From: Paolo Bonzini <pbonzini@redhat.com>
To: "Hervé Poussineau" <hpoussin@reactos.org>, qemu-devel@nongnu.org
Cc: Leon Alrae <leon.alrae@imgtec.com>,
Laurent@Vivier.EU, Aurelien Jarno <aurelien@aurel32.net>
Subject: Re: [Qemu-devel] [PATCH 1/8] rc4030: create custom DMA address space
Date: Thu, 26 Mar 2015 15:27:23 +0100 [thread overview]
Message-ID: <5514174B.9010406@redhat.com> (raw)
In-Reply-To: <1425593606-5909-2-git-send-email-hpoussin@reactos.org>
On 05/03/2015 23:13, Hervé Poussineau wrote:
> +static const MemoryRegionOps rc4030_dma_tt_ops = {
> + .impl.min_access_size = 4,
> + .impl.max_access_size = 4,
> + .impl.max_access_size = 4,
> +};
> +
> +static void rc4030_dma_tt_update(rc4030State *s, uint32_t new_tl_base,
> + uint32_t new_tl_limit)
> +{
> + int entries, i;
> + dma_pagetable_entry *dma_tl_contents;
> +
> + if (s->dma_tl_limit) {
> + /* write old dma tl table to physical memory */
> + memory_region_del_subregion(get_system_memory(), &s->dma_tt);
> + cpu_physical_memory_write(s->dma_tl_limit & 0x7fffffff,
> + memory_region_get_ram_ptr(&s->dma_tt),
> + s->dma_tl_limit);
You would need object_unparent(&s->dma_tt) here.
However, this breaks the rules for memory region lifetime (see
docs/memory.txt).
One solution is to warn here for a large dma_tl_limit and always create
a 4K ROMD region. Here you can create an alias into the actual dma_tt
region and add/remove/unparent that alias. Aliases can be created and
unparented at will.
> @@ -733,7 +793,16 @@ static void rc4030_do_dma(void *opaque, int n, uint8_t *buf, int len, int is_wri
> dma_addr = s->dma_regs[n][DMA_REG_ADDRESS];
>
> /* Read/write data at right place */
> - rc4030_dma_memory_rw(opaque, dma_addr, buf, len, is_write);
> + for (i = 0; i < len; ) {
> + int ncpy = DMA_PAGESIZE - (dma_addr & (DMA_PAGESIZE - 1));
> + if (ncpy > len - i) {
> + ncpy = len - i;
> + }
> + address_space_rw(&s->dma_as, dma_addr, buf + i, ncpy, is_write);
> +
> + dma_addr += ncpy;
> + i += ncpy;
> + }
>
> s->dma_regs[n][DMA_REG_ENABLE] |= DMA_FLAG_TC_INTR;
> s->dma_regs[n][DMA_REG_COUNT] -= len;
The loop should not be necessary, address_space_rw does the same.
Paolo
> @@ -800,6 +869,7 @@ void *rc4030_init(qemu_irq timer, qemu_irq jazz_bus,
> MemoryRegion *sysmem)
> {
> rc4030State *s;
> + int i;
>
> s = g_malloc0(sizeof(rc4030State));
>
> @@ -821,5 +891,15 @@ void *rc4030_init(qemu_irq timer, qemu_irq jazz_bus,
> "rc4030.jazzio", 0x00001000);
> memory_region_add_subregion(sysmem, 0xf0000000, &s->iomem_jazzio);
>
> + memory_region_init(&s->dma_tt, NULL, "dma_tt", 0);
> + memory_region_init(&s->dma_mr, NULL, "dma", INT32_MAX);
> + for (i = 0; i < MAX_TL_ENTRIES; ++i) {
> + memory_region_init_alias(&s->dma_mrs[i], NULL, "dma-alias",
> + get_system_memory(), 0, DMA_PAGESIZE);
> + memory_region_set_enabled(&s->dma_mrs[i], false);
> + memory_region_add_subregion(&s->dma_mr, i * DMA_PAGESIZE,
> + &s->dma_mrs[i]);
> + }
> + address_space_init(&s->dma_as, &s->dma_mr, "rc4030_dma");
> return s;
> }
>
next prev parent reply other threads:[~2015-03-26 14:27 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-03-05 22:13 [Qemu-devel] [PATCH 0/8] net/dp8393x improvements Hervé Poussineau
2015-03-05 22:13 ` [Qemu-devel] [PATCH 1/8] rc4030: create custom DMA address space Hervé Poussineau
2015-03-25 14:45 ` Paolo Bonzini
2015-03-25 19:10 ` Hervé Poussineau
2015-03-26 14:15 ` Paolo Bonzini
2015-03-26 14:27 ` Paolo Bonzini [this message]
2015-03-05 22:13 ` [Qemu-devel] [PATCH 2/8] rc4030: use AddressSpace and address_space_rw in users Hervé Poussineau
2015-03-05 22:13 ` [Qemu-devel] [PATCH 3/8] net/dp8393x: always calculate proper checksums Hervé Poussineau
2015-03-05 22:13 ` [Qemu-devel] [PATCH 4/8] net/dp8393x: do not use old_mmio accesses Hervé Poussineau
2015-03-05 22:13 ` [Qemu-devel] [PATCH 5/8] net/dp8393x: use dp8393x_ prefix for all functions Hervé Poussineau
2015-03-05 22:13 ` [Qemu-devel] [PATCH 6/8] net/dp8393x: QOM'ify Hervé Poussineau
2015-03-05 22:13 ` [Qemu-devel] [PATCH 7/8] net/dp8393x: add PROM to store MAC address Hervé Poussineau
2015-03-05 22:13 ` [Qemu-devel] [PATCH 8/8] net/dp8393x: add load/save support Hervé Poussineau
2015-03-25 14:13 ` [Qemu-devel] [PATCH 0/8] net/dp8393x improvements Leon Alrae
2015-04-29 14:56 ` Leon Alrae
2015-05-14 21:11 ` Hervé Poussineau
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=5514174B.9010406@redhat.com \
--to=pbonzini@redhat.com \
--cc=Laurent@Vivier.EU \
--cc=aurelien@aurel32.net \
--cc=hpoussin@reactos.org \
--cc=leon.alrae@imgtec.com \
--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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).