20: /* SPTR */
> - s->sptr = val;
> + s->sptr = val & 0xf3;
> return;
> case 0x24: /* LSR */
> return;
> @@ -194,6 +224,15 @@
> s->flags |= SH_SERIAL_FLAG_TDE | SH_SERIAL_FLAG_TEND;
>
> break;
> + case 0x14:
> + if (s->rx_cnt > 0) {
> + ret = s->rx_fifo[0];
> + s->rx_cnt--;
> + memmove(&s->rx_fifo[0], &s->rx_fifo[1], s->rx_cnt);
I don't think moving the whole buffer each time a character is read is a
good idea. I would suggest to use a circular buffer instead. Have a look
at hw/serial.c how it is done.
> + if (s->rx_cnt < s->rtrg)
> + s->flags &= ~SH_SERIAL_FLAG_RDF;
> + }
> + break;
> #if 0
> case 0x18:
> ret = s->fcr;
> @@ -219,6 +258,9 @@
> case 0x10:
> ret = 0;
> break;
> + case 0x14:
> + ret = s->rx_fifo[0];
> + break;
> case 0x1c:
> ret = s->sptr;
> break;
> @@ -240,15 +282,30 @@
>
> static int sh_serial_can_receive(sh_serial_state *s)
> {
> - return 0;
> + return s->scr & (1 << 4);
> }
>
> static void sh_serial_receive_byte(sh_serial_state *s, int ch)
> {
> + if (s->feat & SH_SERIAL_FEAT_SCIF) {
> + if (s->rx_cnt < sizeof(s->rx_fifo)) {
> + s->rx_fifo[s->rx_cnt++] = ch;
> + if (s->rx_cnt >= s->rtrg) {
> + s->flags |= SH_SERIAL_FLAG_RDF;
> + if (s->scr & (1 << 6) && s->rxi) {
> + sh_intc_toggle_source(s->rxi, 0, 1);
> + }
> + }
> + }
> + } else {
> + s->rx_fifo[0] = ch;
> + }
> }
>
> static void sh_serial_receive_break(sh_serial_state *s)
> {
> + if (s->feat & SH_SERIAL_FEAT_SCIF)
> + s->sr |= (1 << 4);
> }
>
> static int sh_serial_can_receive1(void *opaque)
> @@ -313,6 +370,7 @@
> s->base = base;
> s->feat = feat;
> s->flags = SH_SERIAL_FLAG_TEND | SH_SERIAL_FLAG_TDE;
> + s->rtrg = 1;
>
> s->smr = 0;
> s->brr = 0xff;
> @@ -323,6 +381,7 @@
> s->fcr = 0;
> }
> else {
> + memset(s->rx_fifo, 0, sizeof(s->rx_fifo));
> s->dr = 0xff;
> }
>
>
>
>
--
.''`. Aurelien Jarno | GPG: 1024D/F1BCDB73
: :' : Debian developer | Electrical Engineer
`. `' aurel32@debian.org | aurelien@aurel32.net
`- people.debian.org/~aurel32 | www.aurel32.net
next prev parent reply other threads:[~2008-09-14 16:56 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-09-07 14:34 [Qemu-devel] [PATCH] SH4: Serial controller improvement and sleep op bug fix Shin-ichiro KAWASAKI
2008-09-14 16:56 ` Aurelien Jarno [this message]
2008-09-15 5:28 ` [PATCH] [RESEND] SH4 : sleep instruction bug fix (was Re: [Qemu-devel] [PATCH] SH4: Serial controller improvement and sleep op bug fix) Shin-ichiro KAWASAKI
2008-09-15 6:42 ` Aurelien Jarno
2008-09-15 5:42 ` [Qemu-devel] [PATCH] [RESEND] SH4 : Serial controller improvement Shin-ichiro KAWASAKI
2008-09-15 7:06 ` Aurelien Jarno
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=20080914165648.GE22422@volta.aurel32.net \
--to=aurelien@aurel32.net \
--cc=kawasaki@juno.dti.ne.jp \
--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.